Explanation task real-time planning method and device, equipment and storage medium

Abstract

This application discloses a method, apparatus, device, and storage medium for real-time planning of explanation tasks, relating to the field of robotics. The method includes: acquiring the instruction intent of the explanation task command; determining the explanation task planning result based on the instruction intent, a preset knowledge graph, and a preset task planning algorithm; and during the explanation process according to the explanation task planning result, using a preset acquisition device to scan and acquire current environmental information in real time, which is then combined with a preset path update algorithm to determine whether to update the explanation route. This method determines the explanation task planning result based on the task command using a preset task planning algorithm and a preset knowledge graph, and scans the current environmental information in real time during the explanation process. This enables the determination of whether to update the explanation route based on the current environmental information, avoiding obstacles in the explanation route from interrupting the explanation. This solves the problems of fixed explanation robot modes and poor flexibility, and optimizes the user experience during the explanation process.

Description

Technical Field

[0001] This application relates to the field of robotics technology, and in particular to a method, apparatus, device, and storage medium for real-time planning of explanatory tasks. Background Technology

[0002] Explanatory robots are intelligent robots with knowledge service and interactive capabilities. They are mainly used in education, science popularization, and service fields, providing information through voice interaction, visual display, and other methods.

[0003] Current explanation robots typically follow fixed routes and processes to provide explanations, and use pre-set programs and data to enable question-and-answer sessions and interactions with people.

[0004] However, when explaining things using a fixed route and process, the robot cannot adjust its route autonomously. This shows that the current explanation robot has a fixed pattern and poor flexibility, resulting in a poor user experience. Summary of the Invention

[0005] The main purpose of this application is to propose a method, apparatus, equipment and storage medium for real-time planning of explanation tasks, which aims to solve the problems of fixed modes and poor flexibility of explanation robots.

[0006] In a first aspect, the present invention provides a real-time planning method for narration tasks, applied to a narration robot, comprising: Receive a narration task instruction, and analyze the instruction intent corresponding to the narration task instruction based on the narration task instruction and a preset natural language processing algorithm; Based on the instruction intent, the preset knowledge graph, and the preset task planning algorithm, the explanation task planning result is determined. The preset task planning algorithm is obtained by training on the task planning sample dataset, which includes: sample instructions and corresponding sample task planning results. The explanation task planning result includes: explanation route, explanation content, and safety requirements. According to the preset identification rules, the target visitor is identified among the visitors. After confirming that the target visitor is in place, the explanation is carried out according to the explanation task planning results, and the current environmental information is obtained in real time by scanning through the preset collection device. Based on the current environment information and the preset path update algorithm, determine whether to update the explanation route.

[0007] In an optional implementation, before receiving the explanation task instruction and analyzing and obtaining the instruction intent corresponding to the explanation task instruction based on the explanation task instruction and a preset natural language processing algorithm, the method further includes: Obtain the knowledge document corresponding to the explanation task, and extract the explanation knowledge corresponding to the explanation task according to the preset recognition algorithm and the knowledge document, wherein the knowledge document includes: explanation script document and professional knowledge document; Based on the explained knowledge and the preset knowledge graph generation algorithm, the preset knowledge graph is generated.

[0008] In an optional implementation, before determining the target visitor, the method further includes: Obtain visitor information, which includes: visitor facial image, visitor name, and visitor identity background; The step of identifying and obtaining target visitors from the current visitors according to preset identification rules includes: Facial images of multiple visitors were captured by scanning with a pre-set data acquisition device. Based on the preset recognition algorithm, the facial images captured, and the visitor's facial images, the visitor information corresponding to each of the visitors entering the venue is determined; The target visitors are determined based on the visitor information and preset visitor priority rules.

[0009] In an optional implementation, the explanation route includes: multiple explanation points; the current environment information includes: obstacle information of the explanation route; the preset path update algorithm includes: route generation algorithm and route optimization algorithm; The step of determining whether to update the explanation route based on the current environment information and the preset path update algorithm includes: Based on the obstacle information of the explanation route, determine the obstacle situation of the explanation route from the current explanation point to the next explanation point; If there is an obstacle between the current explanation point and the next explanation point, the explanation route is updated, and multiple alternative routes between the current explanation point and the next explanation point are generated by the route generation algorithm. Based on multiple alternative routes and the route selection algorithm, an updated route is determined from the current explanation point to the next explanation point, wherein the updated route is the alternative route with the shortest travel time and / or the fewest obstacles.

[0010] In an optional implementation, the current environment information further includes: the visitor's visit behavior, which includes: the target visitor's body posture and the duration of stay at the current explanation point; After completing the explanation according to the updated explanation route, the method further includes: Receive the question instruction from the target visitor, and analyze and obtain the question keywords corresponding to the question instruction based on the question instruction and the preset natural language processing algorithm; Based on the question keywords, the preset knowledge graph, the target visitor's visit behavior, question and answer history, preset knowledge matching algorithm, and preset natural language processing algorithm, the answer content is generated, and the target visitor is answered based on the answer content.

[0011] In an optional implementation, the current environmental information further includes: the real-time location of the target visitor; the safety requirements include: collision restrictions, danger zone restrictions, and touch restrictions; After acquiring the current environmental information in real time through a preset acquisition device, the method further includes: Based on the explained route obstacle information and collision restrictions, the system performs real-time avoidance of moving obstacles during movement. Based on the target visitor's real-time location and the danger zone restrictions, the target visitor's body posture and the touch restrictions, danger zone alerts and touch alerts are issued to the target visitor.

[0012] In an optional implementation, the explanation robot is a humanoid explanation robot, and the explanation task planning result further includes: explaining body postures; The process of identifying target visitors and, after they are in place, conducting the explanation according to the planned explanation task includes: Explain the body postures as described above.

[0013] Secondly, the present invention provides a real-time planning device for instructional tasks, comprising: The receiving module is used to receive the explanation task instruction and analyze the instruction and the preset natural language processing algorithm to obtain the instruction intent corresponding to the explanation task instruction. The planning module is used to determine the explanation task planning result based on the instruction intent, the preset knowledge graph and the preset task planning algorithm. The preset task planning algorithm is obtained by training from the task planning sample dataset. The task planning sample dataset includes: sample instructions and corresponding sample task planning results. The explanation task planning result includes: explanation route, explanation content and safety requirements. The explanation module is used to identify the target visitor among the current visitors according to the preset recognition rules. After confirming that the target visitor is in place, the explanation is carried out according to the explanation task planning results, and the current environmental information is obtained in real time by scanning through the preset acquisition device. The update module is used to determine whether to update the explanation route based on the current environment information and the preset path update algorithm.

[0014] Thirdly, the present invention provides an electronic device, comprising: a processor, a storage medium, and a bus, wherein the storage medium stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the storage medium via the bus, and the processor executes the machine-readable instructions to perform the steps of any of the methods described in the foregoing embodiments.

[0015] Fourthly, the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the method as described in any of the foregoing embodiments.

[0016] The beneficial effects of this application are: The real-time planning method for explanation tasks provided in this application includes: receiving an explanation task instruction; analyzing and obtaining the instruction intent corresponding to the explanation task instruction based on the explanation task instruction and a preset natural language processing algorithm; determining the explanation task planning result based on the instruction intent, a preset knowledge graph, and a preset task planning algorithm, wherein the preset task planning algorithm is trained from a task planning sample dataset, the task planning sample dataset includes: sample instructions and corresponding sample task planning results, and the explanation task planning result includes: explanation route, explanation content, and safety requirements; identifying and obtaining a target visitor among the current visitors according to a preset recognition rule; after confirming that the target visitor is in place, conducting the explanation according to the explanation task planning result, and obtaining current environmental information in real time through a preset acquisition device; and determining whether to update the explanation route based on the current environmental information and the preset path update algorithm. In this embodiment, after analyzing the received task instructions to obtain the corresponding instruction intent, the system uses a preset task planning algorithm based on a preset knowledge graph to determine the explanation task planning result corresponding to the instruction intent. During the explanation process according to the explanation task planning result, the system scans the current environment information in real time to judge obstacles and other situations on the explanation route. This enables the system to determine whether to update the explanation route based on the current environment information, so as to avoid timely adjustments when the current explanation route is affected by the environment. This improves the flexibility of the explanation robot in the face of complex environments and optimizes the user experience during the explanation process. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0018] Figure 1This is a schematic diagram of a real-time planning method for explaining tasks provided in an embodiment of this application; Figure 2 This is a flowchart illustrating a real-time planning method for a tutorial task, provided in another embodiment of this application. Figure 3 A schematic diagram of a real-time planning method for explanatory tasks provided in another embodiment of this application; Figure 4 A schematic diagram of the structure of a real-time planning device for explaining tasks provided in this application embodiment; Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0020] Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0021] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. The terms "comprising," "including," or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0022] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0023] Current explanation robots typically follow fixed routes and procedures, and interact with people through pre-set programs and data. However, when providing explanations along fixed routes and procedures, if obstacles such as people or objects appear, the robot can only provide voice prompts to guide people to avoid them or to move the obstacles. It cannot autonomously adjust its route, and the explanation will be interrupted when the voice prompt is issued. The explanation can only continue after external intervention or removal of the obstacle. Therefore, current explanation robots are not very flexible in complex environments, which may result in a poor user experience during the explanation process.

[0024] To address the aforementioned issues, the main objective of this application is to propose a real-time planning method for explanation tasks, thereby optimizing the user experience during the explanation process.

[0025] Figure 1 This is a schematic flowchart of a real-time planning method for a narration task provided in an embodiment of this application. The executing entity of this method can be, for example, a controller or processor of a narration robot, or other device with computing capabilities. The narration robot can include, for example, an agent (intelligent agent), but is not limited thereto. Figure 1 As shown, the method includes: S101. Receive the explanation task instruction, and analyze and obtain the instruction intent corresponding to the explanation task instruction based on the explanation task instruction and the preset natural language processing algorithm.

[0026] For example, the above-mentioned explanation task instructions may be, but are not limited to, instructions in the form of voice or text. The explanation task instructions may include, for example, the explanation format, the explanation focus, the explanation duration, the scope of related knowledge expansion, etc., and the above-mentioned explanation task instructions may include specific numerical requirements, such as "2-3 minutes for each explanation point". The explanation task instructions may also include vague instructions, such as "The explanation of each point should focus on the function of the exhibit at that point, and may also expand on the explanation of some special technologies used to realize the function of the exhibit", without any specific limitations here.

[0027] The aforementioned preset natural language processing algorithm may refer to an algorithm based on NLP (Natural Language Processing) technology. The above-mentioned analysis of the instruction intent corresponding to the instruction task based on the instruction task and the preset natural language processing algorithm may include extracting keywords from the instruction task through NLP technology, but is not limited thereto. It is understood that the specific algorithm selected for the above-mentioned preset natural language processing algorithm can be determined according to the actual situation, and no specific restrictions are imposed here.

[0028] S102. Based on the above instruction intent, preset knowledge graph and preset task planning algorithm, determine the explanation task planning result.

[0029] The aforementioned preset task planning algorithm is trained using a task planning sample dataset, which includes sample instructions and corresponding sample task planning results. The aforementioned explanation task planning results include explanation routes, explanation content, and safety requirements.

[0030] For example, the aforementioned preset task planning algorithm may include multiple agents that learn from the task planning sample dataset to generate a task model, such as an environmental perception model, a motion control model, or an interactive task model. There are no specific limitations here. The sample task planning results corresponding to the sample instructions may be manually set and labeled, but this is not a limitation. The preset task planning algorithm is obtained by training the task planning sample dataset. The training logic may be, for example, the algorithm logic that enables the explanation robot to form a sample instruction (or task instruction) with certain characteristics based on the task planning sample dataset and determine the corresponding explanation task planning result.

[0031] In addition, before determining the explanation task planning result based on the above-mentioned instruction intent, preset knowledge graph and preset task planning algorithm, the explanation robot can also determine the explanation route based on the preset explanation site map.

[0032] S103. According to the preset identification rules, identify the target visitor among the visitors. After confirming that the target visitor is in place, conduct the explanation according to the above explanation task planning results, and obtain the current environmental information in real time through the preset collection device.

[0033] For example, the aforementioned preset identification rules may include rules that prioritize identifying visiting guests and professionals as target visitors based on their visitor identity, or rules that prioritize identifying group visitors as target visitors based on the number of visitors. However, these rules are not limited to these specific rules and can be adjusted and determined according to actual needs.

[0034] The above-mentioned explanation is carried out according to the results of the explanation task planning. For example, it can refer to guiding visitors to visit in sequence according to the above explanation route, and playing corresponding explanation content in a timely manner according to the visit location. The explanation content can be extracted from the above-mentioned preset knowledge graph according to the instruction intention, and the text content can be converted into speech content through a TTS (TextToSpeech) engine to realize the playback of the explanation content. However, the specific explanation content is not limited to the above-mentioned examples.

[0035] The aforementioned acquisition of current environmental information through real-time scanning using preset data collection devices, such as one or more devices including LiDAR, cameras, and infrared scanners installed on the head of the guide robot, can be used to enable the guide robot to determine the status of surrounding visitors, the location of obstacles, etc., so as to flexibly respond to visitors' questions based on their status, and provide prompts based on safety requirements when visitors enter dangerous areas or touch untouchable exhibits, and / or flexibly adjust the explanation route according to the location of obstacles, etc., without any specific limitations.

[0036] In addition to the preset map, the aforementioned site map can also be constructed by a guided tour robot that enters the site in advance and scans and photographs the site using one or more of the aforementioned LiDAR, camera, and infrared scanner. This map can be a two-dimensional map or a three-dimensional map, which can be selected and determined as needed, and there are no restrictions here.

[0037] S104. Based on the current environment information and the preset path update algorithm, determine whether to update the above-mentioned explanation route.

[0038] For example, the aforementioned preset path update algorithm may include rules for determining whether the guided tour route needs to be updated, and may also have the function of regenerating or selecting a guided tour route. When determining whether to update the guided tour route, the rules can be based on the location, size, and movement status of obstacles in the current environment, but are not limited to these limitations; they can be flexibly adjusted according to the actual situation. If it is determined that the guided tour route needs to be updated, the updated route may be a route with shorter travel time, an easier route to traverse, etc., and there are no specific restrictions here. For example, during peak holiday periods, if visitors linger at a certain point on the current guided tour route for more than a certain period, the plan can be to skip that point first, and then, based on the synchronized routes and dwell times of the visitors ahead, plan the timing of passing that point later. Alternatively, if there is a large obstacle ahead that affects the current tour group's positioning or movement, a route can be planned to bypass the obstacle.

[0039] The real-time planning method for explanation tasks provided in this application includes: receiving an explanation task instruction; analyzing and obtaining the instruction intent corresponding to the explanation task instruction based on the explanation task instruction and a preset natural language processing algorithm; determining the explanation task planning result based on the instruction intent, a preset knowledge graph, and a preset task planning algorithm, wherein the preset task planning algorithm is trained on a task planning sample dataset, the task planning sample dataset includes: sample instructions and corresponding sample task planning results, and the explanation task planning result includes: explanation route, explanation content, and safety requirements; identifying target visitors among the current visitors according to preset recognition rules; after confirming that the target visitors are in place, conducting the explanation according to the explanation task planning result, and acquiring current environmental information in real time through a preset acquisition device; and determining whether to update the explanation route based on the current environmental information and the preset path update algorithm. In this embodiment, after analyzing the received task instructions to obtain the corresponding instruction intent, a preset task planning algorithm is used to determine the explanation task planning result corresponding to the instruction intent based on a preset knowledge graph. During the explanation process according to the explanation task planning result, the current environmental information is scanned in real time to judge obstacles and other situations on the explanation route. This enables the robot to determine whether to update the explanation route based on the current environmental information, so as to avoid obstacles in the explanation route from blocking the robot's progress. When the current explanation route is affected by the environment, it can be adjusted in time to avoid hindering the explanation. This improves the robot's operational autonomy and question-and-answer interaction flexibility in complex environments and optimizes the user experience during the explanation process.

[0040] Optionally, in the above Figure 1 Based on the embodiments, before receiving the explanation task instruction and analyzing and obtaining the instruction intent corresponding to the explanation task instruction according to the explanation task instruction and a preset natural language processing algorithm, the above method may further include: Obtain the knowledge document corresponding to the explanation task, and extract the explanation knowledge corresponding to the explanation task based on the preset recognition algorithm and the above knowledge document. The above knowledge document includes: explanation script document and professional knowledge document.

[0041] Based on the knowledge explained above and the preset knowledge graph generation algorithm, the preset knowledge graph is generated.

[0042] For example, the explanatory knowledge in the aforementioned knowledge document may be in the form of images and text, but is not limited thereto. The extraction of the explanatory knowledge corresponding to the aforementioned explanatory task may refer to extracting the explanatory knowledge in the form of images and text in the aforementioned knowledge document through OCR (Optical Character Recognition) technology, but other technologies may also be used to extract the explanatory knowledge corresponding to the explanatory task.

[0043] The aforementioned preset knowledge graph generation algorithms may include, for example, knowledge extraction algorithms, knowledge fusion algorithms, knowledge reasoning algorithms, and typical application algorithms. These algorithms then perform entity recognition, relation extraction, and attribute extraction on the knowledge being explained, followed by entity disambiguation, knowledge merging, and reasoning completion, ultimately storing the information to form a knowledge graph. However, the specific preset knowledge graph generation process is not limited to the examples described above. For instance, in a history museum explaining a historical artifact, the algorithm can identify its shape, color, and other relevant information from a photograph, combining this with input historical background, artifact background, and related historical figures. This information, combined with the preset knowledge graph generation algorithm, generates the aforementioned preset knowledge graph.

[0044] For example, in an industrial tourism exhibition hall of an automaker, the explanation of a certain car product can be achieved by recognizing relevant knowledge such as the car's exterior design, performance components, etc., through multi-angle photos of the car product. Then, combined with the input vehicle size parameters, body structure, body material, power configuration, intelligent configuration, etc., the aforementioned preset knowledge graph is generated through a preset knowledge graph generation algorithm.

[0045] Figure 2 This is a flowchart illustrating a real-time planning method for an explanatory task, as provided in another embodiment of this application. Figure 2 As shown, optionally, in the above Figure 1 Based on the embodiments, before determining the target visitors, the above method may further include: Obtain visitor information, including: visitor facial image, visitor name, and visitor identity background.

[0046] Specifically, visitors or organizers can input visitor information through the UI interactive interface, binding a specific visit and the corresponding visitor information.

[0047] The process of identifying target visitors from among the current visitors based on preset identification rules may include: S201. Facial images of multiple visitors are captured by scanning with a preset acquisition device.

[0048] For example, the above-mentioned scanning and acquisition of facial images of multiple visitors entering the venue by a preset acquisition device can be achieved by devices including but not limited to the aforementioned camera. Correspondingly, the aforementioned explanation robot can wait at the entrance before visitors begin to enter, so as to scan each visitor before identifying and acquiring the target visitor, but is not limited to this.

[0049] S202. Based on the preset recognition algorithm, the facial images captured above, and the facial images of the visitors above, determine the visitor information corresponding to each of the aforementioned visitors.

[0050] For example, the aforementioned preset recognition algorithm may be a face recognition algorithm. This preset recognition algorithm may extract and compare the facial features in the facial capture image and the visitor's facial image, respectively, so as to determine the visitor information corresponding to the facial capture image through the visitor's facial image matched with the facial capture image.

[0051] S203. Based on the above visitor information and the preset visitor priority rules, determine the above target visitors.

[0052] For example, the aforementioned preset visitor priority rules may include rules that prioritize visiting guests, professionals, tour organizers, tour leaders, etc., based on the visitor's identity and background, or rules that prioritize group visitors based on the number of visitors. After determining the target visitors, the guide robot can greet the visitors and interact with them according to the visitor's address, so that the visitors feel a sense of familiarity during the interaction with the guide robot.

[0053] Figure 3 For a flowchart illustrating the real-time planning method for explanatory tasks provided in another embodiment of this application, please refer to... Figure 3 Optionally, in the above Figure 1 Based on the embodiments, the aforementioned explanation route may include: multiple explanation points. The aforementioned current environment information may include: obstacle information along the explanation route. The aforementioned preset path update algorithm may include: a route generation algorithm and a route optimization algorithm.

[0054] The determination of whether to update the explained route based on the current environment information and the preset path update algorithm may include: S301. Based on the above-mentioned obstacle information of the explanation route, determine the obstacle situation of the explanation route from the current explanation point to the next explanation point.

[0055] For example, obstacles along the tour route from the current tour point to the next tour point may include the position, height, width, and duration of visitors, trolleys transporting or moving exhibits, or other people or objects that could obstruct the tour robot's movement along the tour route. No specific restrictions are imposed here.

[0056] S302. If there is an obstacle between the current explanation point and the next explanation point, the explanation route is updated, and multiple alternative routes between the current explanation point and the next explanation point are generated through the route generation algorithm.

[0057] For example, the existence of an obstacle between the current explanation point and the next explanation point should be understood as the existence of an obstacle on the explanation route between the current explanation point and the next explanation point that is sufficient to prevent the explanation robot from moving along the explanation route, and the duration of the obstacle should be greater than the preset duration, such as more than 10 seconds. For obstacles that are temporarily passed on the explanation route from the current explanation point to the next explanation point or that can be avoided by adjusting the angle and posture of the explanation robot or crossing over them without updating the explanation route, it can be considered that there is no obstacle between the current explanation point and the next explanation point. Of course, the above is only a possible example, and the actual criteria for judging whether there is an obstacle between the current explanation point and the next explanation point can be adjusted and determined according to the actual situation, without any restrictions here.

[0058] The route generation algorithm described above can generate multiple different routes based on different permutations and combinations of the remaining multiple explanation points. Since the next explanation point can be different for different routes, the route generation algorithm described above can generate multiple alternative routes from the current explanation point to the next explanation point.

[0059] S303. Based on the multiple alternative routes and the route selection algorithm, determine the updated route from the current explanation point to the next explanation point, wherein the updated route is the alternative route with the shortest time and / or the fewest obstacles.

[0060] Based on the above-mentioned multiple alternative routes and the above-mentioned route selection algorithm, the updated route from the current explanation point to the next explanation point is determined. The updated route is the alternative route with the shortest time and / or the fewest obstacles. For example, it can refer to the alternative route from the current explanation point to the next explanation point that corresponds to the route with the shortest total time and / or the fewest total obstacles among the various routes generated based on the remaining multiple explanation points as the updated route from the current explanation point to the next explanation point.

[0061] For example, the route selection algorithm described above may include a TSP (Travelling Salesman Problem) model and a particle swarm optimization model. The objective function of this route selection algorithm can be expressed as follows: F(x) = minz1 + βminz2, Wherein, F(x) represents the objective function of the route selection algorithm, z1 represents the minimum total distance, z2 represents the minimum total steering adjustment, and β is a weighting coefficient used to convert the steering adjustment into distance, which can be adjusted and determined according to the actual situation. Therefore: , , Among them, the above The distance from explanation point i to explanation point j is the above. The turning adjustment amount is from explanation point i to explanation point j, and n is the total number of remaining explanation points.

[0062] The above Let be a 0-1 variable. A value of 1 indicates moving from explanation point i to explanation point j, and a value of 0 indicates not moving from explanation point i to explanation point j. Then: , The above v is a set that includes all the teaching points, that is, there are n teaching points in the set v.

[0063] The objective function of the above route selection algorithm is also subject to the following restrictions: , The above restrictions apply to each explanation point in set b. , indicating from There is only one starting path (that is, each explanation point can only start once, and it is impossible to start from the same explanation point multiple times).

[0064] , The above restrictions apply to each explanation point in set b. , indicating arrival There is only one path (that is, each explanation point is reached only once, and it is impossible to reach the same explanation point multiple times).

[0065] , The above constraints are sub-loop elimination constraints, used to ensure that the various routes generated based on the remaining multiple explanation points are all routes that pass through all the remaining explanation points (i.e., the route must pass through every explanation point).

[0066] It is understood that the above algorithm is only one possible example, and the actual route selection algorithm is not limited to the algorithm in the example above.

[0067] Optionally, based on the above embodiments, the current environmental information may further include: the visitor's behavior, which includes: the target visitor's body posture and the duration of stay at the current explanation point.

[0068] After completing the explanation according to the updated explanation route, the above method also includes: Upon receiving the question instruction from the target visitor, the system analyzes and obtains the corresponding question keywords based on the question instruction and the preset natural language processing algorithm.

[0069] Based on the aforementioned question keywords, the aforementioned preset knowledge graph, the aforementioned visitor's visit behavior, question and answer history, preset knowledge matching algorithm, and preset natural language processing algorithm, the system generates answer content and responds to the aforementioned target visitor based on the aforementioned answer content.

[0070] For example, the aforementioned questioning instructions from the target visitor may be, but are not limited to, voice instructions. When receiving the voice from the target visitor, the received voice can be analyzed to determine whether it is a questioning instruction. For instance, for the voice content "What material is this exhibit made of?", the corresponding questioning keywords obtained through a preset natural language processing algorithm may include "exhibit" and "material," thus confirming that the voice content is a questioning instruction. For the voice content "Shall we go to that noodle shop on the second floor for lunch later?", the corresponding questioning keywords obtained through a preset natural language processing algorithm may include "noodle shop" and "lunch," thus confirming that the voice content is not a questioning instruction. Of course, the specific method for receiving the aforementioned questioning instructions from the target visitor and obtaining the corresponding questioning keywords based on the questioning instructions and the preset natural language processing algorithm can be adjusted and determined according to the actual situation, and is not limited to the examples above.

[0071] For example, the above-mentioned question keywords, the above-mentioned preset knowledge graph, the above-mentioned visitor's visit behavior, question and answer history, preset knowledge matching algorithm, and preset natural language processing algorithm are used to generate answer content. The above-mentioned visitor's visit behavior can refer to, for example, the target visitor's body posture and the length of time spent at the current explanation point. By analyzing the visitor's body posture and the length of time spent at the current explanation point, it is possible to analyze the target visitor's level of interest in the current exhibit and explanation content. For example, if the visitor gets close to the exhibit and examines the details carefully, and does not leave after 2 minutes, it can be known that the target visitor is very interested in the current exhibit and explanation content. In this case, the answer content can be more detailed and rich. If the visitor is more than 3 meters away from the exhibit and does not stop, it can be known that the target visitor may have just asked casually. In this case, the answer content can be more concise and clear.

[0072] The aforementioned question-and-answer history can refer to the target visitor's previous questioning instructions and interaction records with the guide robot. By analyzing the target visitor's previous questioning instructions and interaction records with the guide robot, we can identify the target visitor's points of interest and communication habits, and then generate answers based on the target visitor's interests and habits.

[0073] The aforementioned preset knowledge matching algorithm, for example, can search and match the aforementioned question keywords with knowledge in a preset knowledge graph, select knowledge content with a high degree of matching, and generate answer content through a preset natural language processing algorithm.

[0074] Of course, the above are just possible examples. The actual process of generating the answer may be the same as or different from the examples above, and is not limited to the examples above.

[0075] Furthermore, based on the above embodiments, the aforementioned current environmental information may also include: the real-time location of the target visitor. The aforementioned safety requirements include: collision restrictions, hazardous area restrictions, and touch restrictions.

[0076] After acquiring current environmental information in real time through a preset data acquisition device, the above method further includes: Based on the above explanation of the obstacle information and collision restrictions, the system can avoid moving obstacles in real time during movement.

[0077] Based on the aforementioned real-time location of the target visitor, the aforementioned danger zone restrictions, the aforementioned body posture of the target visitor, and the aforementioned touch restrictions, respectively, danger zone warnings and touch warnings are issued to the aforementioned target visitor.

[0078] For example, obstacles on the narration route may not only appear when the narration robot is at a certain narration point, but may also enter the narration route when the narration robot moves from a certain narration point to the next narration point. At this time, the narration robot can no longer avoid the current collision by updating the narration route. Therefore, it can also avoid obstacles by adjusting the robot's angle and posture or crossing and avoiding obstacles as mentioned above.

[0079] Based on the real-time location of the target visitor, the restrictions on dangerous areas, the target visitor's body posture, and the restrictions on touch, the robot issues warnings for dangerous areas and touches to the target visitor. For example, when the target visitor enters a dangerous area or a restricted area, the robot can issue a voice warning asking the target visitor to leave the specific area; or when the target visitor touches an exhibit that cannot be touched, the robot can issue a voice warning asking the visitor to stop touching the exhibit, etc. However, no specific restrictions are imposed here.

[0080] In addition, based on any of the above embodiments, the above-mentioned explanation robot can be a humanoid explanation robot, and the explanation task planning result can also include: explaining body postures.

[0081] The above-mentioned target visitors have been identified. Once the target visitors are in place, the explanation will be conducted according to the above-mentioned explanation task plan, including: The explanation will follow the body postures described above.

[0082] For example, humanoid guide robots, compared to traditional wheeled chassis guide robots, can more flexibly traverse obstacles such as steps, and can avoid obstacles by simulating human side-turning movements. Furthermore, they can combine gesturing with gestures to better deliver the explanations, thus enhancing the experience for the target visitors.

[0083] Figure 4 This is a schematic diagram of a real-time planning device for explanation tasks provided in an embodiment of this application. This device can execute the aforementioned real-time planning method for explanation tasks. It can be integrated into the controller, processor, or other computing devices of the explanation robot. Figure 4 As shown, the device may include: The receiving module 410 is used to receive the explanation task instruction and analyze the instruction intent corresponding to the explanation task instruction based on the explanation task instruction and the preset natural language processing algorithm.

[0084] The planning module 420 is used to determine the explanation task planning result based on the above-mentioned instruction intent, preset knowledge graph and preset task planning algorithm. The above-mentioned preset task planning algorithm is obtained by training from the task planning sample dataset. The task planning sample dataset includes: sample instructions and corresponding sample task planning results. The above-mentioned explanation task planning result includes: explanation route, explanation content and safety requirements.

[0085] The explanation module 430 is used to identify the target visitor among the visitors according to the preset recognition rules. After confirming that the target visitor is in place, the explanation is carried out according to the explanation task planning results, and the current environmental information is obtained in real time by scanning through the preset acquisition device.

[0086] The update module 440 is used to determine whether to update the above-mentioned explanation route based on the above-mentioned current environment information and the above-mentioned preset path update algorithm.

[0087] The real-time planning method for explanation tasks provided in this application includes: receiving an explanation task instruction; analyzing and obtaining the instruction intent corresponding to the explanation task instruction based on the explanation task instruction and a preset natural language processing algorithm; determining the explanation task planning result based on the instruction intent, a preset knowledge graph, and a preset task planning algorithm, wherein the preset task planning algorithm is trained on a task planning sample dataset, the task planning sample dataset includes: sample instructions and corresponding sample task planning results, and the explanation task planning result includes: explanation route, explanation content, and safety requirements; identifying target visitors among the current visitors according to preset recognition rules; after confirming that the target visitors are in place, conducting the explanation according to the explanation task planning result, and acquiring current environmental information in real time through a preset acquisition device; and determining whether to update the explanation route based on the current environmental information and the preset path update algorithm. In this embodiment, after analyzing the received task instructions to obtain the corresponding instruction intent, the system uses a preset task planning algorithm based on a preset knowledge graph to determine the explanation task planning result corresponding to the instruction intent. During the explanation process according to the explanation task planning result, the system scans the current environment information in real time to judge obstacles and other situations on the explanation route. This enables the system to determine whether to update the explanation route based on the current environment information, so as to avoid timely adjustments when the current explanation route is affected by the environment. This improves the flexibility of the explanation robot in the face of complex environments and optimizes the user experience during the explanation process.

[0088] Optionally, the aforementioned real-time planning device for explanation tasks may further include: a generation module, used to acquire knowledge documents corresponding to the explanation tasks, and extract explanation knowledge corresponding to the explanation tasks based on a preset recognition algorithm and the aforementioned knowledge documents, wherein the aforementioned knowledge documents include: explanation script documents and professional knowledge documents. Based on the aforementioned explanation knowledge and a preset knowledge graph generation algorithm, the aforementioned preset knowledge graph is generated.

[0089] Optionally, the aforementioned real-time planning device for the explanation task may further include: an acquisition module for acquiring visitor information, including: visitor facial image, visitor name, and visitor identity background.

[0090] The aforementioned explanation module 430 can specifically be used to scan and acquire facial images of multiple visitors entering the venue using a preset acquisition device. Based on a preset recognition algorithm, the acquired facial images, and the visitor facial images, visitor information corresponding to each of the aforementioned visitors is determined. Based on the visitor information and preset visitor priority rules, the target visitor is determined.

[0091] Optionally, the aforementioned explanation route may include multiple explanation points. The aforementioned current environment information may include obstacle information along the explanation route. The aforementioned preset path update algorithm may include a route generation algorithm and a route optimization algorithm.

[0092] The aforementioned update module 440 is specifically used to determine the obstacle situation of the explanation route from the current explanation point to the next explanation point based on the obstacle information of the explanation route. If there are obstacles between the current explanation point and the next explanation point, an updated explanation route is determined, and multiple candidate routes from the current explanation point to the next explanation point are generated through the route generation algorithm. Based on the multiple candidate routes and the route selection algorithm, an updated route from the current explanation point to the next explanation point is determined, wherein the updated route is the candidate route with the shortest time and / or the fewest obstacles.

[0093] Optionally, the aforementioned current environmental information may also include: the visitor's behavior, which includes: the target visitor's body posture and the duration of stay at the current explanation point.

[0094] The receiving module 410 can also be used to receive the questioning instructions from the target visitor and analyze and obtain the questioning keywords corresponding to the questioning instructions based on the questioning instructions and the preset natural language processing algorithm.

[0095] The aforementioned explanation module 430 can also be used to generate answer content based on the aforementioned question keywords, the aforementioned preset knowledge graph, the aforementioned visitor's aforementioned visit behavior, question and answer history, preset knowledge matching algorithm, and preset natural language processing algorithm, and to answer the aforementioned target visitor based on the aforementioned answer content.

[0096] Optionally, the aforementioned current environmental information may also include: the real-time location of the target visitor. The aforementioned safety requirements include: collision restrictions, hazardous area restrictions, and touch restrictions.

[0097] The aforementioned real-time planning device for the guided tour task may further include: a safety module, used to avoid moving obstacles in real time during movement based on the obstacle information of the guided tour route and the collision restrictions. It also issues danger zone warnings and touch warnings to the target visitor based on the target visitor's real-time location and danger zone restrictions, as well as the target visitor's body posture and touch restrictions.

[0098] Optionally, the aforementioned explanation robot can be a humanoid explanation robot, and the explanation task planning results may also include: explaining body postures.

[0099] The aforementioned explanation module 430 can also be used to explain according to the body postures described above.

[0100] The above-described device is used to execute the method provided in the foregoing embodiments, and its implementation principle and technical effect are similar, so they will not be described again here.

[0101] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device can be a controller, processor, or other device with computing capabilities, such as the robot described above. Figure 5 As shown, the device 500 includes: The processor 510, storage medium 520, and bus 530 are connected in communication via bus 530.

[0102] The storage medium 520 stores machine-readable instructions that can be executed by the processor 510. When the electronic device is running, the processor 510 executes the aforementioned machine-readable instructions to perform the real-time planning method for the task described above.

[0103] It should be understood that, Figure 5 The structure shown is only a schematic diagram of an electronic device; the electronic device may also include components that are larger than those shown. Figure 5 The more or fewer components shown, or having the same Figure 5 The different configurations shown. Figure 5 The components shown can be implemented using hardware, software, or a combination thereof.

[0104] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the real-time planning method for the teaching task described in the above method embodiments.

[0105] Computer-readable storage media can be electronic storage devices such as flash memory, EEPROM (Electrically Erasable Programmable Read-Only Memory), EPROM, hard disk, or ROM. Optionally, computer-readable storage media includes non-transitory computer-readable storage medium. The computer-readable storage medium has storage space for program code that performs any of the method steps described above. This program code can be read from or written to one or more computer program exhibits. The program code can be compressed, for example, in a suitable form.

[0106] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can also be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program exhibits according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0107] In addition, the functional modules in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0108] If the functionality is implemented as a software module and sold or used as an independent exhibit, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software exhibit. This computer software exhibit is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0109] The above description is merely a preferred embodiment of this application and does not limit the patent scope of this application. Any equivalent structural transformations made based on the inventive concept of this application and the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included within the patent protection scope of this application.

Claims

1. A real-time planning method for narration tasks, applied to a narration robot, characterized in that, include: Receive a narration task instruction, and analyze the instruction intent corresponding to the narration task instruction based on the narration task instruction and a preset natural language processing algorithm; Based on the instruction intent, the preset knowledge graph, and the preset task planning algorithm, the explanation task planning result is determined. The preset task planning algorithm is obtained by training on the task planning sample dataset, which includes: sample instructions and corresponding sample task planning results. The explanation task planning result includes: explanation route, explanation content, and safety requirements. According to the preset identification rules, the target visitor is identified among the visitors. After confirming that the target visitor is in place, the explanation is carried out according to the explanation task planning results, and the current environmental information is obtained in real time by scanning through the preset collection device. Based on the current environment information and the preset path update algorithm, determine whether to update the explanation route.

2. The method according to claim 1, characterized in that, Before receiving the explanation task instruction and analyzing and obtaining the instruction intent corresponding to the explanation task instruction based on the explanation task instruction and a preset natural language processing algorithm, the method further includes: Obtain the knowledge document corresponding to the explanation task, and extract the explanation knowledge corresponding to the explanation task according to the preset recognition algorithm and the knowledge document, wherein the knowledge document includes: explanation script document and professional knowledge document; Based on the explained knowledge and the preset knowledge graph generation algorithm, the preset knowledge graph is generated.

3. The method according to claim 1, characterized in that, Before identifying the target visitors, the method further includes: Obtain visitor information, which includes: visitor facial image, visitor name, and visitor identity background; The step of identifying and obtaining target visitors from the current visitors according to preset identification rules includes: Facial images of multiple visitors were captured by scanning with a pre-set data acquisition device. Based on the preset recognition algorithm, the facial images captured, and the visitor's facial images, the visitor information corresponding to each of the visitors entering the venue is determined; The target visitors are determined based on the visitor information and preset visitor priority rules.

4. The method according to claim 1, characterized in that, The explanation route includes: multiple explanation points; the current environment information includes: obstacle information of the explanation route; the preset path update algorithm includes: route generation algorithm and route optimization algorithm; The step of determining whether to update the explanation route based on the current environment information and the preset path update algorithm includes: Based on the obstacle information of the explanation route, determine the obstacle situation of the explanation route from the current explanation point to the next explanation point; If there is an obstacle between the current explanation point and the next explanation point, the explanation route is updated, and multiple alternative routes between the current explanation point and the next explanation point are generated by the route generation algorithm. Based on multiple alternative routes and the route selection algorithm, an updated route is determined from the current explanation point to the next explanation point, wherein the updated route is the alternative route with the shortest travel time and / or the fewest obstacles.

5. The method according to claim 4, characterized in that, The current environmental information also includes: the visitor's behavior, which includes: the visitor's body posture and the duration of stay at the current explanation point; After completing the explanation according to the updated explanation route, the method further includes: Receive the question instruction from the target visitor, and analyze and obtain the question keywords corresponding to the question instruction based on the question instruction and the preset natural language processing algorithm; Based on the question keywords, the preset knowledge graph, the target visitor's visit behavior, question and answer history, preset knowledge matching algorithm, and preset natural language processing algorithm, the answer content is generated, and the target visitor is answered based on the answer content.

6. The method according to claim 5, characterized in that, The current environmental information also includes: the real-time location of the target visitor; the safety requirements include: collision restrictions, danger zone restrictions, and touch restrictions; After acquiring the current environmental information in real time through a preset acquisition device, the method further includes: Based on the explained route obstacle information and collision restrictions, the system performs real-time avoidance of moving obstacles during movement. Based on the target visitor's real-time location and the danger zone restrictions, the target visitor's body posture and the touch restrictions, danger zone alerts and touch alerts are issued to the target visitor.

7. The method according to any one of claims 1-6, characterized in that, The explanation robot is a humanoid explanation robot, and the explanation task planning result also includes: explaining body postures; The process of identifying target visitors and, after they are in place, conducting the explanation according to the planned explanation task includes: Explain the body postures as described above.

8. A real-time planning device for instructional tasks, characterized in that, include: The receiving module is used to receive the explanation task instruction and analyze the instruction and the preset natural language processing algorithm to obtain the instruction intent corresponding to the explanation task instruction. The planning module is used to determine the explanation task planning result based on the instruction intent, the preset knowledge graph and the preset task planning algorithm. The preset task planning algorithm is obtained by training from the task planning sample dataset. The task planning sample dataset includes: sample instructions and corresponding sample task planning results. The explanation task planning result includes: explanation route, explanation content and safety requirements. The explanation module is used to identify the target visitor among the current visitors according to the preset recognition rules. After confirming that the target visitor is in place, the explanation is carried out according to the explanation task planning results, and the current environmental information is obtained in real time by scanning through the preset acquisition device. The update module is used to determine whether to update the explanation route based on the current environment information and the preset path update algorithm.

9. An electronic device, characterized in that, include: The device includes a processor, a storage medium, and a bus, wherein the storage medium stores machine-readable instructions executable by the processor, and when the electronic device is in operation, the processor communicates with the storage medium via the bus, and the processor executes the machine-readable instructions to perform the steps of the method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the steps of the method as described in any one of claims 1-7.

Images