Information processing system

Abstract

The application provides an information processing system. An information processing system, characterized by comprising: a processor; wherein the processor is configured to receive input information of a user, generate prompt text for indicating that a generative artificial intelligence model generates a mountaineering plan; the processor is configured to input the generated prompt text to the generative artificial intelligence model to generate a mountaineering plan; the processor is configured to automatically generate a plan based on the generated mountaineering plan, and submit the plan to an emergency agency.

Description

Technical Field

[0001] The technology disclosed herein relates to an information processing system. Background Technology

[0002] Japanese Patent Application Publication No. 2022-180282 discloses a method for controlling a role-based chatbot executed by at least one processor. The method includes the following steps: receiving a user's speech; adding the user's speech to a prompt word, the prompt word containing instruction statements associated with an explanation of the chatbot's role; encoding the prompt word; and inputting the encoded prompt word into a language model to generate a chatbot speech in response to the user's speech.

[0003] In existing mountaineering activity management methods, climbers typically need to manually search for a large amount of scattered information, such as suitable climbing routes, itinerary schedules, required equipment, weather and risk warnings, making it difficult to obtain personalized and systematic climbing plans in a timely manner. Furthermore, before a climb, climbers often need to write and submit a climbing plan to relevant emergency agencies (such as police and rescue organizations). This process relies on personal experience and subjective judgment, resulting in incomplete content, inaccurate information, and untimely submission, making it difficult to provide useful information to emergency agencies in a timely manner, thus failing to effectively support rescue decisions in the event of an emergency. Moreover, existing technologies for recording location information during the climb largely rely on manual user operation or independent third-party applications, lacking integrated management linked to the climbing plan. This makes it difficult to automatically generate complete movement route records after the climb for safety review, risk analysis, and subsequent itinerary improvement. In summary, the technical challenge this invention aims to address is how to automatically generate and submit climbing plans using generative artificial intelligence models, and how to link these plans with user terminal location information to achieve unified management of pre-climb planning, in-climb information accumulation, and post-climb route recording. Summary of the Invention

[0004] To address the aforementioned technical challenges, this invention provides an information processing system. The system includes a processor configured to receive user input, generate prompt text instructing a generative artificial intelligence (AI) model to generate a mountain climbing plan, input the generated prompt text into the AI ​​model to generate the mountain climbing plan, automatically generate a plan document based on the generated plan, and submit the plan document to emergency agencies. Through this configuration, the system can automatically generate personalized, structured mountain climbing plans using a generative AI model based on user-inputted information such as climbing objectives, climbing level, climbing style, and date. Furthermore, it can automatically generate a mountain climbing plan document that conforms to the format required by emergency agencies, thereby reducing manual writing work for users and improving the completeness and accuracy of the plan document content.

[0005] In a preferred embodiment, the processor is configured to obtain emergency contact information and facility information for the corresponding region based on the mountaineering plan, and provide the information to the user. In this way, the system can automatically obtain contact information and locations of local police agencies, hospitals, rescue organizations, and other emergency facilities from a pre-built database or external information sources based on the mountainous or administrative regions involved in the mountaineering plan, and present this information to the user in list or map form. This allows the user to have complete emergency contact information before climbing, improving the safety and controllability of the mountaineering activity.

[0006] In another preferred embodiment, the processor is configured to periodically acquire location information from the user terminal and generate a movement route record based on the location information after the climb. In this way, the system continuously accumulates climbing trajectory data during the climb by periodically reporting location information from the user terminal; after the climb, the system sorts the location information by time and calculates the trajectory, generating a movement route record that includes the start and end times of the climb, total time, and actual movement path, which can be compared and analyzed with a previously generated climbing plan. Through these methods, the present invention achieves integrated management from pre-climb planning, in-climb trajectory recording to post-climb route review, effectively improving the safety management level and informatization level of climbing activities.

[0007] "System" refers to a whole device or collection of devices consisting of hardware and / or software, used to perform functions such as mountaineering plan generation, automatic generation and submission of plan documents, emergency information provision, location information recording and route generation.

[0008] A processor is an electronic computing unit that can execute program instructions, perform calculations and logical processing on input data, and control the coordinated operation of various functional modules of the system. It can be a single physical processor, a combination of multiple processors, or processing resources including CPU, GPU, application-specific integrated circuits, etc.

[0009] "User" refers to an individual or group member who uses the system to plan a mountain climb, obtain emergency information, record mountain climbing routes, and view mountain climbing records.

[0010] "Input information" refers to data related to mountaineering activities provided by users to the system through their terminals, including but not limited to the target mountain, climbing date, climbing time, climbing level, climbing style, number of people, mode of transportation, and other optional explanatory information.

[0011] "Generative AI models" refer to AI models trained on large-scale data that can automatically generate text content related to the mountain climbing plan based on input prompts, including but not limited to large language models, multimodal generative models, or combinations thereof.

[0012] "Prompt text" refers to text data generated by the processor based on user input information, used to explicitly explain the requirements and constraints for generating a mountain climbing plan to the generative artificial intelligence model, so as to guide the generative artificial intelligence model to output a mountain climbing plan that meets the requirements.

[0013] "Mountain climbing plan" refers to structured or semi-structured plan information for a specific mountain climbing activity, generated by a generative artificial intelligence model based on prompt text. It includes, but is not limited to, mountain climbing schedule, route suggestions, rest stops, equipment recommendations, and risk warnings.

[0014] A “plan” is a formal document automatically generated by the processor based on a mountaineering plan and submitted to emergency agencies. It includes at least the climber’s basic information, climbing date and time, climbing route, number of people in the climbing party, and emergency contact information.

[0015] "Emergency agencies" refer to organizations or units that can provide support services such as rescue, police or medical care in the event of accidents or dangers during mountaineering activities, including but not limited to police agencies, fire and rescue agencies, mountain rescue teams and medical institutions.

[0016] "Emergency contact information" refers to the information required to contact emergency agencies, including but not limited to telephone numbers, fax numbers, email addresses, online contact information, and emergency contact channels.

[0017] "Facility information" refers to information related to facilities provided by emergency agencies or public service agencies that are relevant to mountaineering activities, including but not limited to the name, location, address and service area of ​​hospitals, rescue stations, police stations, shelters or first aid points.

[0018] "User terminal" refers to electronic devices used by users to input information, view results, and report location information, including but not limited to smartphones, tablets, laptops, wearable devices, or other terminal devices equipped with communication and positioning functions.

[0019] "Location information" refers to data related to the geographical location of the user terminal during the mountain climbing process, including at least longitude, latitude and time information corresponding to the location, and may further include altitude information and positioning accuracy.

[0020] "Periodic acquisition of location information" refers to the process by which the processor receives or collects location information from the user terminal multiple times within a preset time interval to form a sequence of location information that changes over time.

[0021] "Movement route record" refers to a data set or visualization result generated by the processor based on the location information obtained during the mountaineering process, after time sorting, trajectory calculation and statistical processing, to reflect the actual route and time information of the mountaineer, including at least the start and end time of the climb, the total time taken and the route trajectory. Attached Figure Description

[0022] Figure 1 This is a conceptual diagram illustrating an example of the configuration of the data processing system according to the first embodiment.

[0023] Figure 2 This is a conceptual diagram illustrating an example of the main functions of the data processing apparatus and smart device according to the first embodiment.

[0024] Figure 3 This is a conceptual diagram illustrating an example of the configuration of the data processing system according to the second embodiment.

[0025] Figure 4 This is a conceptual diagram illustrating an example of the main functions of the data processing device and smart glasses according to the second embodiment.

[0026] Figure 5 This is a conceptual diagram illustrating an example of the configuration of the data processing system according to the third embodiment.

[0027] Figure 6 This is a conceptual diagram illustrating an example of the main functions of the data processing apparatus and head-mounted terminal according to the third embodiment.

[0028] Figure 7 This is a conceptual diagram illustrating an example of the configuration of the data processing system according to the fourth embodiment.

[0029] Figure 8 This is a conceptual diagram illustrating an example of the main functions of the data processing device and robot according to the fourth embodiment.

[0030] Figure 9 This represents an emotion map that maps multiple emotions.

[0031] Figure 10 This represents an emotion map that maps multiple emotions.

[0032] Figure 11 This is a sequence diagram illustrating the processing flow of the data processing system of the first embodiment.

[0033] Figure 12This is a sequence diagram illustrating the processing flow of the data processing system in Application Example 1.

[0034] Figure 13 This is a sequence diagram illustrating the processing flow of the data processing system of the second embodiment.

[0035] Figure 14 This is a sequence diagram illustrating the processing flow of the data processing system in Application Example 2. Detailed Implementation

[0036] Hereinafter, an example of an implementation of the system according to the present disclosure will be described with reference to the accompanying drawings.

[0037] First, let me explain the terminology used in the following instructions.

[0038] In the following embodiments, the processor (hereinafter referred to as "processor") with reference numerals may be a single computing device or a combination of multiple computing devices. Furthermore, the processor may be a single computing device or a combination of multiple computing devices. Examples of computing devices include CPU (Central Processing Unit), GPU (Graphics Processing Unit), GPGPU (General-Purpose computing on Graphics Processing Units), APU (Accelerated Processing Unit), etc.

[0039] In the following embodiments, RAM (Random Access Memory), as indicated in the figures, is a memory that temporarily stores information and is used as working memory by the processor.

[0040] In the following embodiments, the memory, as indicated by the reference numerals, is one or more non-volatile storage devices that store various programs and parameters. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), disks (e.g., hard disks), or magnetic tapes.

[0041] In the following embodiments, the communication I / F (Interface) with reference numerals is an interface that includes a communication processor and an antenna, etc. The communication I / F is responsible for communication between multiple computers. As an example of a communication specification applicable to the communication I / F, wireless communication specifications such as 5G (5th Generation Mobile Communication System), Wi-Fi (wireless fidelity) (registered trademark), or Bluetooth (registered trademark) can be listed.

[0042] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B". That is, "A and / or B" means that it can be only A, only B, or a combination of A and B. Furthermore, in this specification, when "and / or" connects to express more than three items, the same interpretation as "A and / or B" applies.

[0043] First Implementation Method Figure 1 An example of the configuration of the data processing system 10 according to the first embodiment is shown.

[0044] like Figure 1 As shown, the data processing system 10 includes a data processing device 12 and an intelligent device 14. A server can be cited as an example of the data processing device 12.

[0045] The data processing apparatus 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" as understood in this disclosure. The computer 22 includes a processor 28, RAM 30, and memory 32. The processor 28, RAM 30, and memory 32 are connected to a bus 34. Furthermore, the database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0046] The smart device 14 includes a computer 36, a receiving device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, RAM 48, and memory 50. The processor 46, RAM 48, and memory 50 are connected to a bus 52. In addition, the receiving device 38, output device 40, camera 42, and communication I / F 44 are also connected to the bus 52.

[0047] The receiving device 38 includes a touchscreen 38A and a microphone 38B, and receives user input. The touchscreen 38A receives user input via touch by detecting contact with an indicator (e.g., a pen or finger). The microphone 38B receives user input via sound by detecting the user's voice. The control unit 46A in the processor 46 sends data representing the user input received by the touchscreen 38A and microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data representing the user input.

[0048] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting data in a form perceptible to the user 20 (e.g., sound and / or text). The display 40A displays visual information such as text and images according to instructions from the processor 46. The speaker 40B outputs sound according to instructions from the processor 46. The camera 42 is a small digital camera equipped with an optical system such as a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0049] Communication I / F44 is connected to network 54. Communication I / F44 and 26 are responsible for sending and receiving various information between processor 46 and processor 28 via network 54.

[0050] Figure 2 The diagram shows an example of the main functions of the data processing device 12 and the smart device 14.

[0051] like Figure 2 As shown, in the data processing apparatus 12, specific processing is performed by the processor 28. A specific processing program 56 is stored in the memory 32. The specific processing program 56 is an example of a "program" as understood in this disclosure. The processor 28 reads the specific processing program 56 from the memory 32 and executes the read specific processing program 56 on the RAM 30. Specific processing is implemented by the processor 28 operating as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0052] The memory 32 stores a data generation model 58 and an emotion-specific model 59. The data generation model 58 and the emotion-specific model 59 are used by the specific processing unit 290. The specific processing unit 290 can use the emotion-specific model 59 to infer the user's emotions and perform specific processing based on the user's emotions. In the emotion inference function (emotion-specific function) using the emotion-specific model 59, various inferences and predictions related to the user's emotions are performed, including inferences and predictions of the user's emotions, but this is not limited to this example. Furthermore, emotion inference and prediction may also include, for example, emotion analysis (parsing).

[0053] In the smart device 14, the processor 46 performs the acceptance output processing. The memory 50 stores the acceptance output program 60. The acceptance output program 60 is used in conjunction with the data processing system 10 and the specific processing program 56. The processor 46 reads the acceptance output program 60 from the memory 50 and executes the read acceptance output program 60 on the RAM 48. The acceptance output processing is implemented by the processor 46 acting as the control unit 46A according to the acceptance output program 60 executed on the RAM 48. Furthermore, the smart device 14 has the same data generation model and emotion-specific model as the data generation model 58 and the emotion-specific model 59, and these models can also be used to perform the same processing as the specific processing unit 290. The acceptance output processing is implemented by the processor 46 acting as the control unit 46A according to the acceptance output program 60 executed on the RAM 48.

[0054] Alternatively, other devices besides the data processing device 12 may also have the data generation model 58. For example, a server device (e.g., a generation server) may have the data generation model 58. In this case, the data processing device 12 obtains the processing results (prediction results, etc.) using the data generation model 58 by communicating with the server device that has the data generation model 58. Furthermore, the data processing device 12 may be a server device or a user-held terminal device (e.g., a mobile phone, robot, home appliance, etc.). Next, an example of the processing of the data processing system 10 of the first embodiment will be described.

[0055] Example 1 The flow of a specific process in Example 1 will be described. Each part of the system described below is implemented by the data processing device 12 and the smart device 14. Furthermore, the data processing device 12 is referred to as the "server," and the smart device 14 is referred to as the "terminal."

[0056] Traditional mountaineering planning typically relies on users manually retrieving and summarizing information from multiple sources, such as route guides, weather forecasts, equipment recommendations, and local emergency contact information. Even with the introduction of general generative artificial intelligence models, the lack of structured data processing and interaction mechanisms tailored to specific application scenarios still results in the following shortcomings at the computer technology level: (1) The server side usually forwards the user input as simple text directly to the generative artificial intelligence model. It lacks structured representation and unified management of the input information, which leads to unstable expression of prompts and poor controllability and consistency of the generated results, thereby increasing the complexity of the system side in parsing and utilizing the results. (2) The mountaineering plans output by generative artificial intelligence models are mostly long segments of unstructured natural language text. The server lacks an efficient mechanism for differentiation, extraction and reconstruction, making it difficult to modularly display and process them on the terminal side according to dimensions such as "route, time, items carried, and safety precautions". This limits the automatic linkage capability of the plan information with other data sources (such as regional emergency contact information databases). (3) Existing systems are mostly based on one-time generation of results, lacking technical solutions for iterative updates based on subsequent user-added conditions or change requests. The server cannot automatically regenerate prompt statements and stably drive generative artificial intelligence models while maintaining the consistency of structured information, thus making it difficult to achieve an efficient human-machine collaborative optimization process. (4) In scenarios where the network environment is unstable or the terminal is offline, there is a lack of a mechanism for collaborative management of planning information and emergency information between the server and the terminal, which makes it impossible for the terminal to reliably access critical security information in an offline state, and is also not conducive to the server's unified management of user trajectory data and associated storage with planning information. (5) At the computer system level, the existing technology has failed to integrate the data processing flow of “structured management of user input - automatic generation of prompt statements - calling of generative artificial intelligence models - structured reconstruction of results - automatic matching with regional emergency information - multi-round interactive updates - trajectory data association storage”, resulting in scattered server application logic, messy data interfaces, low efficiency of computing resource utilization and low system maintainability.

[0057] Therefore, an improved computer implementation is needed. This can be achieved by introducing structured data management, automatic generation of prompts, automatic differentiation and reconstruction of generated results, and collaborative control mechanisms with external data sources and user terminals on the server side. This would improve the controllability of generative artificial intelligence models and the overall processing efficiency of the system during the generation and management of mountaineering plans, thereby improving the computer technology of information processing in safe mountaineering scenarios.

[0058] The specific processing performed by the specific processing unit 290 of the data processing apparatus 12 in Embodiment 1 is achieved by the following means.

[0059] In this invention, the server includes: a device for acquiring user input information and storing the user input information as structured information; a device for automatically generating prompt statements based on the structured information for input to a generative artificial intelligence model, and sending the prompt statements to the generative artificial intelligence model running on an external information processing device, thereby enabling the generative artificial intelligence model to generate mountain climbing plan information and for the server to acquire the mountain climbing plan information; and a device for performing differentiation and extraction processing on the mountain climbing plan information, reconstructing the mountain climbing plan information into multiple types of plan composition information, including mountain climbing route information, time information, baggage information, and safety precautions information. The system includes: a device for compiling the plan structure information with emergency contact information and facility information stored in the regional information storage unit to form a plan document, automatically generating the plan document into a predetermined format electronic document, sending it to safety-related agencies via a communication network, and distributing it to user terminals; and a device for obtaining additional condition information or change request information from the user terminals, reflecting the additional condition information or change request information in the structured information, regenerating update prompt statements, and re-inputting them into the generative artificial intelligence model to obtain updated mountaineering plan information and iteratively updating the plan structure information. This creates an integrated data processing link on the server side for mountaineering plan scenarios, enabling full-process computer processing from structured management of user input, automatic generation of prompt statements, automatic structured reconstruction of the generative artificial intelligence model output, to automatic linkage with regional emergency information and multi-round interactive updates. This not only improves the controllability and utilization efficiency of the generative artificial intelligence model output content but also improves data synchronization and offline availability between the server and terminals, thereby enhancing the processing performance and reliability of the mountaineering plan generation and safety management system at the computer technology level.

[0060] "System" refers to a combination of devices consisting of one or more information processing devices, storage devices, and communication devices, which collect, process, store, and transmit data through the execution of programs to achieve the functions of mountaineering plan generation and management.

[0061] "User information" refers to the set of data that users input through their terminals and that is obtained by the system to generate a mountaineering plan, including but not limited to information about the target mountain, the climber's skill level, mode of travel, departure time, number of people, and user preferences.

[0062] "Structured information" refers to data that is stored in the form of a specific data structure (such as key-value pairs, records, objects, etc.) after the server has divided the acquired user information into fields, determined the types, and standardized the format.

[0063] "Generative artificial intelligence models" refer to models that are based on machine learning or deep learning techniques and can automatically generate text content based on input prompts by learning from a large amount of training data.

[0064] "Prompt statements" refer to natural language or natural language-like instruction texts automatically generated by the server based on structured information and used as input to generative artificial intelligence models to explicitly require the generative artificial intelligence models to output the content and format of mountain climbing plan information.

[0065] "External information processing device" refers to a computing resource device that is connected to a server via a communication network and runs generative artificial intelligence models on it, including but not limited to computing nodes or remote servers on a cloud computing platform.

[0066] "Mountain climbing plan information" refers to a set of text information related to mountain climbing activities output by a generative artificial intelligence model based on prompts, including at least one or more of the following: mountain climbing route, schedule, suggested items to carry, and safety and weather precautions.

[0067] "Discrimination extraction processing" refers to the process by which the server parses and segments mountaineering plan information according to predetermined rules or features, and identifies and extracts content belonging to different categories from unstructured text.

[0068] "Plan composition information" refers to the set of mountaineering plan elements information that are reconstructed by the server after being differentiated and extracted, and are divided into different functional categories. It includes at least one or more of the following: mountaineering route information, time information, baggage information, and safety precautions information.

[0069] "Mountain climbing route information" refers to the content used to represent the mountain climbing route in the plan's composition information, including but not limited to the start point, end point, places passed through, recommended routes, and return routes.

[0070] "Time information" refers to the content used in the plan's structure information to indicate the time arrangements related to the mountaineering activity, including but not limited to departure time, estimated arrival time, total round-trip time, and the time required for each section of the route.

[0071] "Information on items carried" refers to the information in the plan that relates to the equipment and supplies needed for the mountaineering activity, including but not limited to suggestions or lists of clothing, footwear, lighting tools, food, water, first aid supplies, and other assistive devices.

[0072] "Safety precautions information" refers to the content related to safety risks, preventive measures, and emergency response in the plan's constituent information, including but not limited to weather precautions, warnings of dangerous road sections, evacuation suggestions, and basic emergency handling points.

[0073] "Regional information storage unit" refers to a database or other data storage structure used to store emergency contact information and facility information related to a specific region, either on the server side or in storage resources that can communicate with the server.

[0074] "Emergency contact information" refers to contact data of organizations or institutions stored in the regional information storage unit, used to contact them in case of an emergency during the mountaineering process. This includes, but is not limited to, the organization's name, contact number, communication method, and service hours.

[0075] "Facility information" refers to information stored in the regional information storage unit that relates to facilities around the mountaineering area, including but not limited to the location, type, and basic services of rescue stations, medical institutions, management agencies, and shelters.

[0076] "Plan document information" refers to the documented data structure edited by the server based on the plan composition information and the corresponding results of emergency contact information and facility information, which is used for display or output. It is the content basis for generating electronic documents.

[0077] "Preset format electronic document" refers to an electronic file automatically generated from planned document information according to pre-set layout, field structure and file format rules, including but not limited to text documents, printable documents or other machine-readable documents.

[0078] "Disaster prevention-related agencies" refers to organizations that have functions in disaster prevention, emergency rescue, or safety management within or related to the mountaineering area, including but not limited to emergency management agencies, police agencies, and mountain rescue organizations.

[0079] "User terminal" refers to a terminal device operated by a user and communicating with a server, including but not limited to mobile terminals, computer terminals, or other electronic devices with display and input functions and the ability to access networks.

[0080] "Additional condition information" refers to supplementary data that users can input through their user terminals after the initial mountaineering plan information is generated, in order to modify or refine the conditions for generating the mountaineering plan. This includes, but is not limited to, changing the departure time, route preferences, and intensity requirements.

[0081] "Change Request Information" refers to the data submitted by the user based on the generated mountaineering plan information to adjust or modify the plan content, including but not limited to shortening or extending the trip time, changing the route, or adjusting the requirements for carrying items.

[0082] "Update prompt statement" refers to a prompt statement that the server regenerates based on the original structured information, reflecting additional conditional information or change request information, and is used to re-input into the generative artificial intelligence model to obtain updated mountaineering plan information.

[0083] "Communication network" refers to wired or wireless network infrastructure used for data exchange between servers, external information processing devices and user terminals, including but not limited to local area networks, wide area networks and public data communication networks.

[0084] "Internal storage unit" refers to storage resources located inside a user terminal for locally storing plan structure information or related data, including but not limited to semiconductor memory, magnetic storage media, or their logically partitioned storage areas.

[0085] "Location information" refers to data obtained from the user terminal that indicates the terminal's geographical location, including but not limited to coordinates, timestamps, and related precision information obtained by satellite positioning systems, cellular network positioning, or wireless local area network positioning.

[0086] "Predetermined time interval" refers to a fixed or configurable time period parameter used by the system when collecting location information or performing related processing, such as every few seconds, minutes, or hours.

[0087] "Movement trajectory information" refers to a set of data generated by the server based on location information collected within a predetermined time interval and combined with mountain climbing route information, used to represent the user's movement path and time sequence during the mountain climbing activity.

[0088] In this embodiment, the server functions as the core information processing device. The server can be deployed on a physical server or virtual machine equipped with a multi-core central processing unit (e.g., a general-purpose multi-core CPU) and a graphics processing unit (e.g., a general-purpose GPU accelerator card), running a general-purpose operating system (e.g., a Linux-based server operating system), and communicating with terminals via network server software. The terminal can be a mobile terminal or a computing terminal equipped with a display device and an input device, running a mobile operating system or a desktop operating system. Users interact with the server through the terminal.

[0089] The server pre-stores program modules in the storage device. These modules include: a user information management module, a structured information management module, a prompt statement generation module, a generative artificial intelligence model invocation module, a plan information parsing and reconstruction module, a region information matching module, a document generation module, a terminal synchronization control module, and a location and trajectory management module. Each module executes on the processor as a process or thread, exchanging data through shared memory or message queues.

[0090] When the server receives user input, it receives request data sent by the terminal via a communication network. In this embodiment, the terminal uses graphical user interface components (such as text input boxes, drop-down selection boxes, time selection controls, etc.) to guide the user in inputting information related to mountaineering. The user inputs information such as the target mountain, climber's skill level, climbing style, scheduled departure time, number of participants, and route preferences through the terminal. The terminal locally converts the user input into a field-based data structure and sends it to the server via a security protocol.

[0091] After receiving user input, the server uses a parsing module to convert the request payload from the transmission format into an internal data structure. The server performs data type determination, validity checks, and default value completion for each field. For example, the server normalizes "target mountain" into a standard geographic entity identifier, maps "climber level" to a predefined level enumeration, and maps "climbing style" to a time range and schedule structure. After completing this series of data processing steps, the server stores the results in a structured information management module. This structured information can take the form of a key-value pair set, a record table, or an object-oriented data representation.

[0092] When generating prompt statements, the server maps the fields in the structured information to natural language parameter slots according to a predefined template. The server predefines multiple prompt statement templates for different application conditions; for example, different constraint descriptions are used for scenarios such as single-day hikes, multi-day trips, and hikes with children. Through string concatenation, placeholder replacement, and syntax rule processing, the server transforms abstract data into prompt statements with specific content and structure required from the generative AI model. The server can generate the following example prompt statements: "Please create a day trip plan for Mt. Fuji for an intermediate-level climber, departing around 5:00 AM, with a team of 2 people. Please provide a detailed recommended route (including the starting point, main sections, and return route), round-trip time, a list of recommended equipment, and weather and safety precautions to be aware of." The server can also generate update prompts based on user-added conditions. For example, when a user wants to depart earlier and avoid crowded routes, the server can generate: "Please redesign your mountain climbing plan under the following constraints: the mountain is Mt. Fuji, the climber's skill level is intermediate, the climb is a day climb, the departure time is approximately 04:00, and you prefer routes with fewer people and less steep slopes. Please provide an updated recommended route, schedule, equipment suggestions, and weather and safety precautions." When the server invokes the generative AI model, it accesses the generative AI model service deployed on an external information processing device via a network interface. In this embodiment, the generative AI model is a neural network model based on a transformer structure, containing a multi-layer self-attention encoder-decoder structure to process the sequence features of the prompt statements. The server segments and encodes the prompt statements into vector sequences, which are then used as model input. Internally, the model uses multi-head attention mechanisms, feedforward networks, and layer normalization to calculate the contextual relationships of the input sequences and generate corresponding output sequences. During the training phase, the model uses supervised learning methods, employing a large number of mountain climbing plan descriptions, route instructions, and safety guide texts as training samples. The difference between the predicted and target sequences is measured using a cross-entropy loss function, and the network weights are updated using a gradient descent-based optimization algorithm. Data augmentation methods can be used during training, such as providing text samples with different wording and timing for the same mountain climbing scenario, to improve the model's robustness and generalization ability under diverse inputs.

[0093] When the server invokes the model during the inference phase, it sets inference parameters such as temperature, maximum output length, and weights for penalizing repetitions to control the diversity and consistency of generated text. Because the server explicitly specifies the output content structure and constraints in the prompts, the generative AI model provides relatively stable segmented descriptions of the mountaineering plan information according to dimensions such as route, time, equipment, and safety. This reduces the difficulty of subsequent server parsing and improves overall processing efficiency.

[0094] After obtaining the mountaineering plan information, the server uses a plan information parsing and reconstruction module to perform structured processing on the generated natural language. This module combines rule-based segmentation algorithms with machine learning classifiers: the server first performs coarse segmentation of the text based on expected chapter markers or key phrases (such as "route," "time," "equipment," "precautions," etc.), and then uses a trained text classification model or pattern matching algorithm to classify the content of each segment. The server stores the content of different categories into the corresponding plan composition information fields, forming structured subsets such as mountaineering route information, time information, equipment information, and safety precautions information.

[0095] When performing this structured reconstruction, the server uses explicit data structures to represent each type of information. For example, the server represents mountain route information as an ordered list containing multiple route node segments, each node including parameters such as geographic location identifier, altitude estimate, and route difficulty index; the server represents time information as a vector of start time, arrival time, and time spent on each route segment; the server represents carried item information as a set of entries including item name, quantity, weight, and usage description; and the server represents safety precautions information as a list of risk entries associated with categories such as weather conditions, terrain risks, and health risks. Through this structured representation, the server can perform efficient data calculations in subsequent processing (such as route comparison, time estimation optimization, and risk level assessment), rather than simply displaying text.

[0096] During the regional information matching process, the server accesses a regional information storage unit, which can be implemented using a relational database or a key-value database, recording emergency contact information and facility information corresponding to different mountain areas. Based on the target mountain's regional identifier, the server retrieves records of corresponding emergency organizations, rescue facilities, and medical institutions from the database. The server compares these records with the route nodes and time schedules in the plan's composition information to determine emergency facilities adjacent to or overlapping with the user's actual itinerary. The server appends the matching results to the plan's composition information, establishing a mapping from specific route nodes to corresponding emergency organizations in the memory structure, so that the terminal can quickly locate the appropriate contact in an emergency.

[0097] When generating planning document information, the server integrates the planning structure information with the matched emergency contact information and facility information according to predetermined layout rules to generate the data structure of the planning document information. Based on this, the server calls the document generation module to convert the planning document information into an electronic document in a predetermined format, such as generating a printable document with titles, tables, and lists, or other machine-readable documents based on field content. During the document generation process, the server performs specific operations such as page layout calculations, paragraph pagination, font selection, and character encoding conversion to ensure good readability on different terminals and printing devices.

[0098] When synchronizing with the terminal, the server sends plan composition information and plan document information to the user terminal via the communication network. The terminal parses this information locally and displays it on the user interface in a partitioned manner. The terminal can store some or all of the plan composition information in its internal storage unit so that it can still display routes, times, equipment, and emergency contact information when there is no network connection. When saving data, the terminal uses a local data table structure to associate and store plan identifiers, timestamps, version numbers, and various information fields, enabling the terminal to manage multiple plans and switch versions.

[0099] During operation, the terminal can periodically acquire location information through its positioning module and report this information to the server or local records at predetermined time intervals. Upon receiving the location information, the server compares it with route nodes in the plan's constituent information, generating movement trajectory information and storing it in conjunction with the plan document information. The server can use the trajectory data to correct the planned time estimate or reference actual movement speed and stopping patterns in subsequent plan generation, thereby gradually improving the accuracy of climbing time prediction. This method of optimizing the plan through trajectory feedback differs from traditional static experience-based rules.

[0100] When processing multi-round interactions, the server integrates the new constraint information into the structured information each time the user adds conditions or requests changes, regenerates the prompt statements, and adjusts the constraints and explanations in the prompt statements based on the current structured information and contextual data such as historical trajectories and regional risk data. Compared to simply forwarding user text to the model, the server, through unified management of structured information, can maintain the consistency and integrity of input each time the generative AI model is invoked, thereby reducing the volatility of the model's output and achieving more stable generation behavior. This prompt statement generation and iterative update based on structured information represents a technical improvement to the computer's internal data management methods and model invocation process.

[0101] Throughout the system's operation, the server establishes an organic data processing chain encompassing "user input management—structured information maintenance—automatic generation of prompts—generative AI model inference—structured result parsing—regional information matching—plan document generation—terminal synchronization and offline storage—trajectory feedback and plan optimization." This streamlined process reduces repetitive parsing and data conversion operations between modules, thereby lowering CPU load and memory consumption and increasing processing speed. When the number of users increases or the request frequency rises, this structured data flow design effectively reduces the system's communication load because the server only transmits necessary prompt text when interacting with the generative AI model, while a large amount of intermediate results and structured information is stored in the local database and cache.

[0102] In another implementation, the server can employ different generative AI model structures, such as using a hybrid encoder-decoder structure, incorporating domain-specific feature embedding vectors, or adding feature dimensions representing mountaineering risk levels and weather patterns to the model's input. The server can fine-tune a general language model using mountaineering-related corpora through transfer learning methods. During fine-tuning, a hierarchical learning rate strategy and regularization methods are used to control parameter updates, balancing the model's general language capabilities with its domain-specific capabilities. The server can also add error cases and corrected plan samples to the model's training data, optimizing the distance between the "error plan" and the "corrected plan" by comparing the loss function. This makes it easier for the model to generate corrected results that meet safety requirements when faced with additional user conditions.

[0103] In some implementations, the server can incorporate a rule engine during the plan information parsing phase to perform technical verification on the plans output by the generative AI model. For example, the server can define a set of non-linear rules: when the estimated walking time for a given day exceeds a set threshold, when a change in altitude exceeds a set range, or when a critical item is missing from the necessary equipment list, the server automatically marks it as a risky plan and adds stricter safety constraints to the model through the regeneration of prompt statements. By working in conjunction with the model using rules, the server achieves secondary technical control over the generated results, reduces the burden of relying on manual review, and improves the reliability of safety-related outputs at the system level.

[0104] Through the aforementioned implementation, the server, terminal, and user each assume clearly defined roles within the system: the server handles intensive data processing and model inference control, the terminal handles user interaction and local information display, and the user is responsible for providing requirements and making decisions. By embedding the capabilities of generative artificial intelligence models into the improved computer system architecture through specific data structures and algorithmic processes, the server achieves comprehensive optimization of data management, computational efficiency, and communication load in the mountaineering project scenario. This technical solution not only automates the manual planning process but, more importantly, delivers quantifiable performance and accuracy improvements at the levels of internal computer structure, data flow organization, and model invocation mechanisms, representing an improvement in computer technology itself.

[0105] use Figure 11 The processing flow is explained.

[0106] Step 1: Users input mountaineering-related information using the terminal. Users operate the input controls on the terminal's graphical interface to input information such as the target mountain, the climber's skill level, climbing style, scheduled departure time, number of participants, and route preferences.

[0107] Input: The original input content such as text, options, and time entered by the user in the interface.

[0108] Based on user actions, the terminal maps each input item to internal fields (such as "mountain", "level", "style", "start_time", "members", "preferences", etc.) and constructs a unified data object in memory. The terminal performs local validation on some fields (such as whether they are empty or whether the time format is valid). After successful validation, the terminal serializes the data object into a request message and sends it to the server through the communication network.

[0109] Output: A request data packet containing structured fields, transmitted over the network to the server.

[0110] Step 2: The server parses the request and generates structured information. The server receives request data packets from the terminal, reads the payload, and calls the parsing module to parse the data.

[0111] Input: The request data packet sent by the terminal (containing serialized data of the user input fields).

[0112] The server uses a data parsing process to restore each field in the request body into its internal data structure, determines the field type, normalizes the target mountain name, maps climber skill level and climbing style to predefined enumerated values, and fills in default values ​​for missing optional fields. After completing this data processing, the server stores the results in a structured information storage unit, forming a structured information record.

[0113] Output: Data structures (structured information) stored inside the server, generating unique identifiers for subsequent processing.

[0114] Step 3: The server generates prompt statements based on structured information. The server reads the structured information record corresponding to the current request from the structured information storage unit and calls the prompt statement generation module.

[0115] Input: Structured information including mountain marker, climber level, climbing type, departure time, number of people, and preference criteria.

[0116] The server formats each field into a string based on a predefined language template, filling in information such as "mountain," "level," "type," "time," "number of people," and "preferences" into the corresponding positions in the template. It also adds requirements for the output structure based on the application scenario (e.g., requiring separate information for route, time, equipment, and safety precautions). During this process, the server concatenates the text, adjusts the syntax, and corrects punctuation, completing the conversion from structured data to natural language instructions to form prompts.

[0117] Output: Prompt text for input to generative artificial intelligence models.

[0118] Step 4: The server sends the prompt to the generative artificial intelligence model and retrieves the mountain climbing plan information. The server calls the generative artificial intelligence model service interface on an external information processing device and sends the generated prompt statement as an input parameter.

[0119] Input: Prompt text and model call parameters (including model name, maximum output length, temperature and other inference control parameters).

[0120] The server encapsulates the prompt statement as a request message and sends it to the model server via a network transmission interface. The generative AI model performs data calculations such as sequence encoding, attention calculation, and decoding on an external information processing device to generate natural language text of the climbing plan corresponding to the prompt statement. The server receives the response message returned by the model, extracts the climbing plan text from it, and caches it locally.

[0121] Output: A text representation of the mountaineering plan information in natural language.

[0122] Step 5: The server parses, differentiates, and extracts information from the mountain climbing plan. The server invokes the plan information parsing and reconstruction module to analyze the mountaineering plan information returned by the generative artificial intelligence model.

[0123] Input: The complete mountaineering plan in natural language text.

[0124] The server first segments the text based on pre-defined key phrases or paragraph markers (such as "Route:", "Time:", "Equipment:", "Precautions:", etc.). Then, it applies rule matching and content feature analysis to each segment, such as statistically analyzing keyword occurrences, sentence structure, time expressions, and item names, classifying each segment into categories such as hiking route information, time information, carried item information, and safety precautions. The server maps the segmentation results to structured fields; for example, it splits the route description into a list of nodes, converts the time description into numerical time intervals, and parses the equipment list into a set of entries.

[0125] Output: A structured data of planned activities categorized by type, including at least information on the hiking route, time, items carried, and safety precautions.

[0126] Step 6: The server will match the planned structure information with regional emergency information. The server accesses the regional information storage unit to retrieve emergency contact information and facility information related to the target mountain and its surrounding area.

[0127] Input: Mountain route information (including route nodes and geographic area identifiers) from the plan composition information, as well as emergency contact information and facility records from the regional information database.

[0128] The server uses data calculations such as geographic region matching and route node coordinate range comparison to associate route nodes with nearby rescue organizations, medical institutions, management stations, and other facilities, calculating auxiliary indicators such as distance or estimated arrival time. The server then appends the matched emergency contact information and facility information to the corresponding route node or the entire plan, forming an extended plan structure containing safety auxiliary information.

[0129] Output: Extended plan composition information including emergency contact information and facility information association results.

[0130] Step 7: The server generates planning document information and constructs electronic document content. The server uses the extended plan to edit the plan document information and combines the content of each category into an overall document structure.

[0131] Input: Extended plan composition information (route, time, equipment, safety precautions, and emergency contact and facility information).

[0132] The server assigns chapter titles, paragraph order, and display format to each type of information according to predetermined layout rules, and organizes text content, list content, and table content into a unified data structure. During this process, the server performs necessary formatting on the text, such as standardizing time expressions, adding serial numbers to equipment lists, and adding emphasis marks to key safety prompts. The server then generates a planning document from the formatted content for subsequent conversion into electronic document files or display on terminals.

[0133] Output: A data structure for planning document information used for rendering or exporting.

[0134] Step 8: The server sends plan composition information and plan document information to the terminal. The server calls the terminal synchronization control module, packages the plan composition information and plan document information into a response message, and sends it to the corresponding user terminal through the communication network.

[0135] Input: Plan composition information and plan document information, as well as identification data used to identify user sessions and requests.

[0136] The server serializes and compresses the output data to reduce network transmission load, and sets necessary version numbers or timestamps for local management by the terminal. The server sends the response to the terminal via network protocols and internally records the log information for this distribution.

[0137] Output: The response message transmitted to the terminal, containing plan information that can be parsed and displayed by the terminal.

[0138] Step 9: The terminal parses and displays the plan information and saves offline data. The terminal receives the response message from the server and calls the local parsing module to restore it to its internal data structure.

[0139] Input: Response data containing plan structure information and plan document information.

[0140] The terminal maps the plan's components to interface elements by category: the "Route" area displays the route text and key nodes; the "Time" area shows the departure time and estimated duration of each segment; the "Equipment" area lists the items to be carried; the "Precautions" area displays safety and weather tips; and the "Emergency Contact" area displays contact information for rescue organizations and facilities. The terminal writes the plan's components and emergency information to its internal storage unit according to a preset strategy for access in offline environments.

[0141] Output: A visual display of the mountaineering plan on the terminal screen, as well as offline plan data stored locally.

[0142] Step 10: Based on the displayed results, the user submits additional conditions or requests for changes. After reading the mountain climbing plan on the terminal, users can choose to adjust the conditions according to their personal needs and judgment.

[0143] Input: The plan information displayed on the terminal interface and the user's subjective preferences.

[0144] Users can modify departure time, preferred routes (e.g., routes to avoid crowds or steep inclines), trip intensity, or equipment requirements using interactive controls on the terminal, such as sliders, option buttons, and text input boxes. The terminal updates its local data structure based on user modifications, organizes the added condition information or change request information into a new set of fields, packages it into a request message, and sends it to the server.

[0145] Output: An update request data packet containing additional condition information or change request information.

[0146] Step 11: The server generates an update prompt statement based on additional conditions and retrieves updated mountain climbing plan information. After receiving the update request data packet, the server reads the original structured information and merges the appended condition information or change request information into the structure.

[0147] Input: Existing structured information and new additional conditional information or change request information.

[0148] The server re-executes the prompt generation process on the merged structured information, incorporating the already determined plan constraints and user-added preference descriptions into the new prompts, making them more specific. For example: "Please redesign the mountain climbing plan under the following constraints: the mountain is Mount Fuji, the climber's skill level is intermediate, the climbing method is day climb, the original departure time was 05:00, but we now request that the departure time be adjusted to 04:00, and that routes with fewer people and fewer steep slopes be prioritized. Please provide a revised recommended route, time arrangement, equipment suggestions, and weather and safety precautions, while ensuring safety." The server sends the update prompt statement again to the generative artificial intelligence model to obtain the updated mountain climbing plan text.

[0149] Output: New mountaineering plan information text generated based on the update prompt statement.

[0150] Step 12: The server re-parses and matches the updated mountain climbing plan and synchronizes it to the terminal. The server repeatedly performs the parsing, differentiation, extraction, regional information matching, and document generation processes on the new mountaineering plan information.

[0151] Input: The updated mountaineering plan in natural language and the original area information.

[0152] The server converts the new text into updated plan structure information using the same parsing rules, and re-matches emergency contact information and facility information based on the new route and schedule, generating an updated plan document. The server then sends the updated plan to the terminal in the same manner as the initial generation. Upon receiving the plan, the terminal refreshes its interface, replacing or displaying the old and new plans side-by-side, and updates its local storage data, allowing the user to prepare for the climb based on the latest plan.

[0153] Output: A response message containing updated plan structure information and plan document information, as well as the updated plan display and local storage content on the terminal side.

[0154] Application Example 1 The process flow corresponding to the specific processing in Use Case 1 will be described below. The various parts of the system described below are implemented by the data processing device 12 and the intelligent device 14. Furthermore, the data processing device 12 is referred to as the "server" and the intelligent device 14 is referred to as the "terminal".

[0155] Traditional computer-based information recommendation and plan generation technologies typically employ pre-defined rule engines or simple filtering algorithms to retrieve results from a database based on a limited set of user-input conditions. This type of technology suffers from the following problems: First, servers struggle to automatically construct complex filtering rules and evaluation benchmarks based on multi-dimensional and semantic user needs, resulting in limited understanding of users' natural language requirements and difficulty in accurately matching user intent with recommended results. Second, when filtering and ranking candidate objects, servers rely heavily on manually preset weights and logic, lacking a mechanism for dynamic generation and adjustment based on current needs, leading to non-scalable and costly comprehensive evaluation algorithms. Third, when users adjust conditions multiple times on the terminal, servers often only perform simple condition replacements and repeated searches, failing to utilize intelligent models to reconstruct selection rules and weighting strategies in each round of interaction, thus failing to fully utilize computing resources to improve recommendation quality and interaction efficiency. Fourth, servers typically separate recommendation results from subsequent processing such as behavior trajectories and report documents, failing to organically combine plan generation, behavior execution, path recording, and report document generation within the same technical framework, resulting in insufficient data utilization and difficulty in providing structured feedback data for subsequent system optimization.

[0156] Based on this, it is necessary to provide a system that automatically generates prompts and derives selection information using a generative artificial intelligence model, dynamically constructs evaluation benchmarks and selection rules on the server side, performs weighted evaluation and ranking of candidate information in the information storage device, and automatically generates report documents after multiple rounds of user interaction and behavior execution. This system aims to improve the server's ability to handle complex needs, the flexibility of data retrieval and evaluation algorithms, and the automation and intelligence of the overall information processing flow from a computer technology perspective.

[0157] The specific processing performed by the specific processing unit 290 of the data processing apparatus 12 in Application Example 1 is achieved by the following means.

[0158] In this invention, the server includes means for receiving user demand information, extracting conditions from the demand information, and generating a prompt statement to instruct a generative artificial intelligence model to generate plan information; means for inputting the prompt statement into the generative artificial intelligence model and obtaining selection information containing evaluation criteria and selection rules corresponding to the conditions based on the output of the generative artificial intelligence model; means for performing retrieval and evaluation processing on candidate facility information and candidate project information stored in an information storage device based on the selection information, thereby extracting candidate facilities and candidate projects that meet the conditions and generating plan information containing the candidate facilities and candidate projects; means for sending the generated plan information to a terminal device and updating the conditions based on a user change request obtained from the terminal device, repeating the generation of the prompt statement, the processing of the generative artificial intelligence model, the retrieval processing, and the evaluation processing according to the updated conditions to regenerate the plan information; and means for automatically generating document information for reporting based on the regenerated plan information and sending the document information to an external device. This allows for the automatic construction of evaluation benchmarks and selection rules tailored to current needs on the server side using generative artificial intelligence models and prompts. Combined with weighted retrieval and sorting of candidate information, it enables efficient computer processing of complex, multi-round changing user conditions. Furthermore, it integrates plan generation, result feedback, and document generation into a unified information processing flow, thereby improving the computer system's processing capacity, resource utilization efficiency, and automation level in the field of personalized plan generation.

[0159] A "system" refers to an integrated whole consisting of multiple computer resources, including processing devices, storage devices, and communication devices, used to perform comprehensive information processing tasks such as receiving user information, calling generative artificial intelligence models, retrieving candidate information, and generating plan information and document information.

[0160] A "server" refers to a computing device that executes core program logic in a system. It can receive requests from terminal devices, run applications through a processor, parse user request information, call generative artificial intelligence models, access information storage devices, and generate planning and document information.

[0161] "Terminal device" refers to an electronic device operated by a user and interacting with a server through a communication network, including but not limited to portable information processing devices and fixed information processing devices, used to input user demand information, receive and display plan information, and send change requests.

[0162] "User demand information" refers to a set of data related to the planned generation that is input by the user through a terminal device and sent to the server. It includes at least one or more fields that represent the user's expected conditions, such as category preferences, budget constraints, time conditions, location information, etc.

[0163] "Conditions" refer to parameters or restrictions extracted from user requirements information to constrain plan generation and candidate information filtering, including but not limited to numerical range, category specification, time limit, spatial location and other logical constraints.

[0164] "Generative artificial intelligence models" refer to machine learning models that can automatically generate text or structured information based on input prompts. They are usually trained using deep learning methods and are used to output evaluation benchmarks, selection rules, or other auxiliary information for planning generation.

[0165] "Prompt statements" refer to input content generated by the server based on user needs and conditions, expressed in natural language or structured text, used to instruct generative artificial intelligence models to generate corresponding outputs for specific tasks.

[0166] "Selection information" refers to the data set obtained by the server based on the output of the generative artificial intelligence model, used to filter and evaluate candidate information in the information storage device. It includes at least the evaluation criteria corresponding to the conditions, selection rules, and optional weight parameters.

[0167] "Evaluation benchmark" refers to a set of standards or indicators used to measure the merits of candidate facilities or projects, including but not limited to price, rating, distance, time, and matching degree, which are used to guide the evaluation process and weighted calculation.

[0168] "Selection rules" refer to a set of logic or conditions for filtering and sorting candidate information based on evaluation criteria. These rules are used to determine whether candidate facilities and projects are retained, how they are combined, and how they are prioritized.

[0169] "Information storage device" refers to storage resources used to store candidate facility information, candidate project information, and data related to the planning in a retrievable manner, including but not limited to database devices, file storage devices, or other non-volatile storage media.

[0170] "Candidate facility information" refers to facility-related data records stored in information storage devices that are the objects of the plan, including at least the facility's identification information, location attributes, category attributes, service attributes, and evaluation attributes.

[0171] "Candidate project information" refers to project or service data records that are associated with candidate facilities and stored in the same information storage device, including at least project identifier, content description, cost information, and restrictions.

[0172] "Retrieval processing" refers to the process by which a server queries and retrieves candidate facility information and candidate project information that meet the specified conditions from an information storage device based on the selected information. This is usually achieved through query statements or index searches.

[0173] "Evaluation processing" refers to the process by which the server calculates and scores the retrieved candidate facility and candidate project information based on evaluation criteria and selection rules, thereby obtaining evaluation results for sorting or screening.

[0174] "Planning information" refers to a structured data set generated by the server after performing retrieval and evaluation processing, which is used to provide users with one or more candidate combinations. It includes at least the selected candidate facilities, candidate items and their related attributes and evaluation results.

[0175] "User change request" refers to a request sent to the server by a user after viewing the plan information, through a terminal device, to modify, add, or delete existing conditions. This request is used to trigger the server to regenerate prompt statements, repeatedly call the generative artificial intelligence model, and regenerate plan information.

[0176] "Regenerated plan information" refers to the updated plan information obtained by the server after receiving a user's change request, based on the updated conditions, regenerated prompts, and the output of the generative artificial intelligence model, by performing new retrieval and evaluation processing on the candidate information.

[0177] "Document information used for reporting" refers to document data automatically generated by the server based on generated or regenerated plan information, used to record, submit or share plan content and related execution status, including but not limited to text format, structured data format or visual report format.

[0178] "External device" refers to a device that exchanges data with the system through a communication network and is used to receive document information for reporting, including but not limited to management terminal, archive server or other information processing device.

[0179] "Attribute value" refers to a specific numerical or category value used to describe a certain attribute in candidate facility information and candidate project information, such as price value, rating value, distance value, time length, type label, etc.

[0180] "Weighted operation" refers to the calculation process by which the server, based on the weights specified in the selection information, mathematically combines multiple attribute values ​​in a predetermined proportion to calculate the comprehensive evaluation value of each candidate combination.

[0181] The "comprehensive evaluation value" refers to a single numerical value or indicator obtained by the server after weighting multiple attribute values ​​for each combination of candidate facilities and candidate projects, used for overall evaluation and ranking.

[0182] A “candidate combination” refers to a combination unit consisting of a candidate facility and one or more related candidate projects, which serves as a basic alternative unit to be selected in the planning information.

[0183] "Location information" refers to data related to the location of a user or facility, including but not limited to geographic coordinates, area identifiers, address information, etc., which are used for spatial condition filtering and route calculation.

[0184] "Spatial conditions" refer to geographical range or distance restrictions set based on location information for filtering candidate facility information, such as within a predetermined radius or within a specified area.

[0185] "Movement path information" refers to structured data generated based on the user's historical location information during the execution of the plan, used to represent the user's movement trajectory, including path records consisting of multiple location information points and their time sequence.

[0186] In this embodiment of the invention, the server, terminal, and user each assume different functional roles, collaboratively achieving the generation and processing of planned information and documents based on a generative artificial intelligence model and prompt statements. The system of this invention can be deployed on general-purpose computer infrastructure, such as virtual machines, function execution environments, relational database services, and storage services provided by cloud computing platforms. The following provides a detailed description of the hardware configuration, software configuration, data structure, generative artificial intelligence model structure and learning methods, as well as the data flow and technical effects between modules.

[0187] I. Example of Hardware and Software Configuration A server may include one or more physical computing nodes, each of which includes at least a central processing unit (CPU), random access memory (RAM), non-volatile storage media (such as solid-state drives or disk arrays), and a network interface controller. The server can communicate with terminals and external devices via a data center network.

[0188] Servers can run on cloud computing platforms. Servers can use function execution environments as application runtime containers, such as event-triggered cloud function services; servers can also use virtual machines or container platforms to run backend applications, such as a combination of a Linux operating system and container management software. Servers can use relational database management systems, such as SQL-based database systems, to achieve persistent storage and retrieval of candidate facility and candidate project information.

[0189] The server can run backend applications, which can be implemented using general-purpose programming languages ​​and corresponding web frameworks. For example, an interpreted language with a lightweight web framework, or a compiled language with an enterprise-level web framework. The server can provide a network interface via HTTP / HTTPS protocols to receive user request information and change requests from terminals, and return planning information and reporting documents to the terminals.

[0190] The server can invoke generative AI models through a dedicated inference service or model server. Generative AI models can be deployed on inference servers equipped with graphics processing units (GPUs) or tensor-accelerated hardware. The server can communicate with the model server via REST-style interfaces, RPC interfaces, or message queues.

[0191] The terminal can be a smartphone, tablet, laptop, or desktop computer. It may include a display unit, a touch input unit or keyboard input unit, a position sensor (e.g., using a GPS module or base station positioning module), and a wireless communication module (e.g., a cellular communication module or a wireless LAN module). The terminal may run a mobile operating system or a desktop operating system.

[0192] The terminal can execute the client application of this invention, which can be a native application, a hybrid application, or a web application. The terminal can collect user request information through a graphical user interface, display the plan information and document summary returned by the server, and send change requests.

[0193] Users input their requirements for plan generation through the terminal interface, such as category preferences, budget constraints, time constraints, and location range. Users can view multiple candidate combinations and select or modify them on the interface.

[0194] II. Data Structure and Storage Design The server can use key-value mappings or object structures to represent user demand information in memory. User demand information can include a set of fields such as: category fields (e.g., "cuisine type" or "service type"), numeric fields (e.g., budget amount or time limit), boolean fields (e.g., whether delivery is required or whether certain restrictions are enabled), location fields (e.g., latitude and longitude coordinates or region identifier), and optional preference fields (e.g., taste preference or style preference).

[0195] The server can define multiple data tables or sets in the information storage device to store candidate facility information and candidate project information. Candidate facility information may include facility identifier fields, geographic location fields (longitude and latitude), category fields, rating fields, service time fields, contact fields, and tag fields. Candidate project information may include project identifier fields, affiliated facility identifier fields, price fields, attribute tag fields (e.g., whether it's spicy, whether it contains specific ingredients), and availability time fields.

[0196] The server can index the location and category fields in the database to improve retrieval efficiency based on spatial and category criteria. The server can build intermediate data structures in memory, such as a list of candidate combinations, where each element includes a combination of facility records and one or more project records, along with multiple attribute values ​​associated with that combination (such as overall price, estimated time, distance, rating, etc.).

[0197] The server can use a dedicated data structure to represent selection information, which may include: a set of evaluation benchmarks (each benchmark corresponds to an attribute name and a minimum / maximum limit), a set of selection rules (each rule corresponds to a logical condition expression), and a weight vector (used for multi-attribute weighted calculation).

[0198] III. Structure and Learning Methods of Generative Artificial Intelligence Models The generative artificial intelligence model used by the server can be a sequence-to-sequence neural network based on a self-attention structure. This model can adopt a multi-layer encoder-decoder architecture, where both the encoder and decoder contain multi-layer self-attention modules, feedforward neural network modules, and residual connection and layer normalization modules.

[0199] During the model training phase, the server can prepare a training dataset. Each sample in this dataset includes an input part and an output part. The input part can be a simulated prompt statement describing the user's needs, such as category, budget, time, location, etc.; the output part can be the corresponding rule description text or structured rule representation, such as: "Price must not exceed X", "Distance must not exceed Y kilometers", "Rating must not be lower than Z", "Weights are set to w1, w2, w3", etc.

[0200] During training, the server can use the cross-entropy loss function as the error function. The server can update the model parameters using stochastic gradient descent or adaptive learning rate optimization algorithms. During training, the server can perform data augmentation, such as by replacing synonyms or changing conditional order, to increase the diversity of input samples, enabling the model to better generalize to various natural language expressions.

[0201] During the model deployment phase, the server can set the maximum sequence length, output temperature parameters, and sampling strategy (such as greedy decoding or bundle search decoding) for the generative AI model. When calling the model, the server can convert the prompt statement into a sequence of sub-words or word fragments, map them into vectors through the embedding layer, and then input them into the model's encoder and decoder for forward propagation to obtain the output labeled sequence, which is then assembled into text or structured data.

[0202] The server can constrain the model's output format to make the model generate rule descriptions that are easy for the program to parse. For example, the server can explicitly require the model to output according to the structure of "hard condition list", "soft condition list", and "weight list" in the prompt statement, making subsequent text parsing or JSON parsing more stable and efficient.

[0203] IV. Generation and Examples of Prompt Statements After receiving user request information from the terminal, the server extracts condition fields and populates them into a predefined prompt statement template. The server can pre-store multiple templates for different types of task generation. The following is an example of a prompt statement that can be used in a practical implementation.

[0204] For example, in one application scenario, the server generates the following prompt: "The user prefers Italian cuisine, has a budget of 2000 yen, and requires delivery within 30 minutes. Based on these conditions, please explain how to filter the most suitable options from restaurant and menu data, and provide the filtering rules." Alternatively, the server might generate the following prompt in another scenario: "The user's requirements are as follows: Cuisine: Chinese; Budget limit: 1500 yen; Delivery time requirement: within 30 minutes; Location: the user's current location. Based on this information, please generate a restaurant and menu recommendation strategy, including the conditions that must be met and the sorting rules." For example, the server might generate the following message when dietary restrictions are in place: "The user prefers Chinese cuisine, has a budget of 1500 yen, requires delivery within 30 minutes, and does not eat spicy food or nuts. Based on these conditions, please provide specific rules for database filtering and explain how to remove dishes that do not meet the dietary restrictions." The server uses the aforementioned prompts to unify discrete numerical and semantic conditions into a coherent natural language description, which is then handed over to a generative artificial intelligence model for semantic parsing and rule derivation.

[0205] V. Server-side rule generation and weighted evaluation algorithm After receiving the output text from the generative AI model, the server uses a text parsing module to perform structured processing on the output content. The server can extract hard conditions, soft conditions, and weight parameters through keyword matching, regular expressions, tag segmentation, or by having the model output directly in a labeled format (e.g., using prefixes like "condition:" or "weight:").

[0206] After extracting these conditions, the server maps them into an internal selection information data structure. Hard conditions can be converted into database query constraints, such as price caps, time caps, rating caps, and distance caps; soft conditions can be converted into weights or penalty terms in a multi-attribute scoring function. For example, the server can construct the following comprehensive scoring function based on the weight vectors w_price, w_distance, and w_rating from the selection information: Overall rating = w_price × price score + w_distance × distance score + w_rating × rating score + … The server can calculate a price score (e.g., a normalized score based on the difference between price and budget), a distance score (e.g., mapped to a 0–1 range based on geographical distance), and a rating score (e.g., directly normalizing existing ratings) for each candidate combination. The server can then combine these multiple attribute scores using the aforementioned weighting function to obtain a comprehensive evaluation value for each candidate combination.

[0207] In this way, the server transforms the high-level rules output by the generative artificial intelligence model into precise numerical calculation logic. Compared with traditional systems that rely entirely on manual setting of weights and rules, it can dynamically generate differentiated evaluation functions for different user needs, thereby improving personalization and matching accuracy.

[0208] VI. Interaction with terminals and users and the technical utilization of path information When users view the plan information returned by the server on their terminals, they can make change requests based on the displayed content, such as adjusting the budget, changing categories, adding dietary restrictions, or switching time conditions. After performing basic validation of the user input locally, the terminal sends the new user request information to the server.

[0209] Upon receiving a change request, the server does not simply replace the conditions and repeat the same search; instead, it regenerates new suggestions and invokes the generative AI model again to generate new evaluation criteria and selection rules. This method enables adaptive updates to the evaluation logic across multiple rounds of interaction, avoiding the inefficiency and reduced recommendation quality caused by fixed rules in traditional systems.

[0210] In some implementations, the server can periodically obtain the user's location information from the terminal. Based on this location information, the server can, on the one hand, perform spatial filtering of candidate facilities during the initial plan generation phase, retaining only facilities within a set radius or specified area. This significantly reduces the size of the candidate set during the database retrieval phase, lowers I / O and computational burden, and improves processing speed. On the other hand, after the user executes the plan, the server can generate historical movement path information based on the location information and associate this path information with report documentation information for subsequent analysis and optimization algorithms.

[0211] VII. Report Document Generation and Technical Effects After generating or regenerating planning information, the server can automatically construct document information for reporting based on this planning information. The server can use template filling to embed selected facilities, projects, schedules, budget information, and movement route information into a predefined document structure. The server can generate structured reports, such as document descriptions containing chapter titles, data tables, and charts, and send them to external devices in a standard format (such as text or markup language documents).

[0212] By automatically generating report documents, the server can tightly integrate the planned generation results with subsequent recording, monitoring, or statistical processing, forming a complete data loop. Over long-term operation, the server can accumulate a large amount of data on planning conditions, generation rules, selected results, and actual execution paths. This data can be used to further train or fine-tune generative artificial intelligence models, enabling them to have higher accuracy and stability in subsequent rule generation tasks.

[0213] Because the server in this invention utilizes a generative artificial intelligence model to generate complex evaluation rules adapted to current needs and transforms them into explicit attribute weighting operations and database retrieval constraints, this invention can bring technical benefits in the following aspects: By using a semantic model in the rule generation phase and numerical computation and index retrieval in the execution phase, the server optimizes the retrieval scope and scoring calculation according to specific needs, reduces unnecessary candidate processing, and thus improves processing speed. The server improves the accuracy and stability of candidate combination ranking by integrating multidimensional attributes into a weighted evaluation function and dynamically adjusting the weights, thereby reducing reliance on manual rule maintenance. By filtering data based on location information and recording movement paths, the server controls the data access range to areas related to the user's actual accessibility, thereby reducing communication load and database load, and using path information to optimize subsequent models or strategies. By using a unified data structure and modular processing, the server integrates prompt generation, model reasoning, rule parsing, database operations, and report generation into a scalable data flow pipeline, thereby optimizing the use of computing resources and simplifying system expansion and maintenance.

[0214] Unlike simple "automation of human tasks," the server in this invention not only replaces manual screening but also combines the rule generation capabilities of generative artificial intelligence models with server-side numerical calculation algorithms to achieve a level of dynamism and complexity that traditional rule engines struggle to reach. The server performs intelligent search and optimization on multi-dimensional data spaces using non-human-preset weight and condition combinations. This approach, combining deterministic database operations with probabilistic model reasoning, substantially improves the computer system's ability to handle complex, multi-condition queries, its scalability, and its automatic self-adaptation.

[0215] use Figure 12 The processing flow is explained.

[0216] Step 1: Users input their requirements into the terminal interface.

[0217] Users select or fill in plan-related conditions on the terminal as input, including category (e.g., cuisine type or service type), budget limit, time limit (e.g., expected completion time or delivery time), whether specific restrictions are required (e.g., not spicy, not containing a certain ingredient), and location information (automatically obtained using the terminal's location function or manually selected area).

[0218] The terminal locally structures the above input data into a set of key-value pairs, for example, mapping categories to string fields, budgets to numeric fields, time limits to integer minutes, and location information to latitude and longitude coordinates.

[0219] The terminal performs format validation and validity checks on the input (e.g., whether the budget is positive and whether the time is within the selectable range). After the validation is passed, the structured data is prepared as output and sent to the server.

[0220] Step 2: The terminal sends user request information to the server.

[0221] The terminal takes the structured requirement data from step 1 as input, serializes the data into a text format (e.g., a JSON string) through the communication module, and encapsulates it in a network request.

[0222] The terminal sends the request to the server's specified address via a secure transmission protocol, and simultaneously starts a task locally to wait for a response.

[0223] The terminal outputs a successfully sent network request containing user request information in this step, and waits for the server to return the processing result.

[0224] Step 3: The server receives and parses the user's request information.

[0225] The server takes network requests sent by the terminal as input, receives the request content through the network interface, and reads the text data within it.

[0226] The server uses a parsing library to deserialize text-formatted data into internal data structures (such as dictionaries or objects), from which it extracts category fields, budget fields, time limit fields, location information fields, and other preference fields.

[0227] The server performs data type checks and default value completion on these fields (e.g., setting a default upper limit when there is no time limit), and outputs the extracted conditional data object for use in generating subsequent prompt statements.

[0228] Step 4: The server generates a prompt message.

[0229] The server takes the conditional data object output in step 3 as input, selects a prompt statement template suitable for the current task type, and fills the placeholders in the template with each conditional field.

[0230] The server performs string concatenation and formatting operations, converting numeric fields (such as budgets and times) into text with units, and location information into readable descriptions or simply as "user's current location," combining them into a complete natural language description.

[0231] The server's output at this step is a prompt statement for generative artificial intelligence model processing, for example: "The user prefers Italian cuisine, has a budget of 2000 yen, and requires delivery within 30 minutes. Based on these conditions, please explain how to filter the most suitable options from restaurant and menu data, and provide the filtering rules." Step 5: The server invokes a generative artificial intelligence model to generate selection information.

[0232] The server takes the prompt statement output in step 4 as input, sends the prompt statement to the generative artificial intelligence model through the model interface, and specifies the decoding method (such as beam search decoding, maximum output length, and other parameters).

[0233] Generative AI models internally segment and embed prompts, perform forward propagation through multi-layer self-attention networks and feedforward networks, and generate an output text based on the trained parameters. This text contains rule information such as hard conditions, soft conditions, and attribute weights.

[0234] After receiving the model's output text, the server uses text parsing logic (keyword location, structural label parsing, etc.) to break down the rules into structured data such as an evaluation benchmark list, a selection rule list, and a weight list.

[0235] The server output in this step is a selection information data structure, which includes a set of conditions and weight parameters that can be directly used for database queries and scoring calculations.

[0236] Step 6: The server constructs database query conditions based on the selected information.

[0237] The server takes the selection information obtained in step 5 and the original conditions in step 3 as inputs, and combines them to form the constraint expression required for the database query.

[0238] The server converts hard conditions (such as price not exceeding budget, time less than limit, rating not lower than a certain value, distance not exceeding a certain radius) into conditional clauses in the database query language, mapping category and tag conditions to the corresponding fields.

[0239] In this step, the server outputs one or more structured query expressions to access candidate facility information and candidate project information in the information storage device.

[0240] Step 7: The server retrieves candidate facilities and candidate projects from the information storage device.

[0241] The server takes the query expression generated in step 6 as input, establishes a connection with the information storage device through the database driver, and performs retrieval operations on the stored candidate facility table and candidate project table.

[0242] The server uses an index structure to accelerate condition matching, reading a set of records from the database that meet spatial, category, price, and rating conditions.

[0243] The server converts the query results into a list or set structure in memory. Each facility record and project record stores its attribute values ​​and associates project records with facility records into preliminary candidate combinations based on the facility identifier.

[0244] The server output at this step is a set of candidate combinations, each of which includes at least one facility record and one or more project records.

[0245] Step 8: The server performs weighted evaluation and ranking of candidate combinations.

[0246] The server takes the candidate combination set obtained in step 7, as well as the weight parameters and evaluation benchmarks in step 5, as input, and constructs multiple attribute value vectors (such as price attribute, distance attribute, rating attribute, time attribute, etc.) for each candidate combination.

[0247] The server normalizes each attribute based on the evaluation criteria (e.g., mapping price differences to scores between 0 and 1, scaling distances to the maximum distance, and normalizing scores to the maximum score), resulting in a set of attribute scores.

[0248] The server performs a weighted calculation based on the weights given in the selection information: for each candidate combination, the corresponding attribute scores are multiplied and added according to their weights to obtain a comprehensive evaluation value.

[0249] The server then sorts all candidate combinations by their overall evaluation value, which can be sorted in descending order to determine the priority recommendation order, and combinations below a certain threshold can be discarded.

[0250] The server output in this step is a sorted list of candidate combinations, with each combination accompanied by a comprehensive evaluation value and scores for each component attribute.

[0251] Step 9: The server generates the plan information and sends it to the terminal.

[0252] The server takes the candidate combination list from step 8 as input, selects several combinations with higher comprehensive evaluation values, extracts the required display fields (such as facility name, location, project name, price, estimated time, reasons for recommendation, etc.), and constructs the plan information structure.

[0253] The server translates complex internal attributes and rating results into user-understandable text descriptions (such as "nearby and within budget" or "high rating and short delivery time") as part of the plan information.

[0254] The server serializes the plan information into a text format and sends it to the terminal via a network response.

[0255] The server output at this step is a response containing planning information that includes multiple recommended combinations.

[0256] Step 10: The terminal receives and displays the plan information.

[0257] The terminal takes the plan information response returned by the server as input, deserializes the text-formatted plan information into internal data objects through a parsing function, and maps the data of each candidate combination to the interface components.

[0258] The terminal sets the interface layout according to the attributes of facilities and projects, displays the facility name, distance estimate, project name, price and estimated time, and displays the recommendation reasons when necessary.

[0259] The terminal outputs a graphical list of recommendations at this step, which users can browse and select from.

[0260] Step 11: Users input change requests based on plan information.

[0261] Users use the plan information displayed on the terminal as input reference to judge whether the recommended results meet their needs. If they are not satisfied, they can modify the conditions on the interface, such as reducing the budget, shortening the time limit, changing the category, or adding restrictions such as "no spicy food, no nuts".

[0262] The terminal restructures the user's modified conditions into new user requirement information and triggers a new request sending process.

[0263] The terminal's output in this step is a new round of user demand information containing updated conditions.

[0264] Step 12: The server generates plan information again through multiple rounds of interaction.

[0265] The server takes the new user demand information from step 11 as input and repeatedly performs the data parsing, prompt statement generation, generative artificial intelligence model invocation, selection information parsing, database retrieval and weighted evaluation processes described in steps 3 to 8.

[0266] The server generates a new prompt statement based on the updated conditions, for example: "The user prefers Chinese cuisine, has a budget of 1500 yen, requires delivery within 30 minutes, and does not eat spicy food or nuts. Based on these conditions, please provide specific rules for database filtering and explain how to remove dishes that do not meet the dietary restrictions." The server obtains an updated candidate combination ranking through a new round of model output and weighted calculation, generates new plan information as output, and sends it to the terminal again to achieve multi-round iterative optimization.

[0267] Step 13: The server uses location information to generate movement path information and associates it with documents.

[0268] During the user's execution plan, the server takes the periodically reported location information as input, writes a series of timestamped location information samples into the memory buffer, and then persists them in batches to the information storage device.

[0269] The server sorts these location information in chronological order and generates movement path information, such as path polylines and road segment statistics, based on the distance and time interval between adjacent points.

[0270] The server fills the movement path information and the final adopted plan information into the report document template, automatically generating a document containing the execution trajectory as a technical record of the plan execution. This document can be sent to external devices for storage or for subsequent analysis.

[0271] Alternatively, an emotion engine for inferring user emotions can be combined. That is, the specific processing unit 290 can also use the emotion-specific model 59 to infer user emotions and perform specific processing using user emotions.

[0272] Example 2 The flow of a specific process in Example 2 will be described. Each part of the system described below is implemented by the data processing device 12 and the smart device 14. The data processing device 12 will be referred to as the "server," and the smart device 14 as the "terminal."

[0273] In the field of mountaineering and outdoor activity management, existing computer-based implementations suffer from the following technical shortcomings. First, the processing of user input, location data, and external organization document formats between the terminal and the server is fragmented: user-inputted activity plans are often stored as simple text, and the server lacks standardized data structures and a unified process for automatically generating documents. This results in significant manual intervention in the generation and submission of mountaineering plans to external organizations, hindering the full utilization of computing resources for automation. Second, while existing systems can acquire terminal location information, this is typically used only for simple real-time display or single-point positioning. They lack a systematic mechanism for structured storage, spatial retrieval, and trajectory reconstruction of location data, failing to provide efficient data support for emergency contact point searches, nearest refuge location alerts, and activity record statistics. Third, traditional plan generation processes often rely on fixed templates or simple rule engines, unable to dynamically adjust the generation logic based on the type and incompleteness of user input. Furthermore, they cannot automatically generate high-quality plan content conforming to predetermined document formats using generative artificial intelligence models, thus limiting the server's processing capabilities in natural language processing and document automation.

[0274] Furthermore, in emergency situations, existing technologies often rely on servers to simply forward location information or manual determination of the nearest refuge location based on location. There is a lack of automated algorithms for determining the nearest refuge location based on location databases and spatial computation. This results in low efficiency in utilizing computing resources and data structures when processing emergency requests, leading to unstable response processes and low levels of automation. Moreover, existing activity records are mostly stored as simple "start-end" time records or a few key points, failing to fully mine historical location data and generate structured activity record data comparing "planned and actual" data through path reconstruction and statistical calculations. This limits the server's capabilities in data analysis and subsequent services (such as risk assessment and route optimization).

[0275] In summary, the technical challenge this invention aims to address is to provide a unified data structure and processing flow on the server side, enabling the server to: automatically generate high-quality prompts based on user input and call generative artificial intelligence models to generate activity plans and document data; uniformly store location information uploaded by terminals and basic plan data, and perform spatial retrieval and trajectory reconstruction; automatically determine the nearest refuge location from the refuge location information database when an emergency notification is detected; and generate activity record data containing comparison information between the plan and the actual situation, thereby comprehensively improving the processing efficiency and automation level of the computer system in plan generation, document processing, location data analysis, and emergency response.

[0276] The specific processing performed by the specific processing unit 290 of the data processing apparatus 12 in Embodiment 2 is achieved by the following means.

[0277] In this invention, the server includes: a unit for receiving input information from an information processing device via a computing device and storing the input information as basic data in an information storage device; a unit for generating a prompt statement containing activity plan information based on the basic data via a computing device and inputting the prompt statement into a generative artificial intelligence model to obtain generated result data including activity plan content; a unit for performing format conversion processing on the generated result data via a computing device to generate document data for submission to an external organization and automatically sending the document data to the external organization via a communication device; and a unit for receiving location information data from a mobile terminal via a communication device and associating the location information data with the basic data via a computing device and storing it in the information storage device; and a unit for using... The system comprises: a unit for retrieving emergency contact information and refuge location information from a geographic information database using a computing device based on the location information data, and providing the emergency contact information and refuge location information to the mobile terminal via a communication device; a unit for periodically receiving historical location data from the mobile terminal via a communication device, performing route reconstruction and statistical processing on the historical location data using a computing device to generate activity record data, and providing the activity record data to the mobile terminal via a communication device; and a unit for determining the nearest refuge location from a refuge location information database using a computing device based on the latest location information data when an emergency notification is detected from the mobile terminal, and prompting the mobile terminal with the nearest refuge location via a communication device. This allows for the formation of a unified data structure and processing pipeline within the server, centered on basic data. This enables generative AI model invocation, document format conversion, location data storage and spatial retrieval, trajectory reconstruction and statistical analysis, and emergency shelter location determination to be executed collaboratively within the same computing framework. Consequently, it reduces human intervention, increases the automation of the plan generation and submission process, improves the computational efficiency and reliability of location data processing and emergency response, and achieves overall performance optimization of computer technology in automatic activity plan generation, geographic information processing, and emergency services.

[0278] A "system" refers to a collection of computer devices that consist of computing devices, information storage devices, communication devices, and multiple functional units, used to perform a series of information processing operations such as activity plan generation, document processing, location information processing, and emergency response.

[0279] "Computational device" refers to a processor or collection of processors that executes program instructions to perform logical operations, numerical operations, and control processing on input data, and is used to realize functions such as prompt statement generation, model calling, format conversion, spatial retrieval, and statistical calculation.

[0280] "Information processing device" refers to a computer device that provides user input information or business-related information to a server, including but not limited to user terminals, back-end management terminals, or other computing devices that can interact with the server.

[0281] "Information storage device" refers to a storage medium used to electronically store basic data, location information data, location history data, activity plan data, document data, and various configuration information, including database systems, hard disks, solid-state storage, and combinations thereof.

[0282] "Communication device" refers to a communication interface or communication module used for sending and receiving data between a server and a mobile terminal, external organization, and other devices, including wired network interfaces, wireless communication modules, and the communication protocol stack running on them.

[0283] "Input information" refers to various types of data sent from information processing devices or mobile terminals to the server for generating activity plans, document data, and performing location services. It includes at least the purpose of the activity, time, route, participant information, emergency contact information, and environmental information related to the activity.

[0284] "Basic data" refers to the raw or structured data set that the server extracts from the input information and stores in the information storage device, which is used for subsequent operations such as generating prompt statements, calling generative artificial intelligence models, performing location association processing, and statistical analysis.

[0285] "Prompt statement" refers to a natural language string or its equivalent representation generated by the computing device based on basic data and provided as input to the generative artificial intelligence model, used to instruct the generative artificial intelligence model to output generated result data containing predetermined activity plan content and document structure.

[0286] "Generative artificial intelligence model" refers to a text generation model based on machine learning methods, especially deep learning networks, that can automatically generate natural language output results containing activity plans, explanatory texts, or document content based on prompts.

[0287] "Generated result data" refers to the natural language text data or its structured representation, which includes the content of the activity plan and the document, output by the generative artificial intelligence model after receiving the prompt statement, and is used for further format conversion and submission processing.

[0288] "Activity plan information" refers to the planned content related to the implementation of a specific activity, such as objectives, schedule, route, participant configuration, and safety precautions.

[0289] "Document data" refers to electronic document data that is converted into a format by a computing device based on the generated result data and is used to submit to external organizations, including file data that meets predetermined format, field structure and submission requirements.

[0290] "External entities" refers to data recipients who receive and manage activity planning documents, including management agencies, public institutions, or their information processing systems responsible for security management, administration, or emergency response.

[0291] "Mobile terminal" refers to a portable electronic device carried by a user and capable of exchanging data with a server through a communication device, including at least smart terminal devices with positioning and communication functions.

[0292] "Location information data" refers to geospatial data obtained by the mobile terminal and sent to the server to represent the current location of the mobile terminal, including at least longitude, latitude, and optional parameters such as altitude and timestamp.

[0293] "Location history data" refers to a collection of multiple location information data that the server periodically receives and stores from the mobile terminal, arranged in chronological order, used to represent the movement trajectory throughout the entire activity.

[0294] A “geographic information database” refers to a collection of data that stores geospatial data and related attribute information in a searchable format, and is used to query emergency contact information, shelter information and other geographic-related information based on location information data.

[0295] "Emergency contact information" refers to the collection of information about contact targets related to safety or rescue retrieved by the server from a geographic information database, including organization name, contact information, address, and functional category.

[0296] "Refuge location information" refers to the location information and attribute information of a place that can be used as an emergency refuge or temporary shelter, which is retrieved by the server from a geographic information database or a refuge location information database. This includes geographic coordinates, name, and basic description.

[0297] "Route reconstruction processing" refers to the process by which a computing device connects and generates trajectories from multiple time-ordered location information data based on historical location data in order to restore or approximately reconstruct the actual movement route.

[0298] "Statistical processing" refers to the numerical analysis operations performed by the computing device on historical location data and related activity data, including but not limited to calculation of movement distance, movement time, altitude change, and the acquisition of other statistical indicators.

[0299] "Activity record data" refers to a data set generated by the server after route reconstruction and statistical processing of historical location data, used to represent the actual execution of an activity. It includes at least trajectory information, statistical results, and optional visualization information.

[0300] "Emergency notification information" refers to control information with an emergency status indication and its corresponding location information sent by a mobile terminal to a server when it detects that a user has triggered an emergency operation or that a predetermined emergency condition has been met.

[0301] The "refuge location information database" refers to a data set used to store information on multiple refuge locations and to support retrieval and distance calculation based on location information data. It can be implemented independently of a geographic information database or as part of it.

[0302] "Nearest refuge location" refers to one or more refuge locations that are closest to the latest location information under a predetermined distance metric, obtained by the server through spatial calculations based on the latest location information database after receiving emergency notification information.

[0303] The embodiments of this invention will be described in conjunction with the collaborative work of the server, terminal and user, and will focus on the data structure design, algorithm flow, composition and learning method of generative artificial intelligence model inside the server, as well as the resulting improvement in computer technology.

[0304] I. System Overall Structure and Hardware / Software Environment The server is implemented using a general-purpose computing platform. It employs a computing device with a multi-core central processing unit and main memory, such as a server running the Linux operating system. The server is equipped with an operating system (e.g., a Linux-based server operating system), a database management system (e.g., a relational database management system), a web application framework (e.g., a server-side application framework), geographic information extension components (e.g., a spatial database extension module), and generative artificial intelligence model inference services (e.g., a large-scale language model inference process based on the Transformer architecture).

[0305] The terminal is implemented using a mobile information terminal with satellite positioning components and a wireless communication module. For example, the terminal can be a smart terminal running a mobile operating system, with a built-in global satellite positioning chip, communication module, and graphics display device. The terminal has a client application installed to execute the program of this invention, calling the location service interface, network communication interface, and graphical user interface components.

[0306] Users interact with the server through a terminal, inputting activity-related information, viewing emergency contact information, evacuation location information, and activity record data returned by the server.

[0307] II. Server-side module composition and data structure The server comprises multiple functional modules, which can be physically implemented by different program logics executed by the same processor, and are logically divided as follows.

[0308] 1. Input Information Management Module The server uses an input information management module to receive input information sent by terminals or other information processing devices. The server parses the input information into an internal data structure, such as an object-oriented structure or record structure, organizing fields like "destination," "time range," "route description," "participant list," and "emergency contact" into key-value pairs. The server then stores this structured data as the base data in a "basic data table" or equivalent data set within the information storage device.

[0309] The server uses a relational database to organize data in multiple tables, such as: user table, activity plan table, location record table, emergency contact point table, shelter location table, and activity record table. The activity plan table includes fields for user ID, activity ID, destination ID, start and end time, route text, participant information, emergency contact, and status. The location record table includes fields for activity ID, timestamp, longitude, latitude, altitude, and emergency marker, and uses spatial indexes to accelerate retrieval.

[0310] 2. Prompt Statement Generation Module The server uses a prompt generation module to perform logical processing on the computing device, automatically generating prompts based on the underlying data to invoke the generative artificial intelligence model. The server first checks for missing fields in the underlying data, such as whether route information or emergency contact information is missing. The server then uses a predefined rule matrix to map "field type × missing status" to a "prompt statement structure template." This template includes not only fixed text but also defines the paragraph order, required items, and optional items.

[0311] The server uses string concatenation, placeholder replacement, and conditional insertion algorithms to populate the selected prompt template with basic data. For example, the server generates the following text prompt based on the event destination, date, route, and participant information: You are a mountaineering safety management expert. Based on the following mountaineering plan information, please generate a formal mountaineering plan in Chinese, including the following format: 1. Basic Information (Destination, Date, Route) 2. Participant Information (Name and Phone Number) 3. Emergency Contact 4. Itinerary 5. Safety Precautions Climbing destination: Mt. Fuji Planned dates: August 10, 2026 to August 11, 2026 Planned route: Yoshidaguchi hiking trail Participants: Zhang San, Phone: 090-XXXX-XXXX Li Si, phone number: 080-XXXX-XXXX Emergency contact person: Wang Wu, phone number: 03-XXXX-XXXX Note: Climb to the summit at night to watch the sunrise. Please output the complete body of the plan.

[0312] When generating prompts, the server does not simply concatenate text, but constructs prompts based on pre-stored semantic templates and constraints (such as "date field must appear" and "safety notice must appear"), making it easier for generative AI models to produce structured outputs that conform to a predetermined document format, thereby controlling the generation quality at the model invocation level.

[0313] 3. Generative Artificial Intelligence Model Inference Module The server uses a generative AI model inference module, passing prompts as input to the generative AI model via an HTTP interface or local process call. The generative AI model employs a Transformer-based neural network, including components such as multi-head self-attention layers, feedforward layers, layer normalization, and positional encoding. During the training phase, the model undergoes unsupervised pre-training and supervised fine-tuning using a large amount of Chinese or multilingual text data.

[0314] During inference, the server encodes the prompts as a sequence of sub-words, transforms them into high-dimensional vectors through an embedding layer, and performs matrix multiplication, weighted summation, and non-linear activation operations within a multi-layer self-attention structure to progressively generate output text containing the activity plan. The model employs an autoregressive decoding approach; at each time step, the server inputs the previously generated token sequence into the model and selects the next token based on the output probability distribution. The server can adjust temperature parameters and Top-k or Top-p sampling thresholds to control the diversity and stability of the generated text.

[0315] When training the model, the server uses the cross-entropy loss function to measure the difference between the generated sequence and the reference sequence, calculates the gradient using the backpropagation algorithm, and updates the network weights using stochastic gradient descent or a variant optimizer (such as Adam). To improve robustness, data augmentation techniques, such as synonym replacement and sentence transformation, can be applied during training.

[0316] Through the structured training and inference described above, the server can form statistical expressions for specific document types such as "mountain climbing plan" and "activity plan document" within the model, making the generated results more consistent in format and content. This reduces the server's burden of format correction in the post-processing stage and improves the overall processing speed.

[0317] 4. Document Generation and Format Conversion Module The server uses a document generation and format conversion module to convert the natural language text output by the generative artificial intelligence model into document data that can be submitted to external organizations. The server first performs grammatical analysis on the generated text according to paragraph and heading tags, extracting logical structures such as "basic information paragraphs," "itinerary paragraphs," and "safety precautions paragraphs." Based on the field requirements preset by the external organization, the server maps the corresponding text fragments to form fields, such as "destination field," "date field," and "contact information field."

[0318] The server calls a document generation library (such as a PDF generation library or document template engine), populates these fields into a predefined document template, and adds headers, footers, page numbers, and encoding information. The server saves the generated document as a file in an information storage device and records the file path, generation time, and submission status in a database. Compared to simply storing raw text, this structured document generation method enables the server to quickly manage and retrieve large amounts of documents in a unified manner, improving the technical performance of document management.

[0319] 5. Location Information and Geographic Information Processing Module The server uses a location information and geographic information processing module to receive location information data sent by the terminal through a communication device. The server associates each piece of location information data with a user identifier, activity identifier, and timestamp, writes it into a location record table, and generates a spatial index (e.g., an R-tree index) using a geographic extension component. This index structure allows the server to efficiently retrieve adjacent records from multiple trajectories and a large amount of geographic point data.

[0320] When retrieving emergency contact information and shelter location information, the server uses spatial query functions to perform searches on the geographic information database. For example, using the user's current location as the center, the server calculates the distance from the current location to each emergency contact point or shelter location through range queries or nearest neighbor queries, and sorts them in ascending order of distance. The server employs vectorized computation and index scanning strategies to reduce the number of full table scans, thereby achieving low-latency retrieval on large-scale geographic datasets.

[0321] In emergency notification scenarios, the server performs a search for the nearest refuge location based on the latest location information, selects the refuge location with the shortest distance, and returns it to the terminal. This type of spatial computation relies on the database's internal geometric operators and accelerated indexes, significantly outperforming manual queries or simple linear traversal, demonstrating the technical effectiveness of efficient utilization of computing resources at the geographic information processing level.

[0322] 6. Trajectory Reconstruction and Statistical Analysis Module The server uses a trajectory reconstruction and statistical analysis module to perform path construction and statistical calculations on historical location data. First, the server sorts all location information records corresponding to a specific activity identifier according to timestamps, forming an ordered sequence of points. The server calculates the geographical distance between adjacent points using great circle distance or a planar approximation formula, accumulating the total travel distance. The server calculates the cumulative ascent and descent altitudes based on the difference in the altitude field, and calculates the total activity time based on the start and end times.

[0323] The server can further use trajectory smoothing algorithms (such as moving average filtering and Kalman filtering) to remove positioning noise, thereby reducing jitter points while maintaining the true path. The server stores the reconstructed trajectory in vector form (latitude and longitude sequence) or raster form for generating trajectory maps or further analysis. The server can also compare the statistical results with the original plan content generated by the generative artificial intelligence model, such as comparing the deviation between planned time and actual time, and between planned route and actual route, to generate activity record data containing "planned vs. actual comparison information." This comparison information is not only a business result but also a structured data product that can be used for subsequent model retraining or risk assessment algorithms.

[0324] III. Terminal-side functions and user interaction The terminal uses a client application to present the user with an input interface, a map interface, and an activity log interface. The terminal calls the location service interface provided by the operating system to obtain longitude, latitude, and altitude from the global positioning module. The terminal encapsulates the location information into data packets through the network communication stack and sends them to the server periodically or when triggered by events.

[0325] When the terminal receives emergency contact information and shelter location information from the server, it invokes the map rendering component to draw the user's current location and multiple target points on the graphical interface, displaying information such as distance and estimated time. When the user triggers an emergency operation, the terminal again obtains high-precision location information, adds an emergency marker to it, and sends it to the server to achieve rapid response in emergency situations.

[0326] When displaying activity log data, the terminal retrieves trajectory information and statistical results from the server and presents them in chart form, including total distance, cumulative elevation gain, and height profile. These visualizations rely on structured data generated by the server, enabling users to intuitively understand the activity trajectory and results.

[0327] IV. Characteristics and Technological Improvements of Generative Artificial Intelligence Models When the server invokes the generative AI model, it maps the structured fields (destination, date, route, participant information, etc.) embedded in the prompts to a high-dimensional semantic space. Compared to traditional text splicing based on fixed templates, this model uses a multi-head attention mechanism to simultaneously focus on multiple contextual features, enabling the generated results to maintain information consistency across different paragraphs. For example, it consistently references the same date and route information in the "Itinerary" and "Safety Precautions" paragraphs, thereby reducing logical errors.

[0328] The server analyzes the input information through a prompt generation module and dynamically adjusts the structure of the prompts based on missing information. Essentially, this imposes "structured constraints" on the model at the input level, preventing the model output from becoming too divergent or omitting key fields. By utilizing this adaptive prompt generation method, the server improves the usability of the generated text and the success rate of post-processing within the same model size, thereby reducing reliance on manual review and correction. This represents a significant improvement in the computer's internal processing capabilities, moving from "human-written documents" to "automatic model documentation."

[0329] During model training, the server utilizes trajectory data, statistical results, and historical document data as part of the training corpus, enabling the model to learn domain-specific writing styles and safety prompt patterns. This fine-tuning method, incorporating domain data, improves the model's accuracy and stability in the target application scenario. By defining dedicated loss function weights, the server can impose higher penalties on errors involving "missing key fields," thereby guiding the model to focus more on field completeness during training.

[0330] V. Technical Effects and Causal Relationship The server achieves integrated data flow between internal modules by uniformly managing basic data, prompts, model outputs, and geographic data. Because the server uses structured data representation at each stage and accelerates processing through indexing and spatial computation, the overall system can significantly reduce database queries and network round trips when handling large amounts of user activity, thereby improving processing throughput.

[0331] The server employs efficient vectorized computation and filtering algorithms during trajectory reconstruction and statistical analysis, enabling trajectory calculation to be completed within an acceptable timeframe even when dealing with high-frequency sampling points. Combined with the planned content of the generative artificial intelligence model, the server forms a closed-loop data structure of "planning-actual-statistics," providing a basis for subsequent model improvement and system optimization.

[0332] Through the aforementioned technical means, the server not only automates the original process of manually writing plans and manually searching for refuge locations, but also improves the internal processing methods of computers in terms of text generation, data structure design, spatial computing, and statistical analysis. This enhances the generation speed, decision-making accuracy, and data management efficiency, while reducing the error rate and communication load, thus resulting in substantial progress at the level of computer technology.

[0333] VI. Alternative Implementation Forms and Variations The server can replace the underlying software components in different implementations. For example, the database can be replaced with other database systems that support spatial indexing, and the document generation library can be replaced with a library that supports multiple output formats. However, as long as the server maintains the data processing chain of "basic data management - prompt statement generation - generative artificial intelligence model reasoning - document generation - geographic information retrieval - trajectory reconstruction and statistics - emergency shelter location determination", it can achieve the same technical effect as this implementation.

[0334] The terminal can also be implemented using different hardware platforms or operating systems. As long as the terminal can obtain location information, communicate with the server, and provide a graphical interface to the user, the functions of this invention can be realized. The communication protocol between the server and the terminal can use different versions of secure transmission protocols, but this does not affect the core technology of this invention.

[0335] Through the above-mentioned multiple implementation forms and variations, the server, terminal and user form a complete technical solution in the system of the present invention, which enables the generation of activity plans, document processing and geographic information processing using generative artificial intelligence models and prompts to be realized in a computer in an efficient, accurate and structured manner, providing an improved computer technology infrastructure for the field of mountaineering and outdoor activities.

[0336] use Figure 13 The processing flow is explained.

[0337] Step 1: The user enters activity information on the terminal and submits it. After launching the application on the terminal, users can enter activity information such as destination, start and end dates, planned route, participant names and contact information, and emergency contact information in the graphical interface.

[0338] Input: Text and option data entered by the user through the terminal interface.

[0339] The terminal performs format validation on the input data (e.g., checking date format, telephone length, and required fields), encapsulates each field into a structured data object, and then serializes it into a data packet for network transmission.

[0340] Data processing and computation: The terminal maps the interface form items to a key-value pair structure, removes spaces from the strings, standardizes the date strings (e.g., unify them to YYYY-MM-DD), and generates activity identifiers or temporary identifiers locally.

[0341] Output: The terminal generates request data containing activity information and sends the request to the server through the communication module.

[0342] Step 2: The server receives activity information and generates basic data records. The server receives request data from the terminal at the communication interface and parses and verifies the data packets.

[0343] Input: Structured data sent by the terminal, including fields such as destination, time, route, participants, and emergency contacts.

[0344] The server uses a parsing module to deserialize network data into an internal data structure, verifying field integrity and logical consistency (e.g., end date is later than start date, participant list is not empty), and then writes the data into the underlying data table in the database.

[0345] Data processing and computation: The server truncates text fields, filters illegal characters, converts date and time fields into timestamp format, and stores participant lists and emergency contact information in JSON encoding; the server generates unique event IDs and establishes indexes.

[0346] Output: The server returns response data containing the activity ID and storage success status to the terminal, and generates a basic data record in the database.

[0347] Step 3: The server generates a prompt statement based on the basic data. The server reads the data corresponding to the activity ID from the basic data table and selects an appropriate prompt statement template based on the field completeness.

[0348] Input: Basic data records stored internally on the server (destination, start and end dates, route, participant information, emergency contact information, etc.).

[0349] The server determines whether any fields are missing or incomplete, and determines the paragraphs and items that the prompt statement should include based on a preset rule base; then it calls the string processing module to fill the field content into the placeholder positions in the template, generating a Chinese prompt statement with clear document structure requirements.

[0350] Data processing and computation: The server performs conditional statements (if / else), template selection, placeholder replacement, and string concatenation operations to map structured fields to natural language descriptions while maintaining consistency in paragraph order and heading numbering.

[0351] Output: The server generates a complete prompt text, which is used for subsequent calls to the generative artificial intelligence model.

[0352] Step 4: The server sends prompts to the generative AI model and retrieves the generated result data. The server uses the model invocation module to send prompts as input to the generative artificial intelligence model via a local process interface or an HTTP interface.

[0353] Input: The prompt text generated by the server, and optional model parameters (such as maximum length, temperature value, sampling strategy, etc.).

[0354] Internally, the generative AI model segments or subdivides the prompts into words, mapping them to vector sequences. These sequences are then processed sequentially through a multi-head self-attention layer and a feedforward layer, undergoing matrix multiplication, weighted summation, and nonlinear transformations. Autoregressive decoding is used to generate the event plan text token by token. The server controls the model's inference process at an outer layer, receiving the model's output sequence.

[0355] Data processing and computation: During the call process, the server encodes the prompt statement and maps it into an integer ID sequence according to the model vocabulary. Then, the model performs floating-point matrix operations and probability distribution calculations. The server decodes the output token sequence, restores the integer IDs to natural language text, and can perform simple filtering on the multi-round generated results (e.g., removing obviously incomplete sentences).

[0356] Output: The server receives generated result data containing the activity plan content, i.e., the complete text form of the activity plan content.

[0357] Step 5: The server performs structured parsing and document format conversion on the generated data. The server extracts the logical components from the generated result data, maps them to a predefined document field structure, and generates a submitable electronic document.

[0358] Input: The project proposal text output by the generative artificial intelligence model.

[0359] The server uses a text parsing module to divide the text into paragraphs such as "Basic Information," "Itinerary," and "Safety Precautions" according to title keywords, serial numbers, or line break rules. It then uses simple rules or pattern matching to extract and validate the destination, date, route, and contact information. Subsequently, the server calls a document generation library to populate these paragraphs into PDFs or other document templates, generating document files with headers, footers, and standardized formatting.

[0360] Data processing and operations: The server performs string searching, splitting, pattern matching (such as matching dates and phone numbers using regular expressions), field mapping, layout calculations (such as pagination, line spacing, and font size settings), and converts the document content into a binary file stream.

[0361] Output: The server generates document data files that meet the submission requirements of external organizations and records the document file path and its associated activity ID in the database.

[0362] Step 6: The server submits document data to external organizations and records the submission status. The server selects an appropriate submission channel (API submission or email submission) and automatically sends the document data to external organizations.

[0363] Input: The generated document data file, the communication address information of the external organization (interface URL, email address, etc.), and the activity ID.

[0364] The server generates an email with attachments through the email sending module, or constructs a multi-part form request through the HTTP client module to submit the document and necessary metadata together; after receiving a successful response or confirmation of sending without errors, the server writes the submission status to the database.

[0365] Data processing and computation: The server encodes document files (such as Base64 encoding or block encapsulation), constructs network data packets, executes retry and timeout control logic, and parses and judges response codes.

[0366] Output: The server generates an updated commit status record (e.g., "not committed" changes to "committed") and can return a commit result notification to the terminal.

[0367] Step 7: The terminal obtains its current location and reports the location information data to the server. The terminal can call the positioning interface in the background or foreground to obtain the longitude, latitude and altitude of the current location from the built-in positioning hardware.

[0368] Input: The original location coordinates and timestamp obtained by the terminal from the positioning module.

[0369] The terminal performs simple filtering on the coordinate values ​​(such as discarding invalid values ​​and smoothing nearby points), and packages the activity ID, user identifier, and coordinate data into location reporting data, which is then sent to the server periodically or triggered by events through the communication module.

[0370] Data processing and computation: The terminal performs simple numerical checks, coordinate precision truncation, and timestamp standardization to convert the data into a unified format for the server to parse.

[0371] Output: The terminal outputs a standardized location information data packet to the server.

[0372] Step 8: The server receives location information data and stores and updates the spatial index. The server receives location information data sent by the terminal at the location reporting interface, writes it into the location record table, and updates the relevant spatial index.

[0373] Input: Longitude, latitude, altitude, timestamp, activity ID, and user identifier reported by the terminal.

[0374] The server verifies each location information (coordinate range, time sequence) and then writes it to the database. At the same time, the server calls the geographic extension module to maintain a spatial index for the new record, so that subsequent nearby searches can be performed on the index structure.

[0375] Data processing and computation: The server performs insertion operations, updates spatial index nodes, sets time sorting markers, and marks or ignores invalid or abnormal points.

[0376] Output: The server generates an updated location information database state, providing a foundation for subsequent emergency searches and trajectory reconstruction.

[0377] Step 9: The server retrieves emergency contact information and evacuation location information based on location data. When the server needs to provide emergency information to the user, it reads the user's latest location information from the database and performs a spatial query.

[0378] Input: The user's current or most recent latitude and longitude coordinates, pre-stored emergency contact points, and geographic information of shelter locations.

[0379] The server calculates the distance between the current location and various emergency contact points and refuge locations using spatial functions from the geographic information database, performs nearest neighbor or range filtering, and then selects a number of records in ascending order of distance.

[0380] Data processing and operations: The server performs vector distance calculations, spatial index scans, and sorting operations, and converts the selected records into structured data containing names, contact information, coordinates, and brief descriptions.

[0381] Output: The server generates a list of emergency contact information and a list of refuge locations, and sends them to the terminal via the communication module.

[0382] Step 10: The terminal displays emergency contact information and evacuation location information. The terminal receives emergency contact points and shelter location data returned by the server and displays them visually on the user interface.

[0383] Input: Emergency contact information (name, phone number, address, etc.) and refuge location information (name, coordinates, distance, etc.) sent by the server.

[0384] The terminal calls the map component to draw the user's current location and multiple target points on the map, while displaying information such as name, phone number, and distance in the list area, and binding a one-click call function to the phone number field, and binding navigation or route preview functions to the refuge location.

[0385] Data processing and computation: The terminal performs interface operations such as coordinate projection transformation (e.g., latitude and longitude to screen coordinates), icon drawing, text layout, and click event binding.

[0386] Output: The terminal outputs a graphical interface to the user, which includes map annotations and list information, to assist the user in making emergency contact and evacuation decisions.

[0387] Step 11: In an emergency, the user operates the terminal to send an emergency notification message. When encountering an emergency, users can click the emergency button on the terminal interface.

[0388] Input: The user's emergency operation command.

[0389] After receiving the operation, the terminal displays a confirmation prompt. After the user confirms, the terminal obtains the current location again, adds an emergency marker to the location information, generates an emergency notification data packet, and sends it to the server.

[0390] Data processing and computation: The terminal adds an emergency flag field to the ordinary location reporting data structure and adds the current timestamp and high-precision coordinates to the data packet.

[0391] Output: The terminal outputs location information data with an emergency status marker to the server.

[0392] Step 12: The server determines the nearest refuge location based on the latest location information and returns the result. After detecting an emergency notification, the server extracts the latest latitude and longitude from the data packet and uses this as the center to perform a nearest neighbor search in the evacuation location information database.

[0393] Input: Current location coordinates with an emergency marker and coordinates of all locations in the evacuation site database.

[0394] The server performs nearest neighbor queries on refuge locations using spatial indexes, calculates the distance between each candidate refuge location and the current coordinates, selects one or more locations with the shortest distance, and can call the route service to estimate walking time.

[0395] Data processing and computation: The server performs distance function calculations, candidate set filtering for index optimization, minimum value selection, and optional time estimation formulas (such as converting distance to time based on average walking speed).

[0396] Output: The server generates information about the nearest refuge location (name, coordinates, distance, estimated time required, etc.) and returns it to the terminal through the communication module.

[0397] Step 13: The terminal receives information about the nearest evacuation location and guides the user there. After receiving the information about the nearest refuge location from the server, the terminal highlights that location on the map interface.

[0398] Input: The name, coordinates, distance, and estimated time of the nearest refuge location returned by the server.

[0399] The terminal draws a route preview from the current location to the refuge location on the map (by calling the route planning interface of the map SDK), and displays prompts such as "distance" and "estimated time" on the interface. At the same time, continuous positioning can be enabled to update the user's position relative to the refuge location in real time.

[0400] Data processing and computation: Terminal execution path drawing, interface refresh and voice prompt triggering, etc.

[0401] Output: The terminal provides users with an intuitive navigation interface and prompts to help them quickly reach a refuge location in an emergency.

[0402] Step 14: The server reconstructs and statistically analyzes historical location data and generates activity record data. After the event ends, the server queries the location record table for all historical location data corresponding to the event ID, sorts and analyzes them.

[0403] Input: All latitude, longitude, timestamp, and altitude records associated with a given event ID.

[0404] The server first sorts the records by timestamp to obtain a continuous sequence of points in time. Then, it calculates the distance and height difference between adjacent points, accumulating the total movement distance and the cumulative ascent and descent heights. The server also calculates the total activity time and can generate curve data of height changes over time. The server can use a smoothing algorithm to remove noise points and generate a trajectory line that more closely resembles the real path.

[0405] Data processing and computation: The server executes sorting algorithms, distance calculation formulas, cumulative height difference calculations, smoothing filters, statistical summaries, and structures the results into activity record objects containing multiple indicators.

[0406] Output: The server generates activity log data containing trajectory, statistics, and optional comparison information between the plan and the actual situation, and stores it in the database. It can also be transmitted to the terminal for display as needed.

[0407] Application Example 2 The process flow corresponding to the specific processing in Use Case 2 will be described below. The various parts of the system described below are implemented by the data processing device 12 and the intelligent device 14. In addition, the data processing device 12 is referred to as the "server" and the intelligent device 14 is referred to as the "terminal".

[0408] In existing technologies, computer-based mountaineering support systems, emergency communication systems, and behavior recording systems are mostly independent application modules. On the one hand, itinerary plans are usually manually edited by users or generated through fixed templates, and the server only performs simple storage and retrieval operations. It is unable to automatically construct high-quality prompts based on user context and call generative artificial intelligence models to dynamically generate and adjust the plan. On the other hand, location information collection and emotion information collection are often separated from the plan generation logic, making it difficult for the server to perform integrated processing and calculation optimization of "plan - real-time location - emotional state - post-event recording" under a unified data model.

[0409] Specifically, the existing system has at least the following computer technology problems: (1) At the planning level, the server usually only receives structured fields (such as time and location) submitted by the user, and then generates static planning documents through rules or fixed templates. It lacks a program flow for constructing prompts for generative artificial intelligence models, and cannot flexibly adjust the content of prompts according to user input and environmental status. As a result, the quality of the generated results depends on manual editing, and the computing resources on the server side cannot be intelligently utilized.

[0410] (2) At the emergency response level, although nearby institutions can be queried based on location information, the server lacks a mechanism to jointly calculate location information and emotional state. It cannot infer the "dangerous state" on the server side through rules or models, so as to automatically trigger machine-generated emergency notifications at the appropriate time. At the same time, the emergency notification content in the existing system is mostly manually entered, and there is a lack of a unified data processing path for the server to automatically combine "current location + plan document information + emotional state", which increases the risk of communication delay and error.

[0411] (3) In terms of behavior recording and content generation, traditional systems generally only store trajectory data or simple logs on the server. They lack a computational process to aggregate trajectory data and emotional timelines on the server and integrate them in a time sequence. They also lack a technical solution to automatically construct prompt statements and call generative artificial intelligence models to generate high-level narrative content and multimedia content (such as videos) on this basis. Therefore, they cannot effectively reuse the deep learning reasoning capabilities of the server to achieve automated and structured “behavioral storytelling” output.

[0412] (4) At the overall system architecture level, in the existing technology, the user terminal, server, external communication service and external model service are often just simple point-to-point calls. There is no systematic program design around the unified processing link of "prompt statement - model reasoning - document generation - multi-source data fusion - automatic notification - story generation", which makes it difficult for the server to uniformly schedule and optimize multiple modules, resulting in fragmented computing processes and poor scalability and maintainability.

[0413] Therefore, a new computer implementation is needed to enable the server to: The program logic automatically generates and adjusts prompts for generative artificial intelligence models. The server-side centrally manages and processes user input information, location information, emotional state information, and emergency organization information; When a dangerous situation is detected, the server automatically triggers a communication service to issue an emergency notification. After activities such as mountain climbing are completed, the server automatically generates multi-layered narrative and multimedia content based on a unified dataset, thereby improving the computing process, resource utilization and reliability at the system level, and achieving integrated intelligent support for the entire lifecycle of user behavior.

[0414] The specific processing performed by the specific processing unit 290 of the data processing apparatus 12 in Application Example 2 is achieved by the following means.

[0415] In this invention, the server includes: a module for receiving input information related to a behavior plan from a user and automatically generating text containing prompts on the server side based on the input information to instruct a generative artificial intelligence model to generate the behavior plan; a module for obtaining text data about the behavior plan from the generative artificial intelligence model and automatically embedding the text data into a predetermined format template on the server side to generate an electronic document of the plan for submission to an emergency response agency; an information processing module for controlling the automatic transmission of the electronic document to the predetermined emergency response agency via a communication network and notifying the user information processing terminal of the transmission result via a communication interface; and a module for obtaining location information from the user information processing terminal at predetermined time intervals, persistently storing the location information in the server's data storage in association with the behavior plan, and transmitting the information through a server-side calculation program based on... The system includes: a module for generating behavior path record data from stored location information; a module for retrieving nearby support organizations relative to the user's current location on the server based on the location information, and packaging the contact and location information of the support organizations before sending them to the user information processing terminal; a module for obtaining information about the user's emotional state on the server side or in conjunction with an external emotion recognition service, and dynamically changing the prompts or behavior plans input to the generative artificial intelligence model based on the emotional state in the server's internal logic, so as to achieve automatic adjustment of the behavior plan on the server; and a module for integrating the location information and the emotional state information in chronological order on the server side to generate behavior process record data, and providing the record data as input conditions to the generative artificial intelligence model through new prompts to automatically generate narrative content about the behavior process. This allows for the formation of an end-to-end computing process within the server, encompassing user input, prompt generation, model inference, document generation, location and emotion data fusion, danger detection and emergency notification, and post-event narrative generation. This enables unified management and automated processing of data processing and computation, significantly improving the intelligence, response speed, resource utilization efficiency, and scalability of computer systems in behavioral support scenarios such as mountaineering.

[0416] "User" refers to an individual or group that provides input information to the system through an information processing terminal, receives output results from the system, and is supported by the system during the process of behavior.

[0417] "Behavior plan" refers to a set of structured or semi-structured data about the time, location, route, steps, and related resource arrangements of a user's future behavior process, generated by a generative artificial intelligence model or adjusted by a server based on user input information.

[0418] "Input information" refers to the data provided by the user to the server through the information processing terminal that is related to the generation of the behavior plan, including but not limited to destination, time, difficulty, format, participant information and preference conditions.

[0419] "Prompt statements" refer to natural language text instructions that are automatically generated by the server based on user input, location information, emotional state information, and other contextual information, and are used to instruct generative artificial intelligence models to generate or adjust behavior plans and related content.

[0420] "Generative artificial intelligence models" refer to data processing models trained using machine learning and deep learning techniques that can automatically generate text, speech, or other content based on prompts.

[0421] "Plan document" refers to an electronic document that is automatically generated by a server according to a predetermined format template based on the generated action plan, and is used to submit to emergency response agencies or other management agencies.

[0422] "Electronic documents" refer to file data generated by a server through data processing programs and stored electronically, including but not limited to text documents, spreadsheets, image files, and printable document files.

[0423] "Emergency response agencies" refer to public or private organizations that need to be notified and have the authority to respond when users encounter danger or abnormal situations, including but not limited to rescue organizations, security management organizations, and medical institutions.

[0424] "Information processing terminal" refers to an electronic device carried or used by a user for inputting information, receiving output from a server, and communicating with the server, including but not limited to smartphones, tablets, wearable devices, and computer terminals.

[0425] "Location information" refers to data obtained through positioning technology that represents a user's location in geospatial space, including at least latitude and longitude information, and may further include altitude information and timestamp information.

[0426] "Behavioral path record data" refers to a set of data generated by the server based on location information stored in chronological order, used to represent the user's movement trajectory and spatial distribution characteristics over a period of time.

[0427] "Support organizations" refer to organizations that can provide assistance, services or emergency support to users during user behavior, including but not limited to security organizations, medical institutions and service facility management organizations.

[0428] "Contact information" refers to identifying information used to establish communication with support agencies or emergency response agencies, including but not limited to telephone numbers, network addresses, and communication accounts.

[0429] "Emotional state" refers to the psychological or emotional characteristics of a user at a specific point in time, which are inferred by the system based on the user's voice, facial expressions, text, or other behavioral characteristics. These characteristics include, but are not limited to, fear, tension, sadness, joy, and excitement.

[0430] "Emotional state information" refers to data used to represent emotional states, including emotion category, emotion intensity, time of occurrence, and correspondence with behavioral processes.

[0431] "Behavioral process record data" refers to a data set that comprehensively organizes the user's location information, emotional state information, and related event information in chronological order during the behavioral process, and is used to represent the overall situation of the behavioral process.

[0432] "Narrative content" refers to textual information generated by a generative artificial intelligence model based on behavioral process recording data, which describes the user's behavioral process, emotional changes, and related events in natural language.

[0433] "Image information" refers to visual data represented in the form of static or dynamic images, including but not limited to photographs, icons, map screenshots, and video frames.

[0434] "Audio information" refers to playable digital data recorded in the form of sound waves, including but not limited to voice narration, ambient sounds, and background music.

[0435] "Video content" refers to multimedia data that can be played continuously after integrating image information, audio information, and location information on a timeline.

[0436] "Communication network" refers to the network infrastructure used to transmit data between servers, information processing terminals and external services, including but not limited to wired networks, wireless networks and mobile communication networks.

[0437] "Communication service program" refers to an application or interface component that runs on a server or external communication platform and is used to automatically send messages, voice or data to a predetermined communication target.

[0438] "Emergency notification message" refers to message data automatically generated and sent by the server through the communication service program when a predetermined dangerous situation is presumed to exist, used to notify emergency response agencies or support agencies. It includes at least the user's current location and relevant information about the plan document.

[0439] In various embodiments of this invention, the system mainly consists of a server, a terminal, and a user. The server, as the core information processing device, performs various data processing and calculations, including generating prompts, calling generative artificial intelligence models, generating planning documents, fusing location and emotion data, inferring dangerous states, and generating narrative and video content. The terminal, as the user-side information processing terminal, performs user input collection, location information collection, emotion-related information collection, and communication processing with the server. The user, as the main user of the system, provides input information to the server through the terminal and receives various types of content output by the server.

[0440] I. Overall Hardware and Software Structure The server can be a cluster of computing devices deployed on a cloud computing platform or in a local data center, including at least one central processing unit and an optional graphics processing unit. In one embodiment, the server uses a general-purpose server hardware platform and accelerates the inference of generative artificial intelligence models through the graphics processing unit. The operating system running on the server can be a general-purpose operating system, the server-side application can be implemented using Python or other general-purpose programming languages, and the server-side web framework can be a general-purpose web framework. The server can also use relational or non-relational databases to store user input information, location information, emotional state information, planning document indexes, behavioral path record data, and narrative content, etc.

[0441] The terminal can be a smartphone, tablet, or wearable device, running a mobile operating system and equipped with dedicated applications. It uses the positioning interface provided by the operating system to obtain location information and utilizes camera, microphone, and user interface components to collect emotion-related and textual information. The terminal exchanges data with a server via a wireless communication network.

[0442] In one embodiment, the server invokes an external generative artificial intelligence model service. This model is a multi-layer neural network employing a transformer architecture. The model includes word embedding layers, multi-layer self-attention encoders, and decoder modules, and is pre-trained on a large-scale text corpus. The server sends prompts to the model through a unified interface and receives the text output generated by the model. The server can also invoke an external emotion recognition service, which can also classify emotions from images, audio, or text based on convolutional neural networks, recurrent neural networks, or transformer architectures.

[0443] II. Server-side functional modules and data structures In a preferred embodiment, the server includes the following functional modules, which can be implemented in software as different program modules or microservices, or in hardware as dedicated logic or a coprocessor: 1. Prompt Statement Generation Module The server uses a prompt generation module to extract structured fields from user input, such as destination, difficulty level, behavior type, estimated departure time, and number of participants. The server stores these fields in data using a key-value structure or relational table structure. The prompt generation module then injects these fields into a predefined natural language template to generate prompts adapted to the generative artificial intelligence model.

[0444] In one example, the server generates the following prompt: "The user plans to undertake a mountain climbing activity. Destination: Mt. Fuji; Difficulty: Intermediate; Mode: Day trip; Departure time: 6:00 AM, August 1, 2026; Number of participants: 2. Please act as a professional mountain guide and generate a detailed mountain climbing plan in Chinese, including: recommended route, time allocation (approximate time for each section), required equipment list, risk warnings, and precautions." In another example, the server generates the following prompt statement in order to produce a plan document that can be submitted to emergency agencies: "Please generate a mountaineering plan that can be submitted to the local emergency management agency based on the following information, including departure time, return time, route summary, member information, and emergency contact information: Destination: Mt. Fuji; Difficulty: Intermediate; Mode: Day trip; Departure time: 2026-08-01 06:00; Estimated return time: 2026-08-01 18:00; Number of people: 2." The prompt generation module parameterizes templates through program logic, dynamically inserting additional constraints based on user emotional state, environmental conditions, or historical records, such as "emphasis on safety" or "altitude restriction." The server thus no longer simply performs static rule replacement, but instead encodes multi-source data into prompts through computational rules, thereby improving the relevance and stability of the generative AI model's output.

[0445] 2. Generative Artificial Intelligence Model Interface Module The server communicates with external generative AI model services through a generative AI model interface module. The server packages prompts into request payloads and sets hyperparameters such as temperature, maximum output length, and penalty coefficients. During training, the model employs a cross-entropy loss function and updates weights using gradient descent or its variants; data augmentation strategies (such as synonym replacement and sentence structure transformation) can be used to improve the model's robustness to diverse inputs.

[0446] During inference, the server performs the following data processing: encodes the prompt statement into a vector representation, calls the model for forward propagation, uses a self-attention mechanism to weighted aggregate dependencies between different terms across multiple encoding layers, and then the decoder generates the output sequence based on the attention weights. After receiving the output, the server performs sentence segmentation, removes invalid symbols, and extracts key information. Internally, the server establishes a unified structure for the output text, storing it as logical fields such as titles, paragraphs, and tables to facilitate processing by the subsequent document generation module.

[0447] In another example, the server generates prompts to adjust the mood-guided plan: "The user's original plan was an intermediate day trip to Mount Fuji. The current sentiment analysis shows 'anxiety' and 'nervousness'. Please plan an alternative Mount Fuji climbing plan with a lower elevation gain, wider trails, and easier evacuation, while ensuring safety, and provide your reasons." The server sends the prompt to the generative AI model and replaces or supplements key fields in the original plan based on the returned content. Unlike traditional rule-based adjustments, the server interacts with the model through multiple rounds of prompts, encoding continuous logical constraints into natural language, thereby achieving flexible transformation of the plan generation behavior without altering the underlying code rules.

[0448] 3. Planning Document Generation Module The server uses a template engine and document generation components to embed text data output by a generative artificial intelligence model into a predefined planning document template. The server predefines field placeholders for the planning document template, such as "destination", "departure time", "estimated return time", "main route", "emergency contact information", etc., and organizes the template data in the server's memory in the form of a tree structure or key-value pairs.

[0449] The server parses the generated text into logical paragraphs and performs field matching based on key phrases, such as identifying departure and return times based on time expressions and matching route fields based on route names. The server then invokes a document rendering engine to convert the completed template into an electronic document format and stores it in a file system or remote storage service. The server records the document file path and metadata to support subsequent retrieval and automatic notifications.

[0450] 4. Location and Behavior Path Management Module The server receives location information periodically uploaded by the terminal through the location and behavior path management module. This location information includes a plan identifier, timestamp, longitude, latitude, and an optional altitude value. The server stores these records in a time-series data table structure in the database and uses an index structure to accelerate queries by plan identifier and time range.

[0451] In one embodiment, the server uses a spatial extension component to perform spatial indexing management of location information to support efficient neighbor location search and trajectory generation. When generating behavioral path record data, the server interpolates, denoises, and calculates distances to the time-series coordinates, outputting a path data structure containing fields such as distance, time consumption, and cumulative elevation gain for each interval. The server uses this path data structure for trajectory visualization and subsequent narrative content generation.

[0452] 5. Emotional State Management Module The server receives or requests emotion recognition results through the emotion state management module. The terminal can send images, voice, or text to an external emotion recognition service, which returns emotion tags and intensity scores. The server stores this emotion state information in an emotion record table, along with corresponding timestamps, location information, and plan identifiers.

[0453] In one embodiment, the server sets rules for determining dangerous situations. For example, when the intensity of a certain type of emotion (such as fear or sadness) exceeds a predetermined threshold and the duration of this state exceeds a preset time period, the server will infer that a potential dangerous situation exists. At this time, the server can proactively trigger an emergency notification process without requiring the user to manually call for help. Because the server considers time continuity, emotional intensity, and location signal quality simultaneously in its determination, it has higher accuracy and robustness than rules based on only a single event.

[0454] 6. Emergency Notification and Organization Search Module The server manages emergency response organization information through an emergency notification and organization retrieval module. For each organization, the server stores its type, geographic coordinates, and communication identifiers (such as telephone number and network address) in the database, and uses a spatial indexing mechanism to enable fast proximity queries. When the server needs to provide a list of support organizations to a user, it calculates the distance between the user's current location and the coordinates of each organization, selects the closest few, and returns them to the terminal.

[0455] When the server determines that a user is in a dangerous state, it calls the communication service program interface to generate an emergency notification message. The server extracts the current coordinates from behavior path records and a plan summary from planning documents, combining this information with the user's emotional state to form the notification text. The server then sends an SMS or voice call via an external communication platform interface and records the sending result in the database. This completes the technical process from data collection and rule inference to automatic notification, improving emergency response speed and reducing the chance of human error.

[0456] 7. Behavioral Process Narration and Video Content Generation Module The server integrates path data and emotional timelines through a behavioral process narration and video content generation module to generate advanced narrative and video content. First, the server segments the recorded location information, identifying key time points such as departure, rest stops, approaching the destination, and return. Then, the server aligns the emotional records with time to obtain the emotional state corresponding to each key point.

[0457] The server constructs new prompts and converts this structured data into natural language conditional input, for example: "The following is the trajectory data (including altitude changes) and emotional events of a mid-level day climb of Mount Fuji (e.g., feeling tired around 10:00 AM, reaching the summit around 11:30 AM and recording a strong sense of accomplishment). Please write a mountain climbing review article in Chinese from a first-person perspective, with the following structure: feelings before setting off, difficulties during the climb, the moment of reaching the summit, the descent and reflections, highlighting the sense of accomplishment at the summit and the emotional changes throughout the process." The server sends the prompt to a generative artificial intelligence model, which retrieves the narrative content and stores it in the storage system. Subsequently, the server can invoke a speech synthesis service to convert the narrative content into an audio file, and then use a multimedia processing component to combine user-uploaded photos, trajectory map images, etc., to generate video content. The server indexes and saves the generated video content and provides a playback link to the terminal.

[0458] III. Terminal-side functions and their collaboration with the server In this invention, the terminal primarily undertakes the tasks of data collection and interaction with the server. Users input information such as destination, difficulty level, and behavior type through the terminal interface. The terminal encapsulates this data into structured messages and sends them to the server. The terminal periodically obtains location information through a positioning interface, uploading it in batches when the network is available and caching it locally when the network is unavailable.

[0459] The terminal can activate foreground services upon detecting user initiation behavior, continuously collecting location information in the background. It adaptively adjusts the sampling interval based on system power consumption strategies to reduce energy consumption and communication load. The terminal can also collect images and audio clips via camera and microphone, measure emotional states using external emotion recognition services, and report the results to the server. In emergencies, users can use the terminal interface to make a one-click call to the nearest support organization, and also view server-generated plans, narratives, and video content.

[0460] IV. Technical Effects and Improvements in Computer Technology Through the modular design and data processing flow described above, the server improves computer technology itself in several ways: 1. By generating prompt statements, the server maps structured data and contextual conditions into natural language instructions, enabling the same generative AI model to adaptively output different prompt statements in different tasks. This reduces reliance on multiple dedicated rule engines, thereby simplifying the system code structure, improving maintainability, and reducing the number of model calls and communication load by caching and reusing prompt statements.

[0461] 2. The server stores location information, emotional state information, and behavior plan data in the same database with a unified identifier, and establishes an index structure for time-series data and spatial data. This reduces the query complexity of trajectory retrieval, nearby institution search, and danger state inference, thereby improving response speed.

[0462] 3. The server uses a multi-source feature fusion approach to infer dangerous states internally. Unlike the traditional approach that relies solely on user input, the server combines temporal continuity, emotional intensity, and location information to make a judgment, reducing the probability of false negatives and improving the overall prediction accuracy.

[0463] 4. The server abstracts behavioral path recording data and emotional timelines into a unified data structure, and then drives a generative AI model to generate narrative and video content through prompts, achieving an automatic conversion process from low-level sensor data to high-level semantic content. This process is optimized on the server side through pipelined processing, reducing the overhead of manual editing and data transfer between multiple systems, thereby improving overall computational efficiency.

[0464] 5. When the server interacts with external model services and communication services, it uses asynchronous calls and queuing mechanisms to queue and retry requests, preventing overall system blockage caused by network jitter or temporary errors, thereby improving the robustness and scalability of the system.

[0465] In other implementations, the server can employ a locally deployed generative artificial intelligence model, with its structure and parameters tailored to adapt to different hardware resource conditions. The server can also employ different emotion recognition algorithms or rule-based strategies to suit different application domains. The terminal can simplify its functions, performing only location acquisition and text input, centralizing all emotion recognition and multimedia processing on the server side to further reduce the terminal's burden. All different implementations are based on the generative artificial intelligence model invocation framework centered on prompt statements and the multi-source data fusion strategy based on a unified data structure proposed in this invention, thereby substantially improving the computer's internal data management, computational processes, and resource scheduling while achieving behavioral support functions.

[0466] use Figure 14 The processing flow is explained.

[0467] Step 1: Users input relevant information about their behavior plans through the terminal, which then sends the information to the server.

[0468] Users sequentially input text or optional information such as destination, difficulty level, activity type, estimated departure time, estimated return time, number of participants, and contact information into the terminal's graphical interface. The terminal organizes these fields into structured data (e.g., a set of key-value pairs) and sends it to the server as input via a secure communication protocol. Before sending, the terminal performs format checks on the input data (e.g., time format, mandatory field checks). If the checks pass, it generates a request message containing user and plan identifiers. The terminal's output is the request data containing the user's input fields, which the server then uses as input for subsequent processing.

[0469] Step 2: The server receives the plan input data sent by the terminal and generates an initial plan record in the database.

[0470] The server takes a structured request from the terminal as input and first performs syntax and integrity checks on the data (e.g., checking if the time field is a valid timestamp and if the difficulty level is within a preset enumeration value). The server generates a unique plan identifier for this plan based on the user's identifier, associates the user input fields with this plan identifier, and stores it in the plan table of the database. Data processing performed by the server during this process includes field mapping, type conversion, and default value filling (e.g., inferring an initial value based on empirical rules when no return time is specified). The server's output is a database record containing the plan identifier and the original input fields, along with a confirmation message returned to the terminal (containing the plan identifier and current status).

[0471] Step 3: The server generates prompts for generative artificial intelligence models based on the plan records.

[0472] The server takes planned records stored in the planning table as input, reads fields such as destination, difficulty, activity type, departure time, return time, and number of people, and uses string concatenation or template rendering algorithms to embed these fields into predefined natural language templates to generate prompts. The data processing includes converting structured fields into natural language fragments, assembling them into complete explanatory text according to the template order, and adding security preference or weather condition prompts as needed based on system settings. The server outputs a natural language prompt text, for example: "The user plans to undertake a mountain climbing activity. Destination: Mt. Fuji; Difficulty: Intermediate; Mode: Day trip; Departure time: 6:00 AM, August 1, 2026; Number of participants: 2. Please act as a professional mountain guide and generate a detailed mountain climbing plan in Chinese, including: recommended route, time allocation (approximate time for each section), required equipment list, risk warnings, and precautions." This prompt statement serves as input for the next step of calling the generative artificial intelligence model.

[0473] Step 4: The server invokes a generative artificial intelligence model to generate behavioral plan text based on the prompts.

[0474] The server takes the prompt generated in step 3 as input and constructs a request through the generative AI model interface module. This request includes the prompt, model name, and inference parameters (such as temperature and maximum output length). The server sends the request to the generative AI model deployed on computing resources. Internally, the model performs word segmentation, embedding, and multi-layer self-attention operations on the prompt, generating the plan text word by word through a decoder. After receiving the model output, the server performs data cleaning operations such as removing redundant blank lines, standardizing punctuation, and segmenting. The server's output here is structured plan text data, for example, divided into paragraphs such as "Itinerary Overview," "Recommended Route," "Schedule," "Equipment List," and "Risk Warning," for subsequent document generation.

[0475] Step 5: The server constructs an electronic document for submission based on the generated plan text.

[0476] The server takes the structured plan text output from step 4 as input and calls the document template module to read a predefined plan document template (including the organization header, fixed fields, and table structure). The server maps key fields in the plan text to corresponding positions in the template and generates an intermediate format (such as a markup language document) through the template engine. During data processing, the server automatically fills in the "departure time" and "estimated return time" table cells based on the time expression, fills in the "main route" based on the route description, and fills in user information in the "climber information" area. Subsequently, the server calls the document rendering component to convert the intermediate format into an electronic file format (such as a printable document), outputting a plan document electronic file with a unique file identifier, and writing the file path and associated plan identifier into the database.

[0477] Step 6: The server automatically sends the electronic plan document to the emergency response agency and notifies the terminal.

[0478] The server takes the file path and plan identifier generated in step 5 as input and retrieves the contact identifier (such as electronic address or interface address) of the corresponding emergency response agency from the agency information table based on the destination or region field. The server constructs a send request and sends the plan document as an attachment or data stream to the agency's information processing device via the communication network. Data processing includes: selecting the sending method according to the agency type, constructing the message header and message body, and processing the sending result status code. After receiving success or failure feedback, the server records the sending status in the database and generates a notification message as output, which is sent to the terminal through the downlink interface. The terminal then displays the result "Plan submitted" or "Submission failed, please try again" to the user.

[0479] Step 7: The terminal periodically collects location information and uploads it to the server.

[0480] After detecting that the user has begun executing a behavior plan, the terminal uses the geographic coordinates output by the device's positioning interface as input to acquire the longitude, latitude, optional altitude, and current timestamp of the current location at predetermined time intervals (e.g., every 60 seconds). The terminal combines these sampling results with a plan identifier and a user identifier to form one or more location records and temporarily stores them locally. When the network is available, the terminal periodically queries the local database for records that have not yet been uploaded, packages them into a batch location data list, and sends it to the server over the network. After each successful upload, the terminal marks the corresponding record as uploaded to avoid duplication. The output of the terminal in this step is a standardized set of location record data, which the server receives and uses as input for trajectory calculation.

[0481] Step 8: The server stores location information and generates behavioral path record data.

[0482] The server takes the list of location records uploaded by the terminal as input, performs a validity check on each record (coordinate range, timestamp order, etc.), and writes it to the location storage table, recording the plan identifier, timestamp, and coordinates. Then, based on all location records under the same plan, the server sorts them by time and executes a trajectory generation algorithm: calculating the distance, direction, and height difference between adjacent points, accumulating the total distance and total elevation gain, and identifying stop points (e.g., based on the duration of near-zero speed). The output of the data processing is behavioral path record data, including a list of trajectories segmented by time, the distance and time spent in each segment, and overall statistical indicators. The server saves this behavioral path record data for subsequent narrative content generation and visualization.

[0483] Step 9: The terminal acquires emotion-related data and uploads the emotion state results to the server.

[0484] During the activity, the user actively inputs text (e.g., "I'm a little scared now, the road is slippery") through the terminal, or allows the terminal to collect facial images and audio clips. Using this raw perceptual data as input, the terminal calls the emotion recognition service interface, sending the images, audio, or text to the emotion analysis service. The emotion analysis service returns the emotion category (e.g., fear, tension, joy) and intensity value as output. The terminal combines this with a timestamp, a plan identifier, and its current approximate location to construct an emotion state record, which is then uploaded to the server via the network. Specific actions performed by the terminal in this step include: camera activation and deactivation, microphone capture, text input capture, calling external interfaces, and local caching and upload control.

[0485] Step 10: The server stores emotional state information and determines dangerous states.

[0486] The server takes the emotional state records uploaded by the terminal as input and writes them into an emotional record table. Fields include plan identifier, timestamp, emotional category, and intensity value. The server then analyzes the emotional data over a period of time according to pre-set judgment rules or threshold functions: for example, calculating the average intensity and duration of a certain type of negative emotion within a certain time window. The server combines these statistical results with location information to determine whether dangerous conditions are met (such as "fear intensity greater than 0.8 and lasting for more than 5 minutes, while located in a high-risk area"). If the conditions are met, the server outputs a "dangerous state trigger event" carrying the current location information, a summary of the plan document, and the emotional state, used to trigger emergency notification processing; if the conditions are not met, it is recorded as a normal emotional change, used for subsequent narrative content generation.

[0487] Step 11: In dangerous situations, the server generates emergency notification messages and sends them via communication services.

[0488] The server takes the "dangerous state trigger event" generated in step 10 as input, queries planning documents, extracts fields such as user name, contact information, plan summary, and emergency contact person, and combines them with the current location coordinates and emotion category to construct an emergency notification text. The server's data processing includes concatenating fields from multiple data sources (plan table, location table, emotion table) into a clear descriptive text and formatting it according to a notification template. Subsequently, the server calls the communication service program interface, taking the target organization's contact identifier and the notification text as input, to send an SMS or initiate an automated voice call. The sending result returned by the communication service (success status, message identifier) ​​is recorded in the notification record table as the server's output and can further notify the terminal, informing the user that the system has automatically issued a notification.

[0489] Step 12: After the behavior ends, the server integrates path data and emotion data to generate a narrative of the behavior process.

[0490] The server takes behavioral path data and corresponding emotion data as input, aligns the two types of data according to timestamps, identifies key event nodes (such as departure, long stay, arrival at the destination, return to the starting point, etc.), and associates each node with a corresponding emotional state. Based on this structured event information, the server constructs new prompts, for example: "The following is the trajectory data (including altitude changes) and emotional events of a mid-level day climb of Mount Fuji (e.g., feeling tired around 10:00 AM, reaching the summit around 11:30 AM and recording a strong sense of accomplishment). Please write a mountain climbing review article in Chinese from a first-person perspective, with the following structure: feelings before setting off, difficulties during the climb, the moment of reaching the summit, the descent and reflections, highlighting the sense of accomplishment at the summit and the emotional changes throughout the process." The server sends the prompt to the generative artificial intelligence model, and the review article text output by the model is the server's output. The server stores this text in a record table and prepares it for subsequent video generation and terminal display.

[0491] Step 13: The server generates optional video content based on the narrative and multi-source data and provides it to the terminal.

[0492] The server takes the narrative content generated in step 12, the behavior path recording data, and the images and short videos uploaded by the user as input. First, it calls a speech synthesis service to convert the narrative content into an audio file. Then, based on the trajectory and time points, it selects corresponding photos or map screenshots to form a timeline of the material. The server uses a multimedia processing component to arrange the image frames in chronological order, uses the synthesized speech as the main audio track, and overlays background music when necessary, performing encoding operations to generate a video file. The server's output is a video clip and its storage path or access link. The server returns this link to the terminal, which displays an entry point such as "Watch Mountain Climbing Stories" on the user interface. The user can then play the video on the terminal or share it with other users.

[0493] The specific processing unit 290 sends the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires sound representing user input regarding the result of the specific processing. The control unit 46A sends the sound data representing user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the sound data.

[0494] Data generation model 58 is a so-called generative AI (Artificial Intelligence). Examples of data generation models 58 include ChatGPT (registered trademark) (accessible via the internet (URL: https: / / openai.com / blog / chatgpt)). Data generation model 58 is obtained through deep learning on a neural network. Input to data generation model 58 are prompt words containing instructions, and inference data such as sound data representing sound, text data representing text, and image data representing images (e.g., still image data or animation data). Data generation model 58 infers from the input inference data based on the instructions represented by the prompt words and outputs the inference result in one or more data forms, such as sound data, text data, and image data. Data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or induction. The specific processing unit 290 performs the aforementioned specific processing while using data generation model 58. The data generation model 58 can also be a model finely tuned to output inference results from prompts that do not contain instructions. In this case, the data generation model 58 can output inference results based on prompts that do not contain instructions. The data processing apparatus 12, etc., includes various data generation models 58, including AI other than the generation AI. AI other than the generation AI can be, for example, linear regression, logistic regression, decision trees, random forests, support vector machines (SVM), k-means clustering, convolutional neural networks (CNN), recurrent neural networks (RNN), generative adversarial networks (GAN), or Naive Bayes, and can perform various processes, but is not limited to this example. Furthermore, the AI ​​can also be an AI agent. Furthermore, when the processing of the above-mentioned parts is performed by AI, the processing can be performed partially or entirely by AI, but is not limited to this example. Furthermore, the processing performed by the AI ​​including the generation AI can be replaced by processing in the rule base, and the processing in the rule base can also be replaced by processing performed by the AI ​​including the generation AI.

[0495] Furthermore, the processing of the aforementioned data processing system 10 is performed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the smart device 14, but it can also be performed by both the specific processing unit 290 of the data processing device 12 and the control unit 46A of the smart device 14. Additionally, the specific processing unit 290 of the data processing device 12 acquires or collects information required for processing from the smart device 14 or external devices, and the smart device 14 acquires or collects information required for processing from the data processing device 12 or external devices.

[0496] For example, the collection unit is implemented by the control unit 46A of the smart device 14 or the specific processing unit 290 of the data processing device 12. For example, the acquisition unit uses the camera 42 or communication I / F 44 of the smart device 14 to acquire step data, which is then processed by the specific processing unit 290 of the data processing device 12. For example, the analysis unit is implemented by the specific processing unit 290 of the data processing device 12, which analyzes the data from the collection unit and the acquisition unit. For example, the generation unit is implemented by the specific processing unit 290 of the data processing device 12, which uses a generation AI to generate a menu. For example, the serving unit is implemented by the output device 40 of the smart device 14 or the specific processing unit 290 of the data processing device 12, which provides the generated menu to the user. The correspondence between each unit and the device or control unit is not limited to the above examples and various changes can be made.

[0497] In the above embodiments, examples of specific processing by the data processing device 12 are given, but the technology disclosed herein is not limited to this, and specific processing may also be performed by the smart device 14.

[0498] Second Implementation Method Figure 3 An example of the configuration of the data processing system 210 according to the second embodiment is shown.

[0499] like Figure 3 As shown, the data processing system 210 includes a data processing device 12 and smart glasses 214. A server can be cited as an example of the data processing device 12.

[0500] The data processing apparatus 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" as understood in this disclosure. The computer 22 includes a processor 28, RAM 30, and memory 32. The processor 28, RAM 30, and memory 32 are connected to a bus 34. Furthermore, the database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0501] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, RAM 48, and memory 50. The processor 46, RAM 48, and memory 50 are connected to a bus 52. Furthermore, the microphone 238, speaker 240, camera 42, and communication I / F 44 are also connected to the bus 52.

[0502] Microphone 238 receives instructions from user 20 by receiving sounds emitted by user 20. Microphone 238 captures sounds emitted by user 20 and converts the captured sounds into sound data, which is then output to processor 46. Speaker 240 outputs sound according to instructions from processor 46.

[0503] Camera 42 is a small digital camera equipped with an optical system such as a lens, aperture and shutter, and imaging elements such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, which captures images of the user 20's surroundings (e.g., the field of view defined by an angle equivalent to the field of vision of an average healthy person).

[0504] Communication I / F44 is connected to network 54. Communication I / F44 and 26 are responsible for the transmission and reception of various information between processor 46 and processor 28 via network 54. The transmission and reception of various information between processor 46 and processor 28 using communication I / F44 and 26 is performed in a secure state.

[0505] Figure 4 This illustrates an example of the main functions of the data processing device 12 and the smart glasses 214. For example... Figure 4 As shown, in the data processing device 12, specific processing is performed by the processor 28. The specific processing program 56 is stored in the memory 32.

[0506] The specific processing program 56 is an example of a "program" involved in the technology of this disclosure. The processor 28 reads the specific processing program 56 from the memory 32 and executes the read specific processing program 56 on the RAM 30. Specific processing is implemented by the processor 28 operating as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0507] The memory 32 stores a data generation model 58 and an emotion-specific model 59. The data generation model 58 and the emotion-specific model 59 are used by the specific processing unit 290. The specific processing unit 290 can use the emotion-specific model 59 to infer the user's emotions and perform specific processing based on the user's emotions. In the emotion inference function (emotion-specific function) using the emotion-specific model 59, various inferences and predictions related to the user's emotions are performed, including inferences and predictions of the user's emotions, but this is not limited to this example. Furthermore, emotion inference and prediction may also include, for example, emotion analysis (parsing).

[0508] In the smart glasses 214, the processor 46 performs reception and output processing. The memory 50 stores the reception and output program 60. The processor 46 reads the reception and output program 60 from the memory 50 and executes the read reception and output program 60 on the RAM 48. The reception and output processing is implemented by the processor 46 operating as a control unit 46A according to the reception and output program 60 executed on the RAM 48. Furthermore, the smart glasses 214 has the same data generation model and emotion-specific model as the data generation model 58 and the emotion-specific model 59, and these models can also be used to perform the same processing as the specific processing unit 290.

[0509] Next, the specific processing of the specific processing unit 290 of the data processing device 12 will be described. Each part of the system described below is implemented by the data processing device 12 and the smart glasses 214. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0510] Example 1 The process is the same as that of the specific process described in Embodiment 1 in the first embodiment above, so the description is omitted.

[0511] Application Example 1 The process is the same as that of the specific processing described in Application Example 1 in the first embodiment above, so the description is omitted.

[0512] Example 2 The process is the same as that of the specific process in Embodiment 2 described in the first embodiment above, so the description is omitted.

[0513] Application Example 2 The process is the same as that in the specific processing described in Application Example 2 of the first embodiment above, so the description is omitted.

[0514] The specific processing unit 290 sends the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A outputs the result of the specific processing to the speaker 240. The microphone 238 acquires sound input representing the user's input regarding the result of the specific processing. The control unit 46A sends the sound data representing the user's input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the sound data.

[0515] Data generation model 58 is a so-called generative AI (Artificial Intelligence). Examples of data generation models 58 include ChatGPT (registered trademark) (accessible via the internet (URL: https: / / openai.com / blog / chatgpt)). Data generation model 58 is obtained through deep learning on a neural network. Input to data generation model 58 are prompt words containing instructions, and inference data such as sound data representing sound, text data representing text, and image data representing images (e.g., still image data or animation data). Data generation model 58 infers from the input inference data based on the instructions represented by the prompt words and outputs the inference result in one or more data forms, such as sound data, text data, and image data. Data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or induction. The specific processing unit 290 performs the aforementioned specific processing while using data generation model 58. The data generation model 58 can also be a model finely tuned to output inference results from prompts that do not contain instructions. In this case, the data generation model 58 can output inference results based on prompts that do not contain instructions. The data processing apparatus 12, etc., includes various data generation models 58, including AI other than the generation AI. AI other than the generation AI can be, for example, linear regression, logistic regression, decision trees, random forests, support vector machines (SVM), k-means clustering, convolutional neural networks (CNN), recurrent neural networks (RNN), generative adversarial networks (GAN), or Naive Bayes, and can perform various processes, but is not limited to this example. Furthermore, the AI ​​can also be an AI agent. Furthermore, when the processing of the above-mentioned parts is performed by AI, the processing can be performed partially or entirely by AI, but is not limited to this example. Furthermore, the processing performed by the AI ​​including the generation AI can be replaced by processing in the rule base, and the processing in the rule base can also be replaced by processing performed by the AI ​​including the generation AI.

[0516] Furthermore, the processing of the aforementioned data processing system 10 is performed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the smart glasses 214, but it can also be performed by both the specific processing unit 290 of the data processing device 12 and the control unit 46A of the smart glasses 214. Additionally, the specific processing unit 290 of the data processing device 12 acquires or collects information required for processing from the smart glasses 214 or external devices, and the smart glasses 214 acquires or collects information required for processing from the data processing device 12 or external devices.

[0517] For example, the collection unit is implemented by the control unit 46A of the smart glasses 214 or the specific processing unit 290 of the data processing device 12. For example, the acquisition unit uses the camera 42 or communication I / F 44 of the smart glasses 214 to acquire step data, which is then processed by the specific processing unit 290 of the data processing device 12. For example, the analysis unit is implemented by the specific processing unit 290 of the data processing device 12, which analyzes the data from the collection unit and the acquisition unit. For example, the generation unit is implemented by the specific processing unit 290 of the data processing device 12, which uses a generation AI to generate a menu. For example, the serving unit is implemented by the speaker 240 of the smart glasses 214 or the specific processing unit 290 of the data processing device 12, which provides the generated menu to the user. The correspondence between each unit and the device or control unit is not limited to the above examples and various changes can be made.

[0518] In the above embodiments, examples of specific processing by the data processing device 12 are given, but the technology disclosed herein is not limited to this, and specific processing may also be performed by the smart glasses 214.

[0519] Third Implementation Method Figure 5 An example of the configuration of the data processing system 310 according to the third embodiment is shown.

[0520] like Figure 5 As shown, the data processing system 310 includes a data processing device 12 and a head-mounted terminal 314. A server can be cited as an example of the data processing device 12.

[0521] The data processing apparatus 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" as understood in this disclosure. The computer 22 includes a processor 28, RAM 30, and memory 32. The processor 28, RAM 30, and memory 32 are connected to a bus 34. Furthermore, the database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0522] The head-mounted terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, RAM 48, and memory 50. The processor 46, RAM 48, and memory 50 are connected to a bus 52. Furthermore, the microphone 238, speaker 240, camera 42, display 343, and communication I / F 44 are also connected to the bus 52.

[0523] Microphone 238 receives instructions from user 20 by receiving sounds emitted by user 20. Microphone 238 captures sounds emitted by user 20 and converts the captured sounds into sound data, which is then output to processor 46. Speaker 240 outputs sound according to instructions from processor 46.

[0524] Camera 42 is a small digital camera equipped with an optical system such as a lens, aperture and shutter, and imaging elements such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, which captures images of the user 20's surroundings (e.g., the field of view defined by an angle equivalent to the field of vision of an average healthy person).

[0525] Communication I / F44 is connected to network 54. Communication I / F44 and 26 are responsible for the transmission and reception of various information between processor 46 and processor 28 via network 54. The transmission and reception of various information between processor 46 and processor 28 using communication I / F44 and 26 is performed in a secure state.

[0526] Figure 6 This illustrates an example of the main functions of the data processing device 12 and the head-mounted terminal 314. For example... Figure 6 As shown, in the data processing device 12, specific processing is performed by the processor 28. The specific processing program 56 is stored in the memory 32.

[0527] The specific processing program 56 is an example of a "program" involved in the technology of this disclosure. The processor 28 reads the specific processing program 56 from the memory 32 and executes the read specific processing program 56 on the RAM 30. Specific processing is implemented by the processor 28 operating as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0528] The memory 32 stores the data generation model 58 and the emotion-specific model 59. The data generation model 58 and the emotion-specific model 59 are used by the specific processing unit 290.

[0529] In the head-mounted terminal 314, the processor 46 performs the acceptance / output processing. The memory 50 stores the acceptance / output program 60. The processor 46 reads the acceptance / output program 60 from the memory 50 and executes the read acceptance / output program 60 on the RAM 48. The acceptance / output processing is implemented by the processor 46 operating as a control unit 46A according to the acceptance / output program 60 executed on the RAM 48.

[0530] Next, the specific processing of the specific processing unit 290 of the data processing device 12 will be described. Each part of the system described below is implemented by the data processing device 12 and the head-mounted terminal 314. In the following description, the data processing device 12 will be referred to as the "server" and the head-mounted terminal 314 will be referred to as the "terminal".

[0531] Example 1 The process is the same as that of the specific process described in Embodiment 1 in the first embodiment above, so the description is omitted.

[0532] Application Example 1 The process is the same as that of the specific processing described in Application Example 1 in the first embodiment above, so the description is omitted.

[0533] Example 2 The process is the same as that of the specific process in Embodiment 2 described in the first embodiment above, so the description is omitted.

[0534] Application Example 2 The process is the same as that in the specific processing described in Application Example 2 of the first embodiment above, so the description is omitted.

[0535] The specific processing unit 290 sends the result of the specific processing to the head-mounted terminal 314. In the head-mounted terminal 314, the control unit 46A outputs the result of the specific processing to the speaker 240 and the display 343. The microphone 238 acquires sound input representing the user's input regarding the result of the specific processing. The control unit 46A sends the sound data representing the user's input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the sound data.

[0536] Data generation model 58 is a so-called generative AI (Artificial Intelligence). Examples of data generation models 58 include ChatGPT (registered trademark) (accessible via the internet (URL: https: / / openai.com / blog / chatgpt)). Data generation model 58 is obtained through deep learning on a neural network. Input to data generation model 58 includes prompt words containing instructions, and inference data such as sound data representing sound, text data representing text, and image data representing images (e.g., still image data or animation data). Data generation model 58 infers the input inference data based on the instructions represented by the prompt words and outputs the inference result in one or more data forms such as sound data, text data, and image data. Data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or induction. The specific processing unit 290 performs the aforementioned specific processing while using data generation model 58. The data generation model 58 can also be a model finely tuned to output inference results from prompts that do not contain instructions. In this case, the data generation model 58 can output inference results based on prompts that do not contain instructions. The data processing apparatus 12, etc., includes various data generation models 58, including AI other than the generation AI. AI other than the generation AI can be, for example, linear regression, logistic regression, decision trees, random forests, support vector machines (SVM), k-means clustering, convolutional neural networks (CNN), recurrent neural networks (RNN), generative adversarial networks (GAN), or Naive Bayes, and can perform various processes, but is not limited to this example. Furthermore, the AI ​​can also be an AI agent. Furthermore, when the processing of the above-mentioned parts is performed by AI, the processing can be performed partially or entirely by AI, but is not limited to this example. Furthermore, the processing performed by the AI ​​including the generation AI can be replaced by processing in the rule base, and the processing in the rule base can also be replaced by processing performed by the AI ​​including the generation AI.

[0537] Furthermore, the processing of the aforementioned data processing system 10 is performed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the head-mounted terminal 314, but it can also be performed by both the specific processing unit 290 of the data processing device 12 and the control unit 46A of the head-mounted terminal 314. Additionally, the specific processing unit 290 of the data processing device 12 acquires or collects information required for processing from the head-mounted terminal 314 or external devices, and the head-mounted terminal 314 acquires or collects information required for processing from the data processing device 12 or external devices.

[0538] For example, the collection unit is implemented by the control unit 46A of the head-mounted terminal 314 or the specific processing unit 290 of the data processing device 12. For example, the acquisition unit uses the camera 42 or communication I / F 44 of the head-mounted terminal 314 to acquire step data, which is then processed by the specific processing unit 290 of the data processing device 12. For example, the analysis unit is implemented by the specific processing unit 290 of the data processing device 12 to analyze the data from the collection unit and the acquisition unit. For example, the generation unit is implemented by the specific processing unit 290 of the data processing device 12 to generate a menu using a generation AI. For example, the serving unit is implemented by the speaker 240 and display 343 of the head-mounted terminal 314 or the specific processing unit 290 of the data processing device 12 to provide the generated menu to the user. The correspondence between each unit and the device or control unit is not limited to the above examples and various changes can be made.

[0539] In the above embodiments, examples of specific processing by the data processing device 12 are given, but the technology disclosed herein is not limited to this, and specific processing may also be performed by the head-mounted terminal 314.

[0540] Fourth Implementation Method Figure 7 An example of the configuration of the data processing system 410 according to the fourth embodiment is shown.

[0541] like Figure 7 As shown, the data processing system 410 includes a data processing device 12 and a robot 414. A server can be cited as an example of the data processing device 12.

[0542] The data processing apparatus 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" as understood in this disclosure. The computer 22 includes a processor 28, RAM 30, and memory 32. The processor 28, RAM 30, and memory 32 are connected to a bus 34. Furthermore, the database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0543] Robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a controlled object 443. The computer 36 includes a processor 46, RAM 48, and memory 50. The processor 46, RAM 48, and memory 50 are connected to a bus 52. Furthermore, the microphone 238, speaker 240, camera 42, controlled object 443, and communication I / F 44 are also connected to the bus 52.

[0544] Microphone 238 receives instructions from user 20 by receiving sounds emitted by user 20. Microphone 238 captures sounds emitted by user 20 and converts the captured sounds into sound data, which is then output to processor 46. Speaker 240 outputs sound according to instructions from processor 46.

[0545] Camera 42 is a small digital camera equipped with an optical system such as a lens, aperture and shutter, and imaging elements such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, to photograph the area around robot 414 (e.g., the field of view defined by a perspective equivalent to the field of vision of an average healthy person).

[0546] Communication I / F44 is connected to network 54. Communication I / F44 and 26 are responsible for the transmission and reception of various information between processor 46 and processor 28 via network 54. The transmission and reception of various information between processor 46 and processor 28 using communication I / F44 and 26 is performed in a secure state.

[0547] The controlled object 443 includes a display device, LEDs (light-emitting diodes) for the eyes, and motors for driving the arms, hands, and feet. The posture or movement of the robot 414 is controlled by controlling the motors in the arms, hands, and feet. Some of the robot 414's emotions can be expressed by controlling these motors. In addition, the robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[0548] Figure 8 This illustrates an example of the main functions of the data processing device 12 and the robot 414. For example... Figure 8 As shown, in the data processing device 12, specific processing is performed by the processor 28. The specific processing program 56 is stored in the memory 32.

[0549] The specific processing program 56 is an example of a "program" involved in the technology of this disclosure. The processor 28 reads the specific processing program 56 from the memory 32 and executes the read specific processing program 56 on the RAM 30. Specific processing is implemented by the processor 28 operating as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0550] The memory 32 stores the data generation model 58 and the emotion-specific model 59. The data generation model 58 and the emotion-specific model 59 are used by the specific processing unit 290.

[0551] In robot 414, the processor 46 performs the acceptance and output processing. The memory 50 stores the acceptance and output program 60. The processor 46 reads the acceptance and output program 60 from the memory 50 and executes the read acceptance and output program 60 on RAM 48. The acceptance and output processing is implemented by the processor 46 acting as the control unit 46A according to the acceptance and output program 60 executed on RAM 48.

[0552] Next, the specific processing of the specific processing unit 290 of the data processing device 12 will be described. Each part of the system described below is implemented by the data processing device 12 and the robot 414. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 will be referred to as the "terminal".

[0553] Example 1 The process is the same as that of the specific process described in Embodiment 1 in the first embodiment above, so the description is omitted.

[0554] Application Example 1 The process is the same as that of the specific processing described in Application Example 1 in the first embodiment above, so the description is omitted.

[0555] Example 2 The process is the same as that of the specific process in Embodiment 2 described in the first embodiment above, so the description is omitted.

[0556] Application Example 2 The process is the same as that in the specific processing described in Application Example 2 of the first embodiment above, so the description is omitted.

[0557] The specific processing unit 290 sends the result of the specific processing to the robot 414. In the robot 414, the control unit 46A outputs the result of the specific processing to the speaker 240 and the controlled object 443. The microphone 238 acquires sound input representing the result of the specific processing. The control unit 46A sends the sound data representing the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the sound data.

[0558] Data generation model 58 is a so-called generative AI (Artificial Intelligence). Examples of data generation models 58 include ChatGPT (registered trademark) (accessible via the internet (URL: https: / / openai.com / blog / chatgpt)). Data generation model 58 is obtained through deep learning on a neural network. Input to data generation model 58 are prompt words containing instructions, and inference data such as sound data representing sound, text data representing text, and image data representing images (e.g., still image data or animation data). Data generation model 58 infers from the input inference data based on the instructions represented by the prompt words and outputs the inference result in one or more data forms, such as sound data, text data, and image data. Data generation model 58 includes, for example, text generation AI, image generation AI, and multimodal generation AI. Here, inference refers to, for example, analysis, classification, prediction, and / or induction. The specific processing unit 290 performs the aforementioned specific processing while using data generation model 58. The data generation model 58 can also be a model finely tuned to output inference results from prompts that do not contain instructions. In this case, the data generation model 58 can output inference results based on prompts that do not contain instructions. The data processing apparatus 12, etc., includes various data generation models 58, including AI other than the generation AI. AI other than the generation AI can be, for example, linear regression, logistic regression, decision trees, random forests, support vector machines (SVM), k-means clustering, convolutional neural networks (CNN), recurrent neural networks (RNN), generative adversarial networks (GAN), or Naive Bayes, and can perform various processes, but is not limited to this example. Furthermore, the AI ​​can also be an AI agent. Furthermore, when the processing of the above-mentioned parts is performed by AI, the processing can be performed partially or entirely by AI, but is not limited to this example. Furthermore, the processing performed by the AI ​​including the generation AI can be replaced by processing in the rule base, and the processing in the rule base can also be replaced by processing performed by the AI ​​including the generation AI.

[0559] Furthermore, the processing of the aforementioned data processing system 10 is performed by the specific processing unit 290 of the data processing device 12 or the control unit 46A of the robot 414, but it can also be performed by both the specific processing unit 290 of the data processing device 12 and the control unit 46A of the robot 414. Additionally, the specific processing unit 290 of the data processing device 12 acquires or collects information required for processing from the robot 414 or external devices, and the robot 414 acquires or collects information required for processing from the data processing device 12 or external devices.

[0560] For example, the collection unit is implemented by the control unit 46A of the robot 414 or the specific processing unit 290 of the data processing device 12. For example, the acquisition unit uses the camera 42 or communication I / F 44 of the robot 414 to acquire step data, which is then processed by the specific processing unit 290 of the data processing device 12. For example, the analysis unit is implemented by the specific processing unit 290 of the data processing device 12, which analyzes the data from the collection unit and the acquisition unit. For example, the generation unit is implemented by the specific processing unit 290 of the data processing device 12, which uses a generation AI to generate a menu. For example, the serving unit is implemented by the speaker 240 of the robot 414 and the control object 443 or the specific processing unit 290 of the data processing device 12, which provides the generated menu to the user. The correspondence between each unit and the device or control unit is not limited to the above examples and various changes can be made.

[0561] In the above embodiments, examples of specific processing by the data processing device 12 are given, but the technology disclosed herein is not limited to this, and specific processing may also be performed by the robot 414.

[0562] Furthermore, the emotion-specific model 59, acting as an emotion engine, can determine a user's emotion based on a specific mapping. Specifically, the emotion-specific model 59 can determine a user's emotion based on an emotion graph that serves as a specific mapping (see [reference]). Figure 9 The emotion-specific model 59 can also determine the robot's emotion, and the specific processing unit 290 performs specific processing based on the robot's emotions.

[0563] Figure 9 This is a diagram representing an emotion map 400 that maps multiple emotions. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. The closer to the center of the concentric circles, the more primitive the emotion is. On the outer side of the concentric circles, emotions representing states or behaviors arising from mood are arranged. Emotions are concepts that include feelings and mental states. Emotions generated by reactions occurring in the brain are arranged roughly to the left of the concentric circles. Emotions derived from situational judgments are arranged roughly to the right of the concentric circles. Emotions generated by reactions occurring in the brain and derived from situational judgments are arranged roughly above and below the concentric circles. Furthermore, "pleasant" emotions are arranged above the concentric circles, and "unpleasant" emotions are arranged below them. Thus, in the emotion map 400, multiple emotions are mapped based on the structure that generates emotions, and emotions that are likely to occur simultaneously are mapped close to each other.

[0564] These emotions are distributed at the three o'clock position of the emotion map 400, typically fluctuating between peace and anxiety. In the right half of the emotion map 400, situational awareness dominates over internal sensation, thus resulting in an impression of calm.

[0565] The inner side of the emotion map 400 represents the inner state, while the outer side represents behavior. Therefore, the further outward you are from the emotion map 400, the more visible the emotion becomes (manifested in behavior).

[0566] Here, human emotions are based on various balances such as posture and blood sugar levels. When these balances deviate from an ideal state, it indicates an unpleasant state; when they approach the ideal state, it indicates a pleasant state. Emotions in robots, cars, motorcycles, etc., can also be created in the following way: based on various balances such as posture and remaining battery power, when these balances deviate from an ideal state, it indicates an unpleasant state; when they approach the ideal state, it indicates a pleasant state. Emotion maps can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on a Brain Physiological Signal Analysis System for Voice Emotion Recognition and Emotion, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). In the left half of the emotion map, emotions belonging to the sensory-dominated region, called "response," are arranged. Furthermore, in the right half of the emotion map, emotions belonging to the situational cognition-dominated region, called "situation," are arranged.

[0567] In the emotion map, two types of emotions that promote learning are defined. One is a negative emotion on the situational side, in the middle or peripheral region of "repentance" or "reflection." This occurs when the robot experiences negative emotions such as "I don't want to experience this feeling again" or "I don't want to be blamed again." The other is a positive emotion on the response side, near the "desire" region. This occurs when there are positive feelings such as "wanting more" or "wanting to know more."

[0568] The emotion-specific model 59 inputs user input into a pre-trained neural network to obtain emotion values ​​representing each emotion shown in the emotion map 400, thereby determining the user's emotion. This neural network is pre-trained based on multiple training data sets, which are combinations of user input and emotion values ​​representing each emotion shown in the emotion map 400. Furthermore, this neural network... Figure 10 As shown in the sentiment graph 900, it was trained in a way that sentiments that are configured close to each other have similar values. Figure 10 The text shows examples of emotions such as "peace of mind", "stability", and "reassurance" that have similar emotion values.

[0569] The above description focuses on the functions of the data processing device 12, but the system of this disclosure is not necessarily installed on a server. The system of this disclosure can also be installed as a general information processing system. This disclosure can also be installed, for example, as a software program running on a personal computer, an application running on a smartphone, etc. The method of this disclosure can also be provided to users in the form of SaaS (Software as a Service).

[0570] In the above embodiments, an example of a specific process being performed by a single computer 22 is given. However, the technology disclosed herein is not limited to this, and the specific process can also be distributed among multiple computers, including computer 22. For example, the data generation model 58 can be located on an external device of the data processing apparatus 12, where data is generated based on the input data.

[0571] In the above embodiments, examples of storing a specific processing program 56 in the memory 32 have been described, but the technology disclosed herein is not limited thereto. For example, the specific processing program 56 may also be stored in a portable computer-readable non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed into the computer 22 of the data processing device 12. The processor 28 executes specific processing according to the specific processing program 56.

[0572] Alternatively, a specific processing program 56 may be pre-stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 according to the requirements of the data processing device 12.

[0573] In addition, it is not necessary to store all the specific processing program 56 in the storage device such as the server connected to the data processing device 12 via the network 54 or in the memory 32; a portion of the specific processing program 56 may be stored in advance.

[0574] As hardware resources for performing specific processes, various processors, as shown below, can be used. For example, a CPU can be listed as a processor, which functions as a general-purpose processor that performs specific processes by executing software, i.e., a program. Furthermore, processors can be listed as special-purpose circuits such as FPGAs (Field-Programmable Gate Arrays), PLDs (Programmable Logic Devices), or ASICs (Application-Specific Integrated Circuits), which are processors with circuitry specifically designed to perform specific processes. Each processor has built-in or connected memory, and each processor executes specific processes using that memory.

[0575] The hardware resources for performing a specific process can consist of one of these various processors, or a combination of two or more processors of the same or different types (e.g., a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Alternatively, the hardware resources for performing a specific process can be a single processor.

[0576] As an example of a single processor, there are two approaches: First, a processor is composed of a combination of one or more CPUs and software, which functions as a hardware resource to perform a specific process; second, as represented by a SoC (System-on-a-chip), a processor is used to implement the functionality of the entire system, which includes multiple hardware resources for performing a specific process, using a single IC (Integrated Circuit) chip. In this way, the specific process is implemented by using one or more of the aforementioned processors as hardware resources.

[0577] Furthermore, the hardware architecture of these various processors, more specifically, can utilize circuits that combine semiconductor elements and other circuit components. Moreover, the specific process described above is just one example. Therefore, without departing from the main point, unnecessary steps can certainly be deleted, new steps added, or the processing order changed.

[0578] The descriptions and illustrations above are detailed explanations of a portion of the technology disclosed herein, and are merely one example of the technology disclosed herein. For example, the above descriptions of the structure, function, effect, and results are just one example of the structure, function, effect, and results of a portion of the technology disclosed herein. Therefore, without departing from the spirit of the technology disclosed herein, unnecessary parts may be deleted, new elements added, or replacements may be made to the descriptions and illustrations above. Furthermore, to avoid confusion and facilitate understanding of a portion of the technology disclosed herein, explanations of common technical knowledge that do not require special explanation under the premise of being able to implement the technology disclosed herein have been omitted from the descriptions and illustrations above.

[0579] All documents, patent applications and technical specifications set forth in this specification are incorporated herein by reference to the same extent that each document, patent application and technical specification is specifically and individually described therein and referenced by reference.

[0580] In addition, the following notes are provided in response to the above explanation.

[0581] Example 1 (Note 1) An information processing system, characterized in that it comprises: Devices for obtaining user information; A device for storing acquired user information as structured information and automatically generating prompt statements for input into a generative artificial intelligence model based on the structured information; A device for sending the prompt statement to a generative artificial intelligence model running on an external information processing device, and for the generative artificial intelligence model to generate mountain climbing plan information and acquire the mountain climbing plan information; A device for distinguishing and extracting the acquired mountaineering plan information and reconstructing it into multiple types of plan composition information, including mountaineering route information, time information, carried items information, and safety precautions information; A device for mapping the plan composition information to emergency contact information and facility information stored in the regional information storage unit, thereby editing it into plan document information; A device for automatically generating the plan document information into an electronic document of a predetermined format, and sending it to disaster prevention-related agencies and distributing it to user terminals via an external communication network; An apparatus for obtaining additional condition information or change request information from the user terminal, reflecting the additional condition information or change request information into the structured information, regenerating update prompt statements and re-inputting them into the generative artificial intelligence model, thereby obtaining updated mountain climbing plan information.

[0582] (Note 2) The information processing system according to Appendix 1 is characterized in that, It also includes means for controlling the display of the plan composition information on the user terminal in a differentiated manner, and storing at least a portion of the plan composition information in a storage unit within the terminal so that it can be referenced when communication is unavailable.

[0583] (Note 3) The information processing system according to Appendix 1 is characterized in that, It also includes a device for collecting location information obtained from the user terminal at predetermined time intervals, generating movement trajectory information based on the mountain climbing route information and the location information, and storing the movement trajectory information in association with the plan document information.

[0584] Application Example 1 (Note 1) An information processing system, characterized in that it comprises: A device for receiving user demand information, extracting the conditions contained in the demand information, and generating prompt statements to instruct a generative artificial intelligence model to generate plan information; A device for inputting the prompt statement into the generative artificial intelligence model and obtaining selection information containing evaluation criteria and selection rules corresponding to the conditions based on the output of the generative artificial intelligence model; An apparatus for performing retrieval and evaluation processing on candidate facility information and candidate project data information stored in an information storage device based on the selection information, thereby extracting candidate facilities and candidate projects that meet the conditions and generating planning information containing the candidate facilities and candidate projects; An apparatus for sending the generated plan information to a terminal device, updating the conditions based on a user change request obtained from the terminal device, and repeatedly executing the generation of the prompt statement, the processing of the generative artificial intelligence model, the retrieval processing, and the evaluation processing according to the updated conditions to regenerate the plan information; A means for automatically generating document information for reporting based on the regenerated plan information and sending the document information to an external device.

[0585] (Note 2) The information processing system according to Appendix 1 is characterized in that, The apparatus for performing retrieval and evaluation processing based on the selection information is configured to perform weighted operations on the candidate facility information and the candidate project data information based on a plurality of attribute values ​​according to the evaluation criteria contained in the selection information, thereby calculating a comprehensive evaluation value for multiple candidate combinations respectively, and sorting the candidate combinations according to the comprehensive evaluation value to generate the plan information.

[0586] (Note 3) The information processing system according to Appendix 1 is characterized in that, The system is configured to obtain user information containing location information from the terminal device, filter the candidate facility information based on the location information using spatial conditions, and after executing actions according to the plan information, generate movement path information based on the history of the location information and record the movement path information in conjunction with the document information used for reporting.

[0587] Example 2 (Note 1) An information processing system, characterized in that it comprises: A unit for receiving input information from an information processing device via a computing device and storing the input information as basic data in an information storage device; A unit for generating prompt statements containing activity plan information based on the basic data using a computing device and inputting the prompt statements into a generative artificial intelligence model to obtain generated result data including activity plan content; A unit for performing format conversion processing on the generated result data through a computing device to generate document data for submission to an external organization and for automatically sending the document data to the external organization through a communication device; A unit for receiving location information data from a mobile terminal via a communication device and storing the location information data in an information storage device by associating it with the basic data via a computing device; A unit for retrieving emergency contact information and refuge location information from a geographic information database using a computing device based on the location information data, and providing the emergency contact information and refuge location information to the mobile terminal via a communication device; A unit for periodically receiving location history data from the mobile terminal via a communication device, performing route reconstruction and statistical processing on the location history data via a computing device to generate activity record data, and providing the activity record data to the mobile terminal via the communication device; A unit for determining the nearest refuge location from a refuge location information database based on the latest location information data using a computing device and prompting the mobile terminal with the nearest refuge location via a communication device when an emergency notification is detected from the mobile terminal.

[0588] (Note 2) The information processing system according to Appendix 1 is characterized in that, When generating the prompt statement, the computing device automatically determines the composition and expression of the items contained in the prompt statement based on the type and missing information of the input information, so that the generative artificial intelligence model outputs generated result data that conforms to the predetermined document format.

[0589] (Note 3) The information processing system according to Appendix 1 is characterized in that, The computing device calculates the movement distance, movement time, and altitude change from the location history data, and associates the calculation results with the activity plan content generated by the generative artificial intelligence model to generate activity record data containing information comparing the plan and the actual situation.

[0590] Application Example 2 (Note 1) An information processing system, characterized in that it comprises: A device for receiving input information from a user related to a behavior plan and generating text containing prompt statements based on the input information to instruct a generative artificial intelligence model to generate the behavior plan; A device for obtaining textual data about the action plan from the generative artificial intelligence model and automatically embedding the textual data into a predetermined format template to generate an electronic document of the plan for submission to an emergency response agency; An information processing device for automatically sending the electronic document to a predetermined emergency response agency via a communication network and notifying the user information processing terminal of the sending result; An apparatus for obtaining location information from the user information processing terminal at predetermined time intervals, storing the location information in association with the behavior plan, and generating behavior path recording data based on the stored location information; An apparatus for retrieving nearby support organizations relative to the user's current location based on the location information, and providing the contact information and location information of the support organizations to the user information processing terminal; An apparatus for acquiring information about a user's emotional state and, based on the emotional state, modifying the prompts or the content of the behavior plan input to the generative artificial intelligence model to dynamically adjust the behavior plan; An apparatus for integrating the location information and the emotional state information in chronological order to generate behavioral process record data, and using the record data as input conditions to provide prompts to the generative artificial intelligence model in order to automatically generate narrative content about the behavioral process.

[0591] (Note 2) The information processing system according to Appendix 1 is characterized in that it further includes: An apparatus for generating video content that integrates image information, audio information, and location information display based on the narrative content about the behavioral process and the location information, and distributing the video content to the user information processing terminal.

[0592] (Note 3) The information processing system according to Appendix 1 is characterized in that it further includes: A device for automatically generating an emergency notification message containing the user's current location and information related to the planning document when a predetermined dangerous situation is presumed, based on information about the user's emotional state and the location information, and sending the emergency notification message to the emergency response agency.

Claims

1. An information processing system, characterized in that, include: processor; The processor is configured to receive user input and generate prompt text to instruct a generative artificial intelligence model to generate a mountain climbing plan. The processor is configured to input the generated prompt text into the generative artificial intelligence model to generate a mountain climbing plan; The processor is configured to automatically generate a plan based on the generated mountaineering plan and submit the plan to the emergency agency.

2. The information processing system according to claim 1, characterized in that, The processor is configured to obtain emergency contact information and facility information for the corresponding area based on the mountaineering plan, and to provide the information to the user.

3. The information processing system according to claim 1, characterized in that, The processor is configured to periodically acquire location information from the user terminal and generate a movement route record based on the location information after the mountain climb ends.

Images