Method, device, medium, equipment and product for determining factors causing robot falls
By constructing the error vector and reviewing the data set, selecting the convex vertex set to calculate the error loss score, the problem of difficult to determine the fall factors of the robot is solved, the efficiency and accuracy of factor determination are improved, and the robot control algorithm is optimized to prevent falls.
Patent Information
- Application Number
- CN202410911650.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-09
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2044-07-09
AI Technical Summary
The prior art is difficult to effectively determine the factors that cause drops by robots, which affects robot algorithm improvement and system optimization.
By constructing an error vector based on the actual and expected motion data of the robot controlled unit, a review data set is constructed, a convex vertex set is selected, and an error loss score is calculated to determine the fall factor.
Improves the efficiency and accuracy of the determination of fall factors, helps optimize control algorithms and prevents robots from falling.
Smart Images

Figure CN118456474B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of robotics technology, and more specifically, to a method, device, medium, equipment and product for determining factors that cause a robot to fall. Background Art
[0002] At present, cutting-edge applications in many fields such as industry, medicine and deep space exploration are closely related to robots. A humanoid robot is a robot system that imitates human appearance, behavioral characteristics and biological mechanisms. Bipedal walking is a key technology for humanoid robots to achieve free movement in real environments. Stable bipedal walking is closely related to the real-time feedback of each controlled unit of the robot and the precise control of multiple joints. However, humanoid robots often have ultra-high degrees of freedom and controlled units of multiple different modes. When a humanoid robot falls due to system factors, how to trace the cause of the fall is crucial to the improvement of robot algorithms and system optimization.
[0003] Therefore, how to determine the factors that cause robots to fall has become a technical problem that the industry urgently needs to solve. Summary of the invention
[0004] The present application provides a method, device, medium, equipment and product for determining factors that cause a robot to fall, so as to solve the technical problem of how to determine factors that cause a robot to fall in the prior art.
[0005] In a first aspect, the present application provides a method for determining factors causing a robot to fall, comprising:
[0006] Based on the actual motion data and expected motion data of each controlled unit of the robot that currently falls, an error vector of each controlled unit is obtained;
[0007] Constructing a current review data set of the robot based on the currently set fall review moment and the error vector;
[0008] constructing an initial convex vertex set based on the distances between the review vectors in the current review data set;
[0009] The initial convex vertex set is updated based on the convex hull distance between each review vector in the current review data set and the initial convex vertex set to obtain a convex vertex set;
[0010] The fall-causing factor of the robot is determined based on the error loss score of the controlled unit corresponding to each review vector in the convex vertex set.
[0011] In some embodiments, constructing an initial convex vertex set based on the distances between the review vectors in the current review data set includes:
[0012] Randomly selecting a first vector from the current review data set;
[0013] Obtain a second vector in the current review data set that is farthest from the first vector;
[0014] Obtaining a third vector in the current review data set that is farthest from the second vector;
[0015] The initial convex vertex set is constructed based on the second vector and the third vector.
[0016] In some embodiments, the updating of the initial convex vertex set based on the convex hull distance between each review vector in the current review data set and the initial convex vertex set includes:
[0017] An optimization function for setting parameters to be optimized based on the current initial convex vertex set;
[0018] Obtaining an optimal solution for the parameter to be optimized based on the optimization function;
[0019] Based on the optimal solution, obtain the convex hull distance between each review vector in the current review data set and the current initial convex vertex set;
[0020] Adding the review vector having the largest convex hull distance from the current initial convex vertex set to the current initial convex vertex set to obtain a new initial convex vertex set;
[0021] The current initial convex vertex set is updated based on the new initial convex vertex set until the number of vectors in the current initial convex vertex set is equal to the preset number of convex hull vertices.
[0022] In some embodiments, the optimization function is as follows:
[0023] ;
[0024] in, ; ; is the parameter to be optimized; is the current initial convex vertex set; is the first A review vector.
[0025] In some embodiments, after obtaining the convex vertex set, the method further includes:
[0026] Set the set of control error loss scores in descending order;
[0027] Based on the acquisition order of each review vector in the convex vertex set, the error loss score in the control error loss score set is sequentially assigned to each review vector in the convex vertex set to obtain the error loss score of each review vector in the convex vertex set.
[0028] In some embodiments, determining the fall-causing factor of the robot based on the error loss score of the controlled unit corresponding to each review vector in the convex vertex set includes:
[0029] Set different fall review moments;
[0030] Obtain each error loss score of the controlled unit corresponding to the convex vertex concentrated review vector at different fall review moments;
[0031] Determining the error loss score of each controlled unit based on the cumulative score of each error loss score;
[0032] The controlled unit whose error loss score is greater than or equal to a preset threshold is regarded as a fall-causing factor of the robot.
[0033] In a second aspect, the present application provides a device for determining factors causing a robot to fall, comprising:
[0034] An error module, used for obtaining an error vector of each controlled unit based on actual motion data and expected motion data of each controlled unit of the robot that currently falls;
[0035] A review module, configured to construct a current review data set of the robot based on a currently set fall review moment and the error vector;
[0036] A construction module, configured to construct an initial convex vertex set based on distances between review vectors in the current review data set;
[0037] An updating module, configured to update the initial convex vertex set based on the convex hull distance between each review vector in the current review data set and the initial convex vertex set to obtain a convex vertex set;
[0038] A determination module is used to determine the fall-causing factor of the robot based on the error loss score of the controlled unit corresponding to each review vector in the convex vertex set.
[0039] In a third aspect, the present application provides a non-transitory computer-readable storage medium having a computer program stored thereon, which implements the above method when executed by a processor.
[0040] In a fourth aspect, the present application provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor is configured to implement the above method when executing the program through the computer program.
[0041] In a fifth aspect, the present application also provides a computer program product, including a computer program, which implements the above method when executed by a processor.
[0042] The present application provides a method, apparatus, medium, equipment and product for determining factors that cause a robot to fall. The error vectors of each controlled unit can be constructed through the motion data of the controlled units of the fallen robot, a review data set can be constructed through the error vector and the review time, and a convex vertex set can be obtained by selecting the review vectors in the review data set. The factors that cause the robot to fall can be determined by the error loss scores of the controlled units corresponding to each review vector in the convex vertex set, thereby improving the efficiency and accuracy of determining the factors that cause the robot to fall, so that the control algorithms of these controlled units can be further optimized according to the factors that cause the robot to fall to improve the control of the robot and prevent the robot from falling. BRIEF DESCRIPTION OF THE DRAWINGS
[0043] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.
[0044] In order to more clearly illustrate the technical solutions in the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0045] Figure 1 A schematic diagram of a flow chart of a method for determining factors causing a robot to fall provided in an embodiment of the present application;
[0046] Figure 2 A schematic diagram of the structure of a device for determining factors causing a robot to fall provided in an embodiment of the present application;
[0047] Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0048] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.
[0049] It should be noted that the terms "first", "second" etc. in the present application are used to distinguish similar objects, and need not be used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable in appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, the process, method, system, product or equipment comprising a series of steps or modules need not be limited to those steps or modules clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products or equipment.
[0050] The method for determining robot fall factors provided in the embodiments of the present application is applicable to a terminal, which may be a variety of electronic devices having a display screen and supporting web browsing, including but not limited to servers, smart phones, tablet computers, laptop computers, and desktop computers.
[0051] In the technical solution of this application, the collection, storage, use, processing, transmission, provision and disclosure of the information involved comply with the provisions of relevant laws and regulations, take necessary confidentiality measures, and do not violate public order and good morals.
[0052] Figure 1 A schematic diagram of a flow chart of a method for determining a robot fall-causing factor provided in an embodiment of the present application, such as Figure 1 As shown, the method includes step 110, step 120, step 130, step 140 and step 150. The steps of the method flow are only a possible implementation of the present application.
[0053] Step 110: Based on the actual motion data and expected motion data of each controlled unit of the robot that currently falls, obtain the error vector of each controlled unit.
[0054] Specifically, the robot fall factor determination method provided in the embodiment of the present application is executed by a robot fall factor determination device, which can be a hardware device independently set in the terminal, or a software program running in the terminal. For example, when the terminal is a mobile phone, the robot fall factor determination device can be embodied as an application such as detection software in the mobile phone.
[0055] The robot is a robot that can walk, such as a bipedal or quadrupedal legged robot.
[0056] The controlled units are the various components or subsystems in the robot, such as joints.
[0057] Motion data refers to the state or position information of the controlled unit in actual motion, such as joint angles, sensor readings, etc.
[0058] The error vector is the difference between the actual motion data and the expected motion data and can be represented as a vector.
[0059] For example, the sensor feedback matrix of all controlled units of the robot is ,in, It's the robot The feedback value of the sensor of each controlled unit during the robot's walking. The sensor can be an angle sensor, a current sensor, an inertial sensor, etc. Indicates the number of controlled units; Indicates the length of time, Represents a vector.
[0060] The data of the sensor feedback matrix is the actual motion data of the controlled unit. The system desired state matrix corresponding to the actual motion data is recorded as ,in Indicates that during the control process The data of the system expected state matrix is the expected motion data of the controlled unit.
[0061] From this, the control error data set of all controlled units can be calculated , the calculation formula is as follows:
[0062] ;
[0063] The control error data set includes the error vector of each controlled unit.
[0064] After obtaining the control error data set, the control error data set can be normalized.
[0065] For example, the control error vector The minimum value in , the maximum value is , for the vector Normalization is performed, and the processing formula is as follows:
[0066] ;
[0067] in, ,remember is the normalized control error data set, The code of the controlled unit.
[0068] Step 120: construct a current review data set of the robot based on the currently set fall review time and error vector.
[0069] Specifically, the fall review time refers to the time point from which the robot starts to trace back after falling. The fall review time can be set according to the actual situation, and different fall review times can be set to construct different review data sets.
[0070] The retrospective vector includes the error vector elements of the controlled unit within the retrospective time starting from the retrospective moment of the fall.
[0071] The current review data set is a set of review vectors of each controlled unit within the review time starting from the currently set fall review moment and tracing back.
[0072] For example, by normalizing the control error dataset Building a retrospective dataset .
[0073] For the The length of a controlled unit is T Vector , let the moment when the robot starts to fall during the experiment be , review time is ,by is the review moment of the fall, then the review vector can be defined for:
[0074] ;
[0075] in, Representation vector Middle time T Elements of the dataset .
[0076] Step 130: construct an initial convex vertex set based on the distances between the review vectors in the current review data set.
[0077] Step 140: Update the initial convex vertex set based on the convex hull distance between each review vector in the current review data set and the initial convex vertex set to obtain a convex vertex set.
[0078] Specifically, the retrospective dataset is calculated The two most distant review vectors and , take the two furthest-distance review vectors as the initial convex vertex set , .
[0079] Calculate the convex hull distance between each review vector in the current review data set and the initial convex vertex set, and gradually add the new review vector with the largest convex hull distance to the initial convex vertex set to the initial convex vertex set, so as to continuously update the initial convex vertex set and finally obtain the convex vertex set. The points corresponding to the review vectors in the convex vertex set can be understood as abnormal points.
[0080] Step 150: Determine the fall-causing factor of the robot based on the error loss score of the controlled unit corresponding to each review vector in the convex vertex set.
[0081] Specifically, the error loss score is used to measure the performance of the controlled unit during the current movement. The higher the score, the worse the performance of the controlled unit, and the more likely it is that the controlled unit is the main factor causing the robot to fall.
[0082] Fall-causing factors refer to controlled units that may cause the robot to fall.
[0083] The error loss score of the controlled unit corresponding to each review vector can be set according to the order in which each review vector is added in the convex vertex set.
[0084] The controlled units whose error loss scores are greater than or equal to the preset threshold can be regarded as the factors that cause the current robot to fall. After finding these controlled units that may cause the robot to fall, the control algorithms of these controlled units can be optimized to improve the control of the robot and prevent the robot from falling.
[0085] The method for determining factors that cause a robot to fall provided in an embodiment of the present application can construct error vectors of each controlled unit through the motion data of the controlled units of the fallen robot, construct a review data set through the error vector and the review time, and select the review vector in the review data set to obtain a convex vertex set. The factors that cause the robot to fall can be determined by the error loss scores of the controlled units corresponding to each review vector in the convex vertex set, thereby improving the efficiency and accuracy of determining the factors that cause the robot to fall, so that the control algorithms of these controlled units can be further optimized according to the factors that cause the robot to fall to improve the control of the robot and prevent the robot from falling.
[0086] It should be noted that each implementation method of the present application can be freely combined, the order can be changed, or it can be executed separately, and does not need to rely on or depend on a fixed execution order.
[0087] In some embodiments, step 130 includes:
[0088] Randomly select a first vector in the current review data set;
[0089] Get the second vector farthest from the first vector in the current review data set;
[0090] Get the third vector farthest from the second vector in the current review data set;
[0091] Construct an initial convex vertex set based on the second vector and the third vector.
[0092] Specifically, in the current review dataset Randomly select a review vector from , review vector This is the first vector. In the current review dataset Find the distance The furthest look back vector , review vector This process can be expressed as the following formula:
[0093] ;
[0094] in, Represents the calculation of the bi-norm of the review vector.
[0095] In the current review dataset Find the distance The furthest look back vector , review vector This process can be expressed as the following formula:
[0096] ;
[0097] remember is the initial convex vertex set. The second and third vectors represent the extreme cases in the data set and can cover the boundaries and variation range of the data.
[0098] The method for determining robot fall-causing factors provided in the embodiment of the present application constructs an initial convex vertex set through the distance between review vectors in the current review data set, effectively utilizes geometric characteristics to select abnormal vectors, and improves the efficiency of determining fall-causing factors.
[0099] In some embodiments, step 140 includes:
[0100] An optimization function for setting parameters to be optimized based on the current initial convex vertex set;
[0101] Obtain the optimal solution of the parameters to be optimized based on the optimization function;
[0102] Based on the optimal solution, obtain the convex hull distance between each review vector in the current review data set and the current initial convex vertex set;
[0103] Add the review vector with the largest convex hull distance from the current initial convex vertex set to the current initial convex vertex set to obtain a new initial convex vertex set;
[0104] The current initial convex vertex set is updated based on the new initial convex vertex set until the number of vectors in the current initial convex vertex set is equal to the preset number of convex hull vertices.
[0105] The optimization function is as follows:
[0106] ;
[0107] in, ; ; is the parameter to be optimized; is the current initial convex vertex set; The first A review vector.
[0108] Specifically, the current review data set is calculated one by one The review vector and the initial convex vertex set The convex hull distance between them, select the review vector with the largest convex hull distance , updated to the initial convex vertex set.
[0109] In order to calculate the review vector in the current review dataset With the initial convex vertex set The convex hull distance between them requires solving a constrained quadratic optimization function, which is as follows:
[0110] ;
[0111] ;
[0112] in, Parameters to be optimized Represents a set of linear combination coefficients whose sum is 1, and its mathematical expression is the constraint condition .
[0113] According to the mathematical definition of the convex hull, any vector in the convex hull can be obtained by passing a set of parameters from the points in the convex hull. Linear representation, that is can represent any point in the convex hull. Then the vector The convex hull distance is to find the distance vector in the convex hull The nearest point and calculate its Euclidean distance, expressed as , after arranging, we can get the above optimization function.
[0114] Obtain the optimal solution of the parameters to be optimized by calculating the optimization function , calculate the look-back vector With the current initial convex vertex set The convex hull distance between , the calculation formula is as follows:
[0115] ;
[0116] Select the current review dataset middle The maximum review vector is ,Will Add the initial convex vertex set Update it in . Recalculate the optimization function according to the updated initial convex vertex set to obtain the new optimal solution of the parameters to be optimized, and obtain the new , and update the initial convex vertex set again, repeat this step until the number of vectors of the review vectors in the initial convex vertex set reaches the preset number of convex hull vertices W, and the final convex vertex set is obtained The preset number of convex hull vertices can be set according to actual conditions.
[0117] The method for determining the factors that cause a robot to fall provided in the embodiment of the present application can obtain abnormal review vectors by constructing a convex vertex set, thereby further determining the factors that cause the robot to fall, thereby improving the efficiency of determining the factors that cause a robot to fall.
[0118] In some embodiments, after step 140, the method further includes:
[0119] Set the set of control error loss scores in descending order;
[0120] Based on the acquisition order of each review vector in the convex vertex set, the error loss score in the control error loss score set is assigned to each review vector in the convex vertex set in turn to obtain the error loss score of each review vector in the convex vertex set.
[0121] Step 150 includes:
[0122] Set different fall review moments;
[0123] Obtain each error loss score of the controlled unit corresponding to the convex vertex concentrated review vector at different fall review moments;
[0124] Determining the error loss score of each controlled unit based on the cumulative score of each error loss score;
[0125] The controlled units whose error loss scores are greater than or equal to the preset threshold are regarded as the factors causing the robot to fall.
[0126] Specifically, the control error loss score may be performed on the controlled units corresponding to the elements located in the convex vertex set according to the order in which the convex hull vertices are selected.
[0127] You can set the control error loss score set in descending order from large to small , is the specific error loss fraction. The error loss fraction can be set according to the user's requirements for system error.
[0128] The earlier the review vector is obtained in the convex vertex set, the more abnormal it is. Therefore, the error loss score in the error loss score set can be assigned to the controlled unit corresponding to each review vector in the convex vertex set according to the order in which each review vector is obtained in the convex vertex set, and the control error loss score of all controlled units of the robot system can be obtained. , . is the error loss fraction of each controlled unit.
[0129] Different fall review moments will correspond to different review data sets, and thus may correspond to different convex vertex sets. Therefore, different fall review moments can be selected, for example, This is the new fall review moment, and the review duration is still , the current review vector is updated as:
[0130] ;
[0131] By analogy, different fall review moments are set. Repeat multiple times to obtain the error loss scores of the controlled units corresponding to each review vector in the convex vertex set at different fall review moments, and accumulate these error loss scores to obtain the final error loss score of the controlled unit. ; is the number of repetitions.
[0132] Error loss fraction for each controlled unit Arrange them in descending order. The controlled units at the front are the ones with the largest system control error or control error change rate before the robot falls. They are possible factors that cause the robot to fall. R&D personnel can further optimize the algorithm and improve system design and control based on this ranking.
[0133] The method for determining robot fall-causing factors provided in the embodiment of the present application can reduce the score error and improve the accuracy of determining the robot fall-causing factors by obtaining the accumulated error loss score of each controlled unit.
[0134] In some embodiments, the method for determining a robot fall-causing factor provided in an embodiment of the present application includes:
[0135] Step 1: Calculate the control error data set of all controlled units of the robot.
[0136] Step 2: Normalize the control error data set.
[0137] Step 3: Construct a review dataset based on the normalized control error dataset.
[0138] Step 4: Calculate the two vectors with the farthest distance in the review data set and construct the initial convex vertex set.
[0139] Step 5: successively calculate the convex hull distance between the review vectors in the review data set and the initial convex vertex set, select the vector with the largest convex hull distance and add it to the initial convex vertex set to update the initial convex vertex set.
[0140] Repeat step 5 until the number of vectors in the initial convex vertex set reaches the preset number of convex hull vertices to obtain a convex vertex set.
[0141] Step 6: According to the order in which the convex hull vertices in the convex vertex set are selected, the control error loss score is performed on the controlled unit corresponding to the review vector located in the convex vertex set.
[0142] Step 7: Set a new fall review time, repeat steps 3 to 6, cumulatively update the error loss score of the controlled unit, repeat multiple times, and obtain the final error loss score of the controlled unit.
[0143] Step 8. Arrange the error loss scores of the controlled units finally obtained in descending order. The controlled units at the top are the ones with the largest system control error or control error change rate before the robot falls. They are possible factors that cause the robot to fall. R&D personnel can further optimize the algorithm and improve system design and control based on this ranking.
[0144] The method for determining robot fall-causing factors provided in the embodiment of the present application analyzes the high-dimensional feature space formed by the sensor and control data fed back by the robot system, performs statistical analysis on the features located at the vertices of the convex hull of the data set before the fall, and lists the controlled units that may cause the humanoid robot to fall while walking based on the cumulative scores, so as to assist R&D personnel in conducting analysis, accelerate attribution analysis, and improve the efficiency of algorithm optimization and system improvement.
[0145] The following is a description of a robot fall-causing factor determination device provided in an embodiment of the present application. The robot fall-causing factor determination device described below and the robot fall-causing factor determination method described above can be referenced to each other.
[0146] Figure 2 A schematic diagram of the structure of a device for determining a robot fall-causing factor provided in an embodiment of the present application, such as Figure 2As shown, the apparatus includes an error module 210 , a review module 220 , a construction module 230 , an update module 240 and a determination module 250 .
[0147] An error module, used to obtain an error vector of each controlled unit based on actual motion data and expected motion data of each controlled unit of the robot that currently falls;
[0148] A review module, used to construct a current review data set of the robot based on the currently set fall review moment and error vector;
[0149] A construction module for constructing an initial convex vertex set based on the distances between review vectors in the current review data set;
[0150] An updating module, used for updating the initial convex vertex set based on the convex hull distance between each review vector in the current review data set and the initial convex vertex set to obtain a convex vertex set;
[0151] The determination module is used to determine the fall-causing factor of the robot based on the error loss score of the controlled unit corresponding to each review vector in the convex vertex set.
[0152] Specifically, according to an embodiment of the present application, any multiple modules among the error module, the review module, the construction module, the update module and the determination module can be combined into one module for implementation, or any one of the modules can be split into multiple modules.
[0153] Alternatively, at least part of the functionality of one or more of these modules may be combined with at least part of the functionality of other modules and implemented in one module.
[0154] According to an embodiment of the present application, at least one of the error module, review module, building module, update module and determination module can be at least partially implemented as a hardware circuit, such as a field programmable gate array, a programmable logic array, a system on a chip, a system on a substrate, a system on a package, an application-specific integrated circuit, or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or can be implemented in any one of the three implementation methods of software, hardware and firmware, or in an appropriate combination of any of them.
[0155] Alternatively, at least one of the error module, the review module, the construction module, the update module and the determination module may be at least partially implemented as a computer program module, and when the computer program module is executed, the corresponding function may be performed.
[0156] The device for determining factors causing a robot to fall provided in the embodiment of the present application can construct error vectors of each controlled unit through the motion data of the controlled units of the fallen robot, construct a review data set through the error vector and the review time, and select the review vector in the review data set to obtain a convex vertex set, and the factors causing the robot to fall can be determined through the error loss scores of the controlled units corresponding to each review vector in the convex vertex set, thereby improving the efficiency and accuracy of determining the factors causing the robot to fall, so that the control algorithms of these controlled units can be further optimized according to the factors causing the robot to fall to improve the control of the robot and prevent the robot from falling.
[0157] In some embodiments, the building blocks are specifically used to:
[0158] Randomly select a first vector in the current review data set;
[0159] Get the second vector farthest from the first vector in the current review data set;
[0160] Get the third vector farthest from the second vector in the current review data set;
[0161] Construct an initial convex vertex set based on the second vector and the third vector.
[0162] In some embodiments, the update module is specifically used to:
[0163] An optimization function for setting parameters to be optimized based on the current initial convex vertex set;
[0164] Obtain the optimal solution of the parameters to be optimized based on the optimization function;
[0165] Based on the optimal solution, obtain the convex hull distance between each review vector in the current review data set and the current initial convex vertex set;
[0166] Add the review vector with the largest convex hull distance from the current initial convex vertex set to the current initial convex vertex set to obtain a new initial convex vertex set;
[0167] The current initial convex vertex set is updated based on the new initial convex vertex set until the number of vectors in the current initial convex vertex set is equal to the preset number of convex hull vertices.
[0168] In some embodiments, the optimization function is as follows:
[0169] ;
[0170] in, ; ; is the parameter to be optimized; is the current initial convex vertex set; The first A review vector.
[0171] In some embodiments, the robot fall-causing factor determination device further includes a value assignment module, and the value assignment module is specifically used to:
[0172] Set the set of control error loss scores in descending order;
[0173] Based on the acquisition order of each review vector in the convex vertex set, the error loss score in the control error loss score set is assigned to each review vector in the convex vertex set in turn to obtain the error loss score of each review vector in the convex vertex set.
[0174] In some embodiments, the determination module is specifically configured to:
[0175] Set different fall review moments;
[0176] Obtain each error loss score of the controlled unit corresponding to the convex vertex concentrated review vector at different fall review moments;
[0177] Determining the error loss score of each controlled unit based on the cumulative score of each error loss score;
[0178] The controlled units whose error loss scores are greater than or equal to the preset threshold are regarded as the factors causing the robot to fall.
[0179] It should be noted here that the robot fall-causing factor determination device provided in the embodiment of the present application can implement all the method steps implemented in the above-mentioned robot fall-causing factor determination method embodiment, and can achieve the same technical effect. The parts and beneficial effects of this embodiment that are the same as those in the method embodiment will not be described in detail here.
[0180] Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application, such as Figure 3 As shown, the electronic device may include: a processor (Processor) 310, a communication interface (Communications Interface) 320, a memory (Memory) 330 and a communication bus (Communications Bus) 340, wherein the processor 310, the communication interface 320, and the memory 330 communicate with each other through the communication bus 340. The processor 310 may call the logic command in the memory 330 to execute the above method, which includes:
[0181] Based on the actual motion data and expected motion data of each controlled unit of the robot that currently falls, the error vector of each controlled unit is obtained;
[0182] Construct the robot's current review data set based on the currently set fall review moment and error vector;
[0183] Construct an initial convex vertex set based on the distances between review vectors in the current review dataset;
[0184] The initial convex vertex set is updated based on the convex hull distance between each review vector in the current review data set and the initial convex vertex set to obtain a convex vertex set;
[0185] The fall-causing factors of the robot are determined based on the error loss scores of the controlled units corresponding to each review vector in the convex vertex set.
[0186] In addition, the logical commands in the above-mentioned memory can be implemented in the form of software function modules and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several commands to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc., and other media that can store program codes.
[0187] The processor in the electronic device provided in the embodiment of the present application can call the logic instructions in the memory to implement the above method. Its specific implementation method is consistent with the implementation method of the aforementioned method and can achieve the same beneficial effects, which will not be repeated here.
[0188] An embodiment of the present application also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the methods provided in the above embodiments are implemented.
[0189] Its specific implementation is consistent with the aforementioned method implementation and can achieve the same beneficial effects, so it will not be repeated here.
[0190] An embodiment of the present application provides a computer program product, including a computer program, which implements the above method when executed by a processor.
[0191] The device embodiments described above are merely illustrative, wherein the modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules, i.e., they may be located in one place, or they may be distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative effort.
[0192] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.
[0193] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.
Claims
1. A method for determining factors that cause a robot to fall, characterized in that: include: Based on the actual motion data and expected motion data of each controlled unit of the robot that currently falls, an error vector of each controlled unit is obtained; Constructing a current review data set of the robot based on the currently set fall review moment and the error vector; constructing an initial convex vertex set based on the distances between the review vectors in the current review data set; The initial convex vertex set is updated based on the convex hull distance between each review vector in the current review data set and the initial convex vertex set to obtain a convex vertex set; Determining the fall-causing factor of the robot based on the error loss score of the controlled unit corresponding to each review vector in the convex vertex set; The updating of the initial convex vertex set based on the convex hull distance between each review vector in the current review data set and the initial convex vertex set comprises: An optimization function for setting parameters to be optimized based on the current initial convex vertex set; Obtaining an optimal solution for the parameter to be optimized based on the optimization function; Based on the optimal solution, obtain the convex hull distance between each review vector in the current review data set and the current initial convex vertex set; Adding the review vector having the largest convex hull distance from the current initial convex vertex set to the current initial convex vertex set to obtain a new initial convex vertex set; The current initial convex vertex set is updated based on the new initial convex vertex set until the number of vectors in the current initial convex vertex set is equal to the preset number of convex hull vertices.
2. The method for determining robot fall-causing factors according to claim 1, characterized in that: The constructing an initial convex vertex set based on the distances between the review vectors in the current review data set includes: Randomly selecting a first vector from the current review data set; Obtain a second vector in the current review data set that is farthest from the first vector; Obtaining a third vector in the current review data set that is farthest from the second vector; The initial convex vertex set is constructed based on the second vector and the third vector.
3. The method for determining robot fall-causing factors according to claim 1, characterized in that: The optimization function is as follows: ; in, ; is the parameter to be optimized; is the current initial convex vertex set; is the first A review vector.
4. The method for determining robot fall-causing factors according to claim 1, characterized in that: After obtaining the convex vertex set, the method further includes: Set the set of control error loss scores in descending order; Based on the acquisition order of each review vector in the convex vertex set, the error loss score in the control error loss score set is sequentially assigned to each review vector in the convex vertex set to obtain the error loss score of each review vector in the convex vertex set.
5. The method for determining robot fall-causing factors according to claim 4, characterized in that: The determining of the fall-causing factor of the robot based on the error loss score of the controlled unit corresponding to each review vector in the convex vertex set includes: Set different fall review moments; Obtain each error loss score of the controlled unit corresponding to the convex vertex concentrated review vector at different fall review moments; Determining the error loss score of each controlled unit based on the cumulative score of each error loss score; The controlled unit whose error loss score is greater than or equal to a preset threshold is regarded as a fall-causing factor of the robot.
6. A device for determining factors causing a robot to fall, characterized in that: include: An error module, used for obtaining an error vector of each controlled unit based on actual motion data and expected motion data of each controlled unit of the robot that currently falls; A review module, configured to construct a current review data set of the robot based on a currently set fall review moment and the error vector; A construction module, configured to construct an initial convex vertex set based on distances between review vectors in the current review data set; An updating module, configured to update the initial convex vertex set based on the convex hull distance between each review vector in the current review data set and the initial convex vertex set to obtain a convex vertex set; A determination module, configured to determine a fall-causing factor of the robot based on an error loss score of a controlled unit corresponding to each review vector in the convex vertex set; The updating of the initial convex vertex set based on the convex hull distance between each review vector in the current review data set and the initial convex vertex set comprises: An optimization function for setting parameters to be optimized based on the current initial convex vertex set; Obtaining an optimal solution for the parameter to be optimized based on the optimization function; Based on the optimal solution, obtain the convex hull distance between each review vector in the current review data set and the current initial convex vertex set; Adding the review vector having the largest convex hull distance from the current initial convex vertex set to the current initial convex vertex set to obtain a new initial convex vertex set; The current initial convex vertex set is updated based on the new initial convex vertex set until the number of vectors in the current initial convex vertex set is equal to the preset number of convex hull vertices.
7. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for determining a robot fall-causing factor according to any one of claims 1 to 5 is implemented.
8. An electronic device, comprising a memory and a processor, characterized in that: The memory stores a computer program, and the processor is configured to execute the robot fall-causing factor determination method according to any one of claims 1 to 5 through the computer program.
9. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for determining a robot fall-causing factor according to any one of claims 1 to 5 is implemented.
Citation Information
Patent Citations
Humanoid robot having fall-management capabilities, and method for managing said falls
CN103124669A
Method for detecting tumbling state of humanoid robot based on multi-sensor information
CN104217107A
Moving posture anti-falling protection based robot and control method thereof
CN111113412A
Humanoid robot tumbling prediction method
CN111291831A
Tumbling detection method and system for crawling robot
CN113848965A
