A humanoid robot posture intelligent control method and system

Abstract

The present invention provides a posture intelligent control method and system for a humanoid robot, which relates to the field of intelligent control technology. The method comprises: performing posture change planning based on initial posture information of a target humanoid robot to generate a first posture stability coefficient, sensing the real-time posture information of the target humanoid robot in real time through an inertial measurement unit to generate a second posture stability coefficient, comparing the second posture stability coefficient with the first posture stability coefficient to generate adjustment amplitude data, performing simulation training on the posture control of the target humanoid robot based on the adjustment amplitude data in a virtual simulation environment, generating a posture simulation training result to formulate a posture control strategy, and performing intelligent control on the posture of the target humanoid robot, thereby solving the technical problem in the prior art of lacking control over the robot posture, resulting in low efficiency of the robot in completing tasks, thereby improving the operation accuracy of the robot and better coping with work tasks in complex environments.

Description

Technical Field

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

[0002] With the continuous development of science and technology, humanoid robot intelligent control technology is also being continuously researched and practiced. This technology is becoming increasingly mature and has become an important part of future intelligent production and services.

[0003] Humanoid robots have a wide range of applications in various fields, including healthcare, household use, and industry. In the medical field, humanoid robots can assist doctors in performing surgeries, achieving precise treatment while reducing risks and errors. In the domestic field, humanoid robots can serve as home assistants, performing household chores and caring for the elderly and children. In the industrial field, humanoid robots can assist workers in completing tedious and repetitive tasks, improving production efficiency while reducing labor intensity and safety risks. However, existing technologies lack control over the robot's posture, resulting in low efficiency in completing tasks. Summary of the Invention

[0004] The present application provides a method and system for intelligent posture control of a humanoid robot, which is used to solve the technical problem in the prior art of lack of control over the robot's posture, resulting in low efficiency of the robot in completing tasks.

[0005] In view of the above problems, the present application provides a method and system for intelligent posture control of a humanoid robot.

[0006] In the first aspect, the present application provides a method for intelligent posture control of a humanoid robot, the method comprising: extracting real-time environmental information of a target humanoid robot for intelligent interaction, and determining the initial posture information of the target humanoid robot; using the dynamic control unit to perform posture change planning for the initial posture information of the target humanoid robot according to the motion path, and generating a first posture stability coefficient; using the inertial measurement unit to perceive the real-time posture information of the target humanoid robot in real time, and generating a second posture stability coefficient; comparing the second posture stability coefficient with the first posture stability coefficient, and generating adjustment amplitude data; constructing a virtual simulation environment based on the real-time environmental information, and performing simulation training on the posture control of the target humanoid robot based on the adjustment amplitude data through the virtual simulation environment, and generating a posture simulation training result; formulating a posture control strategy based on the posture simulation training result, and performing intelligent posture control of the target humanoid robot according to the posture control strategy.

[0007] In the second aspect, the present application provides a posture intelligent control system for a humanoid robot, the system comprising: an intelligent interaction module, the intelligent interaction module being used to extract the real-time environmental information of the target humanoid robot for intelligent interaction, and determine the initial posture information of the target humanoid robot; a planning module, the planning module being used to use the dynamic control unit to perform posture change planning on the initial posture information of the target humanoid robot according to the motion path, and generate a first posture stability coefficient; a first perception module, the first perception module being used to perceive the real-time posture information of the target humanoid robot in real time through the inertial measurement unit, and generate a second posture stability coefficient; a comparison module, the comparison module being used to compare the second posture stability coefficient with the first posture stability coefficient, and generate adjustment amplitude data; a simulation training module, the simulation training module being used to construct a virtual simulation environment based on the real-time environmental information, and to perform simulation training on the posture control of the target humanoid robot based on the adjustment amplitude data through the virtual simulation environment, and generate posture simulation training results; an intelligent control module, the intelligent control module being used to formulate a posture control strategy based on the posture simulation training results, and to perform intelligent control of the posture of the target humanoid robot according to the posture control strategy.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0009] The present application provides a method and system for intelligent control of the posture of a humanoid robot, which relates to the field of intelligent control technology. It solves the technical problem in the prior art of lacking control over the robot's posture, resulting in low efficiency of the robot in completing tasks, and improves the robot's operating accuracy to better cope with work tasks in complex environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 This application provides a flow chart of a method for intelligent posture control of a humanoid robot;

[0011] Figure 2 A schematic diagram of the structure of an intelligent posture control system of a humanoid robot is provided for this application.

[0012] Explanation of the accompanying drawings: intelligent interaction module 1, planning module 2, first perception module 3, comparison module 4, simulation training module 5, intelligent control module 6. DETAILED DESCRIPTION

[0013] The present application provides a method and system for intelligent posture control of a humanoid robot, which is used to solve the technical problem in the prior art of lack of control over the robot's posture, resulting in low efficiency of the robot in completing tasks.

[0014] Example 1

[0015] like Figure 1 As shown, an embodiment of the present application provides a method for intelligent posture control of a humanoid robot, the method comprising:

[0016] Step A100: extracting the real-time environmental information of the target humanoid robot for intelligent interaction, and determining the initial posture information of the target humanoid robot;

[0017] Furthermore, step A100 of the present application also includes:

[0018] Step A110: sensing and collecting basic information of objects in the target humanoid robot environment according to the sensor device group;

[0019] Step A120: extracting features from the basic information of the object using a deep learning algorithm to generate an object feature set;

[0020] Step A130: constructing an environment map based on the object feature set, and determining the real-time environment information according to the environment map.

[0021] In the present application, an intelligent control method for the posture of a humanoid robot provided in an embodiment of the present application is applied to an intelligent control system for the posture of a humanoid robot. The intelligent control system for the posture of a humanoid robot is communicatively connected to a sensor device group, and the sensor device group is used to obtain information on the position, distance, shape, color, sound characteristics, etc. of objects, thereby helping the robot understand the environment.

[0022] In order to better perform intelligent posture control on the target humanoid robot, it is first necessary to extract the surrounding environment information of the target humanoid robot and conduct intelligent interaction with the real-time environmental information of the target humanoid robot. This means that the basic information of objects in the target humanoid robot's environment is perceived and collected through a group of sensor devices. By performing target detection, recognition and tracking on images, the robot can determine objects, people or obstacles in the environment, and further analyze their position, movement and attributes. Using deep learning algorithms, the robot can extract the features and information of the basic information of objects from images or videos to generate an object feature set. Furthermore, based on the object feature set, sensors such as lidar or cameras are used to build an environmental map, and the positioning algorithm is used to determine the position of the target humanoid robot itself in the map. Then, the map and positioning information are used to plan the path, avoid obstacles or navigate to a specific location. On this basis, the real-time environmental information of the target humanoid robot is determined. Finally, by traversing the historical posture information of the target humanoid robot with the real-time environmental information as the basic reference data, the initial posture information of the target humanoid robot is generated, which serves as an important reference for the later realization of intelligent posture control of the humanoid robot.

[0023] Step A200: using the dynamics control unit to perform posture change planning on the initial posture information of the target humanoid robot according to the motion path, and generating a first posture stability coefficient;

[0024] Furthermore, step A200 of the present application also includes:

[0025] Step A210: generating a plurality of movement key points based on the target task, and generating the motion path of the target humanoid robot according to the plurality of movement key points;

[0026] Step A220: sequentially matching the first moving key point, the second moving key point ... the Nth moving key point in the motion path according to the initial posture information of the target humanoid robot to generate first posture information, second posture information ... the Nth posture information;

[0027] Step A230: Calculating, by the dynamics control unit, expected posture information of the first posture information, the second posture information, ..., the Nth posture information at corresponding time steps to generate an expected posture change data set;

[0028] Step A240: Perform stability evaluation based on the expected posture change data set to generate the first posture stability coefficient.

[0029] In the present application, in order to judge the posture stability of the target humanoid robot during movement, it is necessary to use a dynamic control unit that is communicatively connected to a posture intelligent control system of a humanoid robot to perform posture change planning for the initial posture information of the target humanoid robot according to the motion path. The dynamic control unit is used to collect kinematic parameters. First, the target task of the target humanoid robot is used as a benchmark. The target task includes the target movement posture data, target position data, target action data, etc. of the target humanoid robot. Each static posture point is recorded as a movement key point, and multiple movement key points are generated accordingly. At the same time, the movement key points are connected in time sequence to generate the movement path of the target humanoid robot. Further, the posture formed by each node position of the target humanoid robot in the initial posture information of the target humanoid robot is matched with the posture data of the first movement key point, the second movement key point...the Nth movement key point in the motion path in turn, and the posture data of the first movement key point, the second movement key point...the Nth movement key point are generated respectively. The first posture information, the second posture information... the Nth posture information correspond in sequence. The first posture information, the second posture information... the Nth posture information all correspond to the position information of each node of the target humanoid robot. Furthermore, the first posture information, the second posture information... the Nth posture information are calculated by the dynamic control unit. The expected posture information of the corresponding time step refers to the calculation of the joint angle or the position of the end effector by the dynamic control unit using the inverse kinematics or forward kinematics method. After calculating the expected posture information of the robot at each time step, an expected posture change data set is generated. Finally, the posture stability of the target humanoid robot is evaluated based on the expected posture change data set. This refers to quantifying the posture stability of the target humanoid robot by the distance from the zero torque point in the boundary of the predetermined stable area, judging the stability of the dynamic posture and the static posture of the target humanoid robot in the process of changing according to the expected posture change data set, and recording it as the first posture stability coefficient for output, thereby ensuring the realization of intelligent control of the posture of the humanoid robot.

[0030] Step A300: sensing the real-time posture information of the target humanoid robot in real time through the inertial measurement unit to generate a second posture stability coefficient;

[0031] In this application, in order to improve the accuracy of posture control adjustment of the target humanoid robot in the later stage, the real-time posture information of the target humanoid robot is first perceived through an inertial measurement unit. The inertial measurement unit is communicated with a posture intelligent control system of the humanoid robot. The inertial measurement unit is a device for measuring the acceleration and angular velocity of an object.

[0032] Based on the inertial measurement unit (IMU), the acceleration and angular velocity data of the target humanoid robot are collected. At the same time, the acceleration and angular velocity data of the target humanoid robot are filtered and calibrated to remove noise and errors to obtain accurate posture data. Using a posture solution algorithm, such as the quaternion method or the Euler angle method, the acceleration and angular velocity data are converted into the posture information of the target humanoid robot, including the posture angle and direction. Further, based on the posture information of the target humanoid robot, an evaluation index for measuring the stability of the posture is defined. The evaluation index may include posture deviation, posture change rate, etc., and a posture stability coefficient is generated based on the numerical value of the evaluation index. This coefficient is used to represent the degree of posture stability of the target humanoid robot. For example, 0 represents complete instability and 1 represents complete stability, thereby generating a second posture stability coefficient, which lays a solid foundation for the subsequent intelligent control of the posture of the humanoid robot.

[0033] Step A400: comparing the second posture stability coefficient with the first posture stability coefficient to generate adjustment amplitude data;

[0034] Furthermore, step A400 of the present application also includes:

[0035] Step A410: Obtaining a first variation range of the first posture stability coefficient and a second variation range of the second posture stability coefficient respectively;

[0036] Step A420: Identify the boundary extreme value of the first change range, and determine whether the boundary extreme value of the second change range is included in the boundary extreme value of the first change range;

[0037] Step A430: If not, adjusting the second variation range according to the first variation range to generate the adjustment range data.

[0038] In the present application, in order to more accurately perform intelligent control on the posture of the target humanoid robot, it is necessary to compare the error of the second posture stability coefficient generated above with the first posture stability coefficient. First, the maximum value and minimum value of the first posture stability coefficient in the posture change planning of the target humanoid robot are demarcated and the change amplitude corresponding to the first posture stability coefficient is obtained, and it is recorded as the first change amplitude. Further, the maximum value and minimum value of the second posture stability coefficient in the real-time posture information are demarcated and the change amplitude corresponding to the second posture stability coefficient is obtained, and it is recorded as the second change amplitude. The boundary extreme value of the first change amplitude is marked based on the maximum boundary value and the minimum boundary value within the first change amplitude, and the second change amplitude is judged according to the marked extreme value. Whether the boundary extreme value of the change amplitude is included in the boundary extreme value of the first change amplitude refers to judging whether the maximum value of the second change amplitude is less than the maximum value of the first amplitude, and whether the minimum value of the second amplitude is greater than the minimum value of the first amplitude. If both are satisfied, it is regarded that the boundary extreme value of the second change amplitude is included in the boundary extreme value of the first change amplitude, and the second change amplitude is included in the first change amplitude. If any one of them is not satisfied, it is regarded that the boundary extreme value of the second change amplitude is not included in the boundary extreme value of the first change amplitude, and the real-time posture change amplitude of the target humanoid robot is determined to be abnormal, and the second change amplitude needs to be deviated and adjusted according to the first change amplitude to generate adjustment amplitude data, which has a limited effect on realizing intelligent control of the posture of the humanoid robot.

[0039] Step A500: constructing a virtual simulation environment based on the real-time environment information, performing simulation training on the posture control of the target humanoid robot based on the adjustment amplitude data through the virtual simulation environment, and generating a posture simulation training result;

[0040] Furthermore, step A500 of the present application also includes:

[0041] Step A510: using simulation software to integrate the real-time environment information to construct a virtual simulation environment;

[0042] Step A520: controlling the posture data of the target humanoid robot according to the adjustment amplitude data based on the virtual simulation environment;

[0043] Step A530: defining evaluation indicators of the target humanoid robot, evaluating the posture data according to the evaluation indicators, and generating a posture evaluation result;

[0044] Step A540: adjusting the posture of the target humanoid robot according to the posture evaluation result to perform simulation training and generate the posture simulation training result.

[0045] Furthermore, step A540 of the present application includes:

[0046] Step A541: generating a training record log by recording the posture simulation training results;

[0047] Step A542: Calculating the posture error of the target humanoid robot according to the training record log to generate a posture deviation data set;

[0048] Step A543: performing deviation correction on the posture deviation data set to generate posture correction information;

[0049] Step A544: Correct the posture simulation training result based on the posture correction information and update the posture simulation training result.

[0050] In the present application, a model is built based on the terrain information, object information, etc. contained in the real-time environmental information of the target robot, and simulation software is used for integrated design to construct a virtual simulation environment. Then, the posture data of the target humanoid robot is controlled according to the adjustment amplitude data based on the virtual simulation environment. First, the posture data of the target humanoid robot is divided into multiple ranges, and the evaluation indicators of the target humanoid robot can be defined according to the posture deviation, posture change rate, posture stability, etc. At the same time, the posture data is evaluated according to the evaluation indicators to generate a posture evaluation result. Furthermore, the posture of the target humanoid robot is adjusted according to the posture evaluation result for simulation training, which means that the positions of multiple nodes of the target humanoid robot are simulated and moved based on the multiple node position evaluation coefficients contained in the posture evaluation result to generate a posture simulation training result.

[0051] When an anomaly occurs during the simulation training of the target humanoid robot, the posture simulation training results of the target humanoid robot need to be corrected. Furthermore, the entire training process is recorded according to the sequence of time nodes for the posture simulation training results, a training record log is prepared based on the recorded data, and the posture error of the target humanoid robot is calculated based on the training record log. This means comparing the posture data at each time step and calculating the error between the actual posture and the expected posture. The error can be calculated using the Euclidean distance metric. For example, for posture angles, the difference between two angles can be calculated; for posture directions, the difference between two direction vectors can be calculated, thereby generating a posture deviation data set. Deviation correction is performed on the posture deviation data set to generate posture correction information. This is done by analyzing data using methods such as machine learning and obtaining a deviation law function. The posture data in the posture deviation data set is analyzed to find the deviation law of the posture. The posture correction information is calculated based on the deviation law function. Exemplarily, the deviation law can be modeled as a mapping function, the actual posture is input into the function, and the posture correction information is output, so that the actual posture is corrected to be close to the expected posture. Finally, the posture data in the posture simulation training results are corrected through the posture correction information, and the posture simulation training results are updated according to the correction results, so as to serve as reference data for the later intelligent control of the posture of the humanoid robot.

[0052] Step A600: formulating a posture control strategy based on the posture simulation training result, and performing intelligent posture control on the target humanoid robot according to the posture control strategy.

[0053] Furthermore, step A600 of the present application also includes:

[0054] Step A610: Retrieving the correlation factor between the environment information set and the target humanoid robot posture;

[0055] Step A620: extracting first environmental information corresponding to the posture of the target humanoid robot in the posture simulation training result based on the correlation factor;

[0056] Step A630: matching historical posture data of the target humanoid robot based on the first environmental information;

[0057] Step A640: Formulate the posture control strategy according to the first environmental information and the historical posture data.

[0058] In the present application, the posture simulation training results are used as basic reference data to determine the posture control strategy of the target humanoid robot, which means that the correlation factor between the environmental information set and the posture of the target humanoid robot is extracted by calling the environmental information set contained in the big data, which means that the posture data generated by the target humanoid robot under different environmental information is different. Therefore, when a value is taken in the posture of the target humanoid robot, the environmental information set has one and only one corresponding value, and when a value is taken in the environmental information set, the posture of the target humanoid robot can have multiple values corresponding. Furthermore, extracting the first environmental information corresponding to the posture of the target humanoid robot in the posture simulation training result based on the correlation factor means extracting the first environmental information corresponding to the posture of the target humanoid robot based on the correlation factor. Traverse the posture simulation training results, determine the corresponding environmental information associated with the current posture of the target humanoid robot, and record it as the first environmental information. Then, use the first environmental information as the basic matching data to traverse and match it with the historical posture data set of the target humanoid robot in sequence, and regard the historical posture data with a matching degree greater than 80% as the successfully matched posture data, which serves as the benchmark data that the target humanoid robot needs to change according to the historical posture data in the first environment, thereby generating the posture control strategy of the target humanoid robot, and finally performing intelligent control of the posture changes of the target humanoid robot according to the posture control strategy, thereby improving the accuracy of the intelligent control of the posture of the humanoid robot in the later stage.

[0059] In summary, the embodiment of the present application provides a method for intelligent posture control of a humanoid robot, which includes at least the following technical effects, thereby improving the robot's operating accuracy and better coping with work tasks in complex environments.

[0060] Example 2

[0061] Based on the same inventive concept as the method for intelligent posture control of a humanoid robot in the aforementioned embodiment, Figure 2 As shown, the present application provides a posture intelligent control system for a humanoid robot, the system comprising:

[0062] An intelligent interaction module 1 is used to extract real-time environmental information of the target humanoid robot for intelligent interaction and determine initial posture information of the target humanoid robot;

[0063] Planning module 2, the planning module 2 is used to use the dynamic control unit to perform posture change planning on the initial posture information of the target humanoid robot according to the motion path, and generate a first posture stability coefficient;

[0064] a first perception module 3, configured to perceive the real-time posture information of the target humanoid robot in real time through the inertial measurement unit, and generate a second posture stability coefficient;

[0065] a comparison module 4, configured to compare the second posture stability coefficient with the first posture stability coefficient to generate adjustment amplitude data;

[0066] A simulation training module 5 is configured to construct a virtual simulation environment based on the real-time environment information, perform simulation training on the posture control of the target humanoid robot based on the adjustment amplitude data through the virtual simulation environment, and generate a posture simulation training result;

[0067] The intelligent control module 6 is used to formulate a posture control strategy based on the posture simulation training result, and perform intelligent control on the posture of the target humanoid robot according to the posture control strategy.

[0068] Furthermore, the system also includes:

[0069] a second perception module, configured to perceive and collect basic information of objects in the target humanoid robot environment based on the sensor device group;

[0070] A first extraction module, configured to extract features from the basic information of the object using a deep learning algorithm to generate an object feature set;

[0071] A first construction module is used to construct an environment map based on the object feature set, and determine the real-time environment information according to the environment map.

[0072] Furthermore, the system also includes:

[0073] A path generation module, the path generation module is used to generate a plurality of movement key points based on the target task, and generate the motion path of the target humanoid robot according to the plurality of movement key points;

[0074] a first matching module, configured to sequentially match a first moving key point, a second moving key point ... an Nth moving key point in the motion path according to the initial posture information of the target humanoid robot to generate first posture information, second posture information ... an Nth posture information;

[0075] a first calculation module, configured to calculate, by the dynamic control unit, expected posture information of the first posture information, the second posture information, ..., the Nth posture information at corresponding time steps, and generate an expected posture change data set;

[0076] A first evaluation module is configured to perform stability evaluation based on the expected posture change data set to generate the first posture stability coefficient.

[0077] Furthermore, the system also includes:

[0078] a variation range acquisition module, the variation range acquisition module being used to respectively acquire a first variation range of the first posture stability coefficient and a second variation range of the second posture stability coefficient;

[0079] a first judgment module, configured to identify a boundary extreme value of the first variation range and determine whether the boundary extreme value of the second variation range is included in the boundary extreme value of the first variation range;

[0080] The second judgment module is used for adjusting the second change range according to the first change range if no, and generating the adjustment range data.

[0081] Furthermore, the system also includes:

[0082] A second building module, the second building module is used to use simulation software to integrate the real-time environment information to build a virtual simulation environment;

[0083] a first control module, configured to control the posture data of the target humanoid robot according to the adjustment amplitude data based on the virtual simulation environment;

[0084] a second evaluation module, the second evaluation module being used to define evaluation indicators of the target humanoid robot, evaluate the posture data according to the evaluation indicators, and generate a posture evaluation result;

[0085] A training module is used to adjust the posture of the target humanoid robot according to the posture evaluation result to perform simulation training and generate the posture simulation training result.

[0086] Furthermore, the system also includes:

[0087] A recording module, the recording module is used to record the posture simulation training results and generate a training record log;

[0088] a second calculation module, configured to calculate a posture error of the target humanoid robot according to the training record log to generate a posture deviation data set;

[0089] a correction module, the correction module being used to perform deviation correction on the posture deviation data set to generate posture correction information;

[0090] A correction module is used to correct the posture simulation training result based on the posture correction information and update the posture simulation training result.

[0091] Furthermore, the system also includes:

[0092] A factor retrieval module, wherein the factor retrieval module is used to retrieve the correlation factor between the environmental information set and the target humanoid robot posture;

[0093] a second extraction module, configured to extract first environmental information corresponding to the posture of the target humanoid robot in the posture simulation training result based on the correlation factor;

[0094] a second matching module, configured to match historical posture data of a target humanoid robot based on the first environmental information;

[0095] A strategy formulation module is used to formulate the posture control strategy according to the first environmental information and the historical posture data.

[0096] Through the above-mentioned detailed description of the posture intelligent control method of a humanoid robot in this specification, those skilled in the art can clearly understand the posture intelligent control system of a humanoid robot in this embodiment. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0097] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for intelligent posture control of a humanoid robot, characterized in that: The method is applied to a posture intelligent control system of a humanoid robot, wherein the posture intelligent control system of the humanoid robot is communicatively connected with a dynamics control unit, an inertial measurement unit, and a sensor device group, and the method comprises: Extract the real-time environmental information of the target humanoid robot for intelligent interaction and determine the initial posture information of the target humanoid robot; Using the dynamics control unit to perform posture change planning on the initial posture information of the target humanoid robot according to the motion path to generate a first posture stability coefficient; sensing the real-time posture information of the target humanoid robot in real time through the inertial measurement unit to generate a second posture stability coefficient; comparing the second posture stability coefficient with the first posture stability coefficient to generate adjustment amplitude data; Constructing a virtual simulation environment based on the real-time environmental information, performing simulation training on the posture control of the target humanoid robot based on the adjustment amplitude data through the virtual simulation environment, and generating a posture simulation training result; Formulate a posture control strategy based on the posture simulation training results, and perform intelligent posture control on the target humanoid robot according to the posture control strategy; The method includes: performing posture change planning on the initial posture information of the target humanoid robot according to the motion path using the dynamic control unit to generate a first posture stability coefficient. generating a plurality of movement key points based on the target task, and generating the motion path of the target humanoid robot according to the plurality of movement key points; sequentially matching a first moving key point, a second moving key point ... an Nth moving key point in the motion path according to the initial posture information of the target humanoid robot to generate first posture information, second posture information ... an Nth posture information; Calculating, by the dynamic control unit, expected posture information of the first posture information, the second posture information, ..., the Nth posture information at corresponding time steps to generate an expected posture change data set; A stability assessment is performed based on the expected posture change data set to generate the first posture stability coefficient.

2. The method according to claim 1, wherein Extracting real-time environmental information of the target humanoid robot for intelligent interaction and determining initial posture information of the target humanoid robot, the method includes: Perceiving and collecting basic information of objects in the target humanoid robot environment according to the sensor device group; Using a deep learning algorithm to extract features from the basic information of the object to generate an object feature set; An environment map is constructed based on the object feature set, and the real-time environment information is determined according to the environment map.

3. The method according to claim 1, wherein Comparing the second posture stability coefficient with the first posture stability coefficient to generate adjustment amplitude data, the method includes: respectively acquiring a first variation range of the first posture stability coefficient and a second variation range of the second posture stability coefficient; Identifying the boundary extreme value of the first change range, and determining whether the boundary extreme value of the second change range is included in the boundary extreme value of the first change range; If not, the second change range is adjusted according to the first change range to generate the adjustment range data.

4. The method according to claim 1, wherein Constructing a virtual simulation environment based on the real-time environmental information, performing simulation training on the posture control of the target humanoid robot based on the adjustment amplitude data through the virtual simulation environment, and generating a posture simulation training result, the method comprising: Using simulation software to integrate the real-time environment information to build a virtual simulation environment; Controlling the posture data of the target humanoid robot according to the adjustment amplitude data based on the virtual simulation environment; defining evaluation indicators of the target humanoid robot, evaluating the posture data according to the evaluation indicators, and generating a posture evaluation result; The posture of the target humanoid robot is adjusted according to the posture evaluation result to perform simulation training, and the posture simulation training result is generated.

5. The method according to claim 4, wherein Methods include: Generate a training record log by recording the posture simulation training results; Calculating a posture error of the target humanoid robot according to the training record log to generate a posture deviation data set; performing deviation correction on the posture deviation data set to generate posture correction information; The posture simulation training result is corrected based on the posture correction information, and the posture simulation training result is updated.

6. The method according to claim 1, wherein Formulating a posture control strategy based on the posture simulation training results, the method includes: Retrieving the correlation factor between the environmental information set and the target humanoid robot posture; Extracting first environmental information corresponding to the posture of the target humanoid robot in the posture simulation training result based on the correlation factor; matching historical posture data of the target humanoid robot based on the first environmental information; The posture control strategy is formulated according to the first environmental information and the historical posture data.

7. An intelligent posture control system for a humanoid robot, characterized in that: The humanoid robot's posture intelligent control system is communicatively connected to a dynamics control unit, an inertial measurement unit, and a sensor device group, and the system includes: An intelligent interaction module, the intelligent interaction module is used to extract real-time environmental information of the target humanoid robot for intelligent interaction and determine initial posture information of the target humanoid robot; a planning module, the planning module being configured to perform posture change planning on the initial posture information of the target humanoid robot according to a motion path using the dynamics control unit to generate a first posture stability coefficient; a first perception module, configured to perceive real-time posture information of the target humanoid robot in real time through the inertial measurement unit, and generate a second posture stability coefficient; a comparison module, configured to compare the second posture stability coefficient with the first posture stability coefficient to generate adjustment amplitude data; A simulation training module, the simulation training module is used to construct a virtual simulation environment based on the real-time environment information, perform simulation training on the posture control of the target humanoid robot based on the adjustment amplitude data through the virtual simulation environment, and generate a posture simulation training result; An intelligent control module, the intelligent control module being used to formulate a posture control strategy based on the posture simulation training results, and to perform intelligent posture control on the target humanoid robot according to the posture control strategy; The system also includes: A path generation module, the path generation module is used to generate a plurality of movement key points based on the target task, and generate the motion path of the target humanoid robot according to the plurality of movement key points; a first matching module, configured to sequentially match a first moving key point, a second moving key point ... an Nth moving key point in the motion path according to the initial posture information of the target humanoid robot to generate first posture information, second posture information ... an Nth posture information; a first calculation module, configured to calculate, by the dynamic control unit, expected posture information of the first posture information, the second posture information, ..., the Nth posture information at corresponding time steps, and generate an expected posture change data set; A first evaluation module is configured to perform stability evaluation based on the expected posture change data set to generate the first posture stability coefficient.

Images