A control method and system for an artificial intelligence robot

Abstract

The present invention discloses a control method and system for an artificial intelligence robot, which relates to the field of robot control technology, and includes the following steps: obtaining dynamic data and user behavior data of the robot's surrounding environment, analyzing and summarizing a first analysis model, performing a first analysis on the dynamic data and user behavior data through the first analysis model, and obtaining a first control instruction; obtaining static data and historical task data of the robot, analyzing and summarizing a second analysis model, performing a second analysis on the static data and historical task data through the second analysis model, and obtaining a second control instruction; combining the first control instruction and the second control instruction to obtain a final control instruction, and completing the control and task prediction of the intelligent robot. The present invention effectively solves the problems of insufficient task association analysis and insufficient historical data utilization in the prior art through an innovative task association analysis mechanism and a comprehensive historical data utilization strategy.

Description

Technical Field

[0001] The present invention relates to the field of robot control technology, and in particular to a control method and system for an artificial intelligence robot. Background Art

[0002] Robot control methods refer to a series of processes and strategies that guide and manage robot behavior through various techniques and algorithms. These methods usually involve steps such as sensing the environment, processing information, making decisions, and executing actions. Advanced robot control systems are designed to enable robots to adapt to complex and changing environments, understand and respond to user needs, and complete various tasks efficiently. These systems usually combine sensor data processing, machine learning algorithms, decision models, and execution mechanisms to achieve intelligent and autonomous robot behavior.

[0003] However, existing robot control methods have some significant disadvantages:

[0004] Insufficient task association analysis: Most existing methods lack effective mechanisms to analyze and identify the associations between tasks. They usually only focus on the execution of the current task, while ignoring the possible temporal and logical relationships between tasks, resulting in the robot being unable to proactively provide relevant task suggestions or optimize the order of task execution.

[0005] Insufficient use of historical data: Existing systems often fail to make full use of historical task data to learn user behavior patterns and task execution rules, which limits the robot's ability to predict user needs and related tasks, making the control system lack foresight and intelligence. Summary of the invention

[0006] In view of the problems existing in the control and system of the existing artificial intelligence robot, the present invention is proposed.

[0007] Therefore, the problem to be solved by the present invention is: insufficient task relevance and insufficient use of historical data.

[0008] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0009] In a first aspect, an embodiment of the present invention provides a control method for an artificial intelligence robot, which comprises the following steps:

[0010] Acquire dynamic data of the robot's surrounding environment and user behavior data, analyze and summarize a first analysis model, perform a first analysis on the dynamic data and user behavior data through the first analysis model, and obtain a first control instruction;

[0011] Obtain static data and historical task data of the robot, analyze and summarize a second analysis model, perform a second analysis on the static data and historical task data using the second analysis model, and obtain a second control instruction;

[0012] The first control instruction and the second control instruction are combined to obtain a final control instruction to complete the control and task prediction of the intelligent robot.

[0013] As a preferred solution of the control method of the artificial intelligence robot of the present invention, wherein: the dynamic data includes real-time environmental images, sound signals, distance sensor data, time, special event markers, and the current state of the robot;

[0014] The user behavior data includes user instructions, instruction time, and instruction frequency;

[0015] The static data includes robot hardware parameters, mechanical structure information, preset task objectives, and operation constraints;

[0016] The historical task data includes the task sequence, task time, and task frequency executed in the past.

[0017] As a preferred solution of the control method of the artificial intelligence robot of the present invention, the first analysis model is based on a comprehensive analysis of user behavior data and dynamic data by a recurrent neural network, specifically:

[0018] Analyze user behavior data based on the output of the recurrent neural network to obtain a preliminary task prediction T t 1 ; Combined with environmental information in dynamic data E t and the robot's current state data S t , adjust the preliminary task forecast and get the revised task forecast T t 2 ;

[0019] By correcting the task prediction T t 2 , generate the first control instruction C t .

[0020] As a preferred solution of the control method of the artificial intelligence robot of the present invention, wherein: the preliminary task prediction T t 1 It is expressed by the formula:

[0021] T t 1 =RNN(U t , H t-1 );

[0022] Where U t Represented as user behavior data, H t-1 Represented as the previous hidden state, the overall formula represents the use of a recurrent neural network to generate preliminary task prediction results based on the current user behavior data and the previous hidden state;

[0023] The modified task prediction T t 2 It is expressed by the formula:

[0024] T t 2 =T t 1 ·(1+α·E t )+β·(S t -S ref );

[0025] In the formula, α and β are weight factors, which respectively control the influence of environmental data and state data on the prediction of the correction task. ref It is represented as the reference state, and the overall formula represents the combination of the preliminary task prediction with the current environment data and the current state of the robot to generate a revised task prediction result;

[0026] The first control instruction C t It is expressed by the formula:

[0027] C t =f(T t 2 , E t );

[0028] Wherein, function f represents the process of converting task prediction and environmental data into control instructions, and the overall formula represents combining the revised task prediction and current environmental data to generate the first control instruction.

[0029] As a preferred solution of the control method of the artificial intelligence robot of the present invention, the second analysis model considers the influence of dynamic data when generating control instructions, specifically:

[0030] According to the historical task data D and the preset task target G, the task probability distribution P is obtained. T ;

[0031] By the task probability distribution P T and the robot hardware parameters, and obtain the task sequence S after considering the hardware limitations 1 ;

[0032] Combining environmental information in dynamic data t and the robot's current state data S t, adjust the task sequence and get the corrected task sequence S 2 ;

[0033] Based on the modified task sequence S 2 and the operation constraints of the robot, and obtain the second control instruction C w .

[0034] As a preferred solution of the control method of the artificial intelligence robot of the present invention, wherein: the task probability distribution P T It is expressed by the formula:

[0035] P T = g(D, G);

[0036] In the formula, function g represents the process of converting historical data and task objectives into task probability distribution, and the overall formula represents the calculation of task probability distribution based on historical task data and preset task objectives;

[0037] The task sequence S after considering hardware limitations 1 It is expressed by the formula:

[0038] S 1 =h( PT， H);

[0039] In the formula, function h represents the process of generating task sequences after considering hardware limitations, and the overall formula represents the generation of preliminary task sequences based on task probability distribution and robot hardware parameters;

[0040] The modified task sequence S 2 It is expressed by the formula:

[0041] S 2 =S 1 +γ·(E t -E ref )+δ·(S t -S ref );

[0042] In the formula, γ and δ are weight coefficients, which respectively control the correction effect of environmental data and state data on the task sequence. ref It is represented as the reference environment data, which is used to measure the difference with the current environment and state. The overall formula is represented as combining the preliminary task sequence with the current environment data and the current state data of the robot to generate a revised task sequence;

[0043] The second control instruction C w It is expressed by the formula:

[0044] C w = k(S 2, L, E t );

[0045] Where L represents the operation constraint, function k represents the process of converting the modified task sequence and constraint into control instructions, and the overall formula represents combining the modified task sequence, operation constraint and current environment data to generate the second control instruction.

[0046] As a preferred solution of the control method of the artificial intelligence robot of the present invention, wherein: the final control instruction C z The formula is obtained as,

[0047] C z =ω·C t +(1-ω)·C w ;

[0048] In the formula, ω represents the dynamic weight factor, and its value range is 0≤ω≤1;

[0049] When the final control instruction C is determined z After that, it determines whether there is a related task based on the historical task data. If the judgment result is yes, it actively asks the user whether to execute the related task. If the judgment result is no, it does nothing.

[0050] The determination methods of associated tasks include:

[0051] Collect user historical task data, including task type, execution order, and time interval information;

[0052] Identify pairs of tasks that occur simultaneously or are performed in a specific order (R 1 , R 2 );

[0053] Computational task pair (R 1 , R 2 ) is associated with a probability P(R 1 , R 2 ), when the associated probability P(R 1 , R 2 ) is greater than the probability setting value, the task pair is associated and the time association calculation is performed, otherwise no association is performed;

[0054] The time-related calculation includes calculating the task pair (R 1 , R 2 ) and obtain the number of times the time difference is less than the standard value. When the number is greater than the time setting value, the task pair (R 1 , R 2 ) to associate, otherwise cancel the association;

[0055] Finally, the correlation degree is obtained, and its calculation formula is expressed as:

[0056] G(R 1 , R 2 )=μ 1 ·p(R 1 , R 2 )+μ 2 ·η(R 1 , R 2 );

[0057] In the formula, G(R 1 , R 2 ) is expressed as the correlation degree, μ 1 and μ 2 Expressed as a weight coefficient, η(R 1 , R 2 ) is expressed as time correlation;

[0058] When the correlation G(R 1 , R 2 ) is greater than the threshold θ, then the task R is judged 2 For task R 1 It is an associated task, otherwise it is not an associated task.

[0059] In a second aspect, an embodiment of the present invention provides a control system for an artificial intelligence robot, which includes a data acquisition module, a data calculation module, and a command output module;

[0060] The data acquisition module is used to obtain dynamic data of the robot's surrounding environment and user behavior data, and obtain static data and historical task data of the robot;

[0061] The data calculation module is used to generate a first analysis model and a second analysis model;

[0062] The instruction output module is used to obtain a first control instruction and a second control instruction, and merge the two control instructions to obtain a final control instruction.

[0063] In a third aspect, an embodiment of the present invention provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the processor executes the computer program, it implements any step of the above-mentioned artificial intelligence robot control method.

[0064] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the above-mentioned artificial intelligence robot control method is implemented.

[0065] The beneficial effects of the present invention are as follows: through an innovative task association analysis mechanism and a comprehensive historical data utilization strategy, the problems of insufficient task association analysis and insufficient historical data utilization in the prior art are effectively solved.

[0066] In terms of task association analysis, this method introduces an identification mechanism that comprehensively considers task type, execution order and time interval. By calculating the association probability and time correlation of task pairs, it significantly improves the accuracy and comprehensiveness of associated task identification, which enables the system to provide related task suggestions more intelligently and optimize the task execution order, thereby improving user experience and work efficiency.

[0067] In terms of historical data utilization, this method fully utilizes historical task data at multiple levels, including generating task probability distribution, generating task sequences considering hardware limitations, and dynamically combining historical data analysis results with current environment and status information. This multi-level and multi-angle historical data utilization method greatly improves the system's predictive ability for user needs and environmental adaptability. By integrating short-term dynamic data and long-term historical data, this method achieves a balance between rapid response to complex environments and long-term stability, significantly improving the intelligence, efficiency and robustness of the robot control system. This not only provides users with a smarter and more considerate service experience, but also lays a solid foundation for achieving more advanced human-machine collaboration. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0069] Figure 1 Scenario diagram for the control method of the artificial intelligence robot. DETAILED DESCRIPTION

[0070] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0071] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0072] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0073] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0074] At the same time, in the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "upper, lower, inner and outer" are based on the directions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention. In addition, the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0075] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0076] Example 1

[0077] Reference Figure 1 , which is the first embodiment of the present invention, and provides a control method for an artificial intelligence robot, comprising the following steps:

[0078] S1. Acquire dynamic data of the robot's surrounding environment and user behavior data, analyze and summarize a first analysis model, perform a first analysis on the dynamic data and user behavior data through the first analysis model, and obtain a first control instruction.

[0079] The dynamic data includes real-time environment images, sound signals, distance sensor data, time, special event markers, and the current state of the robot;

[0080] The user behavior data includes user instructions, instruction time, and instruction frequency;

[0081] The static data includes robot hardware parameters, mechanical structure information, preset task objectives, and operation constraints;

[0082] The historical task data includes the task sequence, task time, and task frequency executed in the past.

[0083] The first analysis model is based on a comprehensive analysis of user behavior data and dynamic data by a recurrent neural network, specifically:

[0084] Analyze user behavior data based on the output of the recurrent neural network to obtain a preliminary task prediction T t 1 ; Combined with environmental information in dynamic data E t and the robot's current state data S t , adjust the preliminary task forecast and get the revised task forecast T 1 t ;

[0085] By correcting the task prediction T t 2 , generate the first control instruction C t .

[0086] The initial task prediction T t 1 It is expressed by the formula:

[0087] T t 1 =RNN(U t , H t-1 );

[0088] Where U t Represented as user behavior data, H t-1 Represented as the previous hidden state, the overall formula represents the use of a recurrent neural network to generate preliminary task prediction results based on the current user behavior data and the previous hidden state;

[0089] The modified task prediction T t 2 It is expressed by the formula:

[0090] T t 2 =T t 1 ·(1+α·Et )+β·(S t -S ref );

[0091] In the formula, α and β are weight factors, which respectively control the influence of environmental data and state data on the prediction of the correction task. ref It is represented as the reference state, and the overall formula represents the combination of the preliminary task prediction with the current environment data and the current state of the robot to generate a revised task prediction result;

[0092] The first control instruction C t It is expressed by the formula:

[0093] C t =f(T t 2 , E t );

[0094] In the formula, function f represents the process of converting task prediction and environmental data into control instructions, and the formula as a whole represents combining the revised task prediction and current environmental data to generate the first control instruction;

[0095] For example, this control method is a control method for a smart home robot, which is equipped with a camera, a microphone, a distance sensor, and other necessary hardware devices. The following is the workflow of the robot based on the provided control method:

[0096] Dynamic data includes:

[0097] Real-time environment image: Capture the real-time image of the room through the camera;

[0098] Sound signal: Use the microphone to receive environmental sounds and user voice commands;

[0099] Distance sensor data: detects the distance of surrounding obstacles;

[0100] Time: record the current time, such as 3:00 p.m. on Monday;

[0101] Special event markers: such as pet activity detection;

[0102] The current status of the robot: such as battery power 80%, cleaning brush wear degree 20%.

[0103] User behavior data includes:

[0104] User instructions: such as "clean the living room";

[0105] Instruction time: record the time when each instruction is issued;

[0106] Instruction frequency: Statistics on the usage frequency of various instructions.

[0107] Use a recurrent neural network to process user behavior data. For example, if a user often issues the command "clean the living room" at 3 p.m. on Monday, the model will predict that the task of cleaning the living room may need to be performed at this time. At this time, formula T t 1 =RNN(U t , H t-1 ), U t Indicates the current user instruction (clean the living room);

[0108] Adjustments are made based on the environmental information and the robot status. For example, if someone is detected in the living room or the robot is low on power, the initial prediction will be revised. Formula T t 2 =T t 1 ·(1+α·E t )+β·(S t -S ref ), E t is the environmental information (personnel activity detected), S t The current status data of the robot (low battery);

[0109] Therefore, the modified first control instruction is, for example, an instruction to "start cleaning the living room after a delay of 30 minutes, avoiding areas where people are active."

[0110] S2. Obtain static data and historical task data of the robot, analyze and summarize a second analysis model, perform a second analysis on the static data and historical task data through the second analysis model, and obtain a second control instruction.

[0111] The second analysis model considers the influence of dynamic data when generating control instructions, specifically:

[0112] According to the historical task data D and the preset task target G, the task probability distribution P is obtained. T ;

[0113] By the task probability distribution P T and the robot hardware parameters, and obtain the task sequence S after considering the hardware limitations 1 ;

[0114] Combining environmental information in dynamic data t and the robot's current state data S t , adjust the task sequence and get the corrected task sequence S 2 ;

[0115] Based on the modified task sequence S 2and the operation constraints of the robot, and obtain the second control instruction C w .

[0116] The task probability distribution P T It is expressed by the formula:

[0117] P T = g(D, G);

[0118] In the formula, function g represents the process of converting historical data and task objectives into task probability distribution, and the overall formula represents the calculation of task probability distribution based on historical task data and preset task objectives;

[0119] The task sequence S after considering hardware limitations 1 It is expressed by the formula:

[0120] S 1 =h(P T , H);

[0121] In the formula, function h represents the process of generating task sequences after considering hardware limitations, and the overall formula represents the generation of preliminary task sequences based on task probability distribution and robot hardware parameters;

[0122] The modified task sequence S 2 It is expressed by the formula:

[0123] S 2 =S 1 +γ·(E t -E ref )+δ·(S t -S ref );

[0124] In the formula, γ and δ are weight coefficients, which respectively control the correction effect of environmental data and state data on the task sequence. ref It is represented as the reference environment data, which is used to measure the difference with the current environment and state. The overall formula is represented as combining the preliminary task sequence with the current environment data and the current state data of the robot to generate a revised task sequence;

[0125] The second control instruction C w It is expressed by the formula:

[0126] C w = k(S 2 , L, E t );

[0127] Where L represents the operation constraint, function k represents the process of converting the modified task sequence and constraint into control instructions, and the overall formula represents combining the modified task sequence, operation constraint and current environment data to generate the second control instruction.

[0128] Assume that the robot is a multifunctional smart home robot that can perform multiple tasks such as cleaning, cooking, and security. The robot is equipped with various necessary hardware devices and has rich historical task data. The following is the workflow of the robot based on the provided control method:

[0129] Static data includes:

[0130] Robot hardware parameters: such as cleaning power, cooking temperature range, camera resolution, etc.

[0131] Mechanical structure information: such as robot size, range of motion of joints, etc.

[0132] Preset task objectives: such as daily cleaning area, types of dishes to be cooked, frequency of security patrols, etc.

[0133] Operational restrictions: such as maximum working hours, noise limits, etc.;

[0134] Historical mission data includes:

[0135] The sequence of tasks performed in the past: such as “cleaning the living room > cooking dinner > night security patrol”;

[0136] Task time: record the start and end time of each task;

[0137] Task frequency: Statistics on the execution frequency of various tasks.

[0138] The second analysis model generates a task probability distribution. For example, based on historical data, at 6 p.m. on weekdays, the probability of the task "cooking dinner" is 80%, the probability of the task "cleaning the living room" is 15%, and the probability of the task "security patrol" is 5%;

[0139] Combine the task probability distribution and the robot hardware parameters to generate a preliminary task sequence. Considering that the robot's cooking function requires preheating time, the system may arrange the task sequence of "preheating cooking equipment > cleaning the living room > cooking dinner > security patrol";

[0140] The task sequence is modified according to the current environment and robot status. If the system detects that the living room is particularly dirty, it may increase the cleaning time or adjust the cleaning order. If it finds that the kitchen is short of ingredients, it may postpone or cancel the cooking task.

[0141] Based on the revised task sequence and operation constraints, specific control instructions are generated. For example, a detailed instruction sequence is generated: "Start cleaning the living room at 16:30 (estimated 45 minutes), start preheating cooking equipment at 17:15, start cooking dinner at 17:30 (estimated 60 minutes), and start security patrol at 19:00".

[0142] S3. Combining the first control instruction and the second control instruction to obtain a final control instruction, and completing the control and task prediction of the intelligent robot.

[0143] The final control instruction C z The formula is obtained as,

[0144] C z =ω·C t +(1-ω)·C w ;

[0145] In the formula, ω represents the dynamic weight factor, and its value range is 0≤ω≤1;

[0146] When outputting the final instruction, for example, if ω is 0.4, the final instruction will be output as:

[0147] 16:30-16:50, quickly clean the living room, avoiding areas where people and pets are active;

[0148] 16:50-17:00, preheat cooking equipment;

[0149] 17:00-18:00, cooking dinner;

[0150] 18:00-18:45, thoroughly clean the living room and kitchen;

[0151] 21:00, night security patrols begin;

[0152] Note: Keep the noise level below 60 decibels and pay attention to the battery level.

[0153] When the final control instruction C is determined z After that, it determines whether there is a related task based on the historical task data. If the judgment result is yes, it actively asks the user whether to execute the related task. If the judgment result is no, it does nothing.

[0154] The determination methods of associated tasks include:

[0155] Collect user historical task data, including task type, execution order, and time interval information;

[0156] Identify pairs of tasks that occur simultaneously or are performed in a specific order (R 1 , R 2 );

[0157] Computational task pair (R 1 , R 2 ) is associated with a probability P(R 1 , R 2 ), when the associated probability P(R 1 , R 2 ) is greater than the probability setting value, the task pair is associated and the time association calculation is performed, otherwise no association is performed;

[0158] The time-related calculation includes calculating the task pair (R 1 , R 2 ) and obtain the number of times the time difference is less than the standard value. When the number is greater than the time setting value, the task pair (R 1 , R 2 ) to associate, otherwise cancel the association;

[0159] Finally, the correlation degree is obtained, and its calculation formula is expressed as:

[0160] G(R 1 , R 2 )=μ 1 ·p(R 1 , R 2 )+μ 2 ·η(R 1 , R 2 );

[0161] In the formula, G(R 1 , R 2 ) is expressed as the correlation degree, μ 1 and μ 2 Expressed as a weight coefficient, η(R 1 , R 2 ) is expressed as time correlation;

[0162] When the correlation G(R 1 , R 2 ) is greater than the threshold θ, then the task R is judged 2 For task R 1 It is an associated task, otherwise it is not an associated task.

[0163] During implementation, historical data was collected: Living room cleaning and kitchen cleaning often occur in pairs;

[0164] Identify task pairs: [living room cleaning, kitchen cleaning];

[0165] Calculate the association probability: Assume P(kitchen clean | living room clean) = 0.8;

[0166] Time correlation ηη(kitchen cleaning|living room cleaning) = 0.8;

[0167] Correlation calculation: G(R 1 , R 2 )=μ 1 ·p(R 1 , R 2 )+μ 2 ·η(R 1 , R 2 );

[0168] Assume that the weight coefficient μ 1 and μ 2 are equal to 0.6 and 0.4 respectively;

[0169] P = 0.8 (association probability), η = 0.8 (temporal correlation);

[0170] The calculated correlation degree is equal to 0.64. Assuming the threshold value θ = 0.5, since 0.64>0.5, it is determined that kitchen cleaning is an associated task of living room cleaning. Therefore, the robot starts to perform the living room cleaning task. After completion, it will actively ask the user: "I have completed the living room cleaning. According to past records, you usually clean the kitchen after cleaning the living room. Do you need me to clean the kitchen now?"

[0171] In summary, through the innovative task correlation analysis mechanism and comprehensive historical data utilization strategy, the problems of insufficient task correlation analysis and insufficient historical data utilization in the existing technology are effectively solved.

[0172] In terms of task association analysis, this method introduces an identification mechanism that comprehensively considers task type, execution order and time interval. By calculating the association probability and time correlation of task pairs, it significantly improves the accuracy and comprehensiveness of associated task identification, which enables the system to provide related task suggestions more intelligently and optimize the task execution order, thereby improving user experience and work efficiency.

[0173] In terms of historical data utilization, this method fully utilizes historical task data at multiple levels, including generating task probability distribution, generating task sequences considering hardware limitations, and dynamically combining historical data analysis results with current environment and status information. This multi-level and multi-angle historical data utilization method greatly improves the system's predictive ability for user needs and environmental adaptability. By integrating short-term dynamic data and long-term historical data, this method achieves a balance between rapid response to complex environments and long-term stability, significantly improving the intelligence, efficiency and robustness of the robot control system. This not only provides users with a smarter and more considerate service experience, but also lays a solid foundation for achieving more advanced human-machine collaboration.

[0174] Example 2

[0175] On the basis of the first embodiment, this embodiment further provides a control system of an artificial intelligence robot, including a data acquisition module, a data calculation module, and a command output module;

[0176] The data acquisition module is used to obtain dynamic data of the robot's surrounding environment and user behavior data, and obtain static data and historical task data of the robot;

[0177] The data calculation module is used to generate a first analysis model and a second analysis model;

[0178] The instruction output module is used to obtain a first control instruction and a second control instruction, and merge the two control instructions to obtain a final control instruction.

[0179] This embodiment also provides a computer device suitable for the control method of an artificial intelligence robot, including a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the control method of the artificial intelligence robot proposed in the above embodiment.

[0180] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0181] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the control method of the artificial intelligence robot proposed in the above embodiment is implemented.

[0182] The storage medium proposed in this embodiment and the data storage method proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be found in the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0183] Example 3

[0184] On the basis of the previous two embodiments, this embodiment provides a control method for an artificial intelligence robot. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0185] Experimental environment: simulate a smart home environment, including a living room, kitchen, bedroom and bathroom.

[0186] Robot tasks: cleaning, cooking assistance, item organization, temperature regulation, etc.

[0187] Experiment period: simulate 30 days of usage.

[0188] Comparison method: Give the execution content through instructions.

[0189] Evaluation indicators: task prediction accuracy; user satisfaction; task completion efficiency; number of proactive services.

[0190] Data generation: Create a dataset that simulates user behavior, including various task requests and execution records within 30 days, and add some regular task pairs to the data, such as "clean the kitchen after cooking" and "adjust the room temperature before going to bed."

[0191] Method implementation: Implement the task prediction and execution logic of this method and the comparison method respectively.

[0192] Experimental execution: For each day of simulated data, let the two methods perform task prediction and execution respectively, and record the performance of each method.

[0193] Data collection: Collect performance data of the two methods on various evaluation indicators.

[0194] Task prediction accuracy: This method: 90%; Comparison method: 0;

[0195] Analysis: This method can more accurately predict the user's next possible task by considering the task relevance, while the comparison method cannot achieve this step.

[0196] User satisfaction (out of 100): This method: 92; Comparative method: 78;

[0197] Analysis: Because this approach can better predict and proactively provide services, user satisfaction has increased significantly.

[0198] Task completion efficiency: This method: 120 minutes; Comparative method: 150 minutes;

[0199] Analysis: This method reduces the waiting time between tasks and improves overall efficiency by predicting related tasks.

[0200] Active service times: This method: 45 times; Comparison method: 0 times;

[0201] Analysis: This method can better identify related tasks, so more services are provided proactively, improving user experience.

[0202] Specific case analysis:

[0203] Take the task pair “cooking-cleaning the kitchen” as an example:

[0204] Association probability calculation: In the 30 days of data, there were 25 days of cooking tasks, and 22 days of kitchen cleaning after cooking.

[0205] P(cleaning the kitchen|cooking) = 22 / 25 = 0.88;

[0206] Time correlation calculation: Among the 22 "cooking-cleaning the kitchen" task pairs, the time interval of 20 times was less than 30 minutes;

[0207] Time correlation: T = 20 / 22 = 0.91;

[0208] The correlation calculation result is 0.892;

[0209] Related task judgment: Assuming the threshold value θ = 0.7, since 0.892>0.7, cleaning the kitchen is determined to be a related task of cooking.

[0210] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A control method for an artificial intelligence robot, characterized in that: The following steps are included: Acquire dynamic data of the robot's surrounding environment and user behavior data, analyze and summarize a first analysis model, perform a first analysis on the dynamic data and user behavior data through the first analysis model, and obtain a first control instruction; Obtain static data and historical task data of the robot, analyze and summarize a second analysis model, perform a second analysis on the static data and historical task data using the second analysis model, and obtain a second control instruction; Combining the first control instruction and the second control instruction to obtain a final control instruction, completing the control and task prediction of the intelligent robot; The first analysis model is based on a comprehensive analysis of user behavior data and dynamic data by a recurrent neural network, specifically: Analyze user behavior data based on the output of the recurrent neural network to obtain preliminary task predictions ; Combine environmental information in dynamic data and the robot's current status data , adjust the preliminary task forecast to obtain the revised task forecast ; Through the revised task prediction , generate the first control instruction ; The initial mission forecast It is expressed by the formula: ； In the formula, Represented as user behavior data, Represented as the previous hidden state, the overall formula represents the use of a recurrent neural network to generate preliminary task prediction results based on the current user behavior data and the previous hidden state; The revised mission prediction It is expressed by the formula: ； In the formula, and are expressed as weight factors, respectively controlling the impact of environmental data and state data on the prediction of the correction task. It is represented as the reference state, and the overall formula represents the combination of the preliminary task prediction with the current environment data and the current state of the robot to generate a revised task prediction result; The first control instruction It is expressed by the formula: ； In the formula, the function represents the process of converting the task prediction and the environment data into the control instruction. The formula as a whole represents combining the revised task prediction and the current environment data to generate the first control instruction; The second analysis model considers the influence of dynamic data when generating control instructions, specifically: Obtain the task probability distribution based on historical task data D and preset task target G ; Through the task probability distribution and robot hardware parameters to obtain the task sequence after considering hardware limitations ; Incorporating environmental information in dynamic data and the robot's current status data , adjust the task sequence to obtain the corrected task sequence ; Based on the revised task sequence And the robot's operating constraints, get the second control instruction .

2. The control method of the artificial intelligence robot according to claim 1, characterized in that: The dynamic data includes real-time environment images, sound signals, distance sensor data, time, special event markers, and the current state of the robot; The user behavior data includes user instructions, instruction time, and instruction frequency; The static data includes robot hardware parameters, mechanical structure information, preset task objectives, and operation constraints; The historical task data includes the task sequence, task time, and task frequency executed in the past.

3. The control method of the artificial intelligence robot according to claim 2, characterized in that: The task probability distribution It is expressed by the formula: ; In the formula, the function It is expressed as the process of converting historical data and task objectives into task probability distribution. The overall formula is expressed as calculating task probability distribution based on historical task data and preset task objectives. The task sequence after considering hardware limitations It is expressed by the formula: ; In the formula, the function It is expressed as the process of generating a task sequence after considering the hardware limitations. The overall formula is expressed as generating a preliminary task sequence based on the task probability distribution and the robot hardware parameters; The revised task sequence It is expressed by the formula: ； In the formula, and are expressed as weight coefficients, which respectively control the correction effect of environmental data and state data on the task sequence. It is represented as the reference environment data, which is used to measure the difference with the current environment and state. The overall formula is represented as combining the preliminary task sequence with the current environment data and the current state data of the robot to generate a revised task sequence; The second control instruction It is expressed by the formula: ； Where L represents the operation constraint, function k represents the process of converting the modified task sequence and constraint into control instructions, and the overall formula represents combining the modified task sequence, operation constraint and current environment data to generate the second control instruction.

4. The control method of the artificial intelligence robot according to claim 3, characterized in that: The final control instruction The formula is obtained as, ； In the formula, Expressed as a dynamic weight factor, the value range is ; When the final control order is determined After that, it determines whether there is a related task based on the historical task data. If the judgment result is yes, it actively asks the user whether to execute the related task. If the judgment result is no, it does nothing. The determination methods of associated tasks include: Collect user historical task data, including task type, execution order, and time interval information; Identify pairs of tasks that occur simultaneously or in a specific order ; Computational tasks The probability of association , when the associated probability When it is greater than the probability setting value, the task pair is associated and the time association calculation is performed, otherwise no association is performed; The time-related calculation includes calculating the task pair The time difference is calculated and the number of times the time difference is less than the standard value is obtained. When the number is greater than the time setting value, the task is To associate, otherwise to cancel the association; Finally, the correlation degree is obtained, and its calculation formula is expressed as: ； In the formula, Expressed as the degree of association, and Expressed as weight coefficient, It is expressed as time correlation; When the correlation Greater than threshold When For the task It is an associated task, otherwise it is not an associated task.

5. A control system for an artificial intelligence robot, based on the control method for an artificial intelligence robot according to any one of claims 1 to 4, characterized in that: It includes a data acquisition module, a data calculation module, and an instruction output module; The data acquisition module is used to obtain dynamic data of the robot's surrounding environment and user behavior data, and obtain static data and historical task data of the robot; The data calculation module is used to generate a first analysis model and a second analysis model; The instruction output module is used to obtain a first control instruction and a second control instruction, and merge the two control instructions to obtain a final control instruction.

6. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the control method of the artificial intelligence robot described in any one of claims 1 to 4 are implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the control method of the artificial intelligence robot described in any one of claims 1 to 4 are implemented.

Images