Robot adaptability testing method and platform in multiple scenarios

Abstract

The present application provides a robot adaptability testing method and platform in multiple scenarios, which relates to the field of robot testing technology, including: constructing a simulated fire scene and obtaining its scene information, further dividing the sub-scenes into ruins space, narrow space and smoke environment; based on the scene complexity rating of these sub-scenes, generating basic rating identifiers, and establishing mapping tasks; configuring a monitoring sensor group according to the mapping task, monitoring the robot in real time and obtaining a real-time data set; constructing an evaluation network, including three sub-networks, corresponding to three sub-scenes, and performing adaptability test evaluation; finally, controlling and managing the robot according to the adaptability evaluation results. This application can solve the technical problem that the existing technology lacks consideration of the diversity of environments and tasks, resulting in inaccurate adaptability evaluation of robots in complex and changeable scenarios, improves the comprehensiveness and accuracy of robot evaluation, enhances the adaptability of robots, and optimizes the control and management of robots.

Description

Technical Field

[0001] The present application relates to the field of robot testing technology, and in particular to a robot adaptability testing method and platform in multiple scenarios. Background Art

[0002] With the high-level development of artificial intelligence, computer science, electronic technology and control theory, robotics has made remarkable progress. The integration of these technologies has promoted the rapid development of intelligent robotics, enabling robots to have stronger perception, decision-making and execution capabilities. At the same time, robotics is not only widely used in traditional fields such as manufacturing and agriculture, but also shows great potential in medical care, education, services and other fields. In multiple scenarios, the adaptability of robots has become one of the important indicators to measure their performance. Adaptability testing aims to evaluate the performance of robots in different environments and conditions to ensure that they can operate stably and reliably in various complex environments. Through adaptability testing, problems and deficiencies in robot design can be discovered, and then optimized and improved. This is of great significance to improving the overall performance of robots and expanding their scope of application.

[0003] At present, existing technical solutions have limitations in many aspects. Specifically, when dealing with complex and changing environments and tasks, existing technologies often use a single evaluation standard and method, which lacks pertinence and flexibility. At the same time, the collection and use of real-time monitoring data are also insufficient, resulting in an incomplete and inaccurate evaluation of robot performance.

[0004] In summary, the existing technology adopts a single evaluation standard and method and lacks consideration of the diversity of environments and tasks, resulting in inaccurate evaluation of the robot's adaptability in complex and changing scenarios, which further affects the technical issues of robot control management and overall performance improvement. Summary of the invention

[0005] The purpose of this application is to provide a robot adaptability testing method and platform in multiple scenarios, so as to solve the technical problem that the existing technology lacks consideration of the diversity of environments and tasks, resulting in inaccurate robot adaptability evaluation in complex and changeable scenarios, improve the comprehensiveness and accuracy of robot evaluation, enhance the adaptability of the robot, and optimize the control and management of the robot.

[0006] In view of the above problems, the present application provides a robot adaptability testing method and platform in multiple scenarios.

[0007] In the first aspect, the present application provides a robot adaptability testing method in multiple scenarios, which is implemented by a robot adaptability testing platform in multiple scenarios, wherein the robot adaptability testing method in multiple scenarios includes: establishing a simulated fire scene, and obtaining scene information of the simulated fire scene, dividing the sub-scenes according to the scene information, and generating a ruin space sub-scene, a narrow space sub-scene, and a smoke environment sub-scene; establishing a scene complexity rating for each sub-scene based on the scene information, and generating a basic rating identifier; establishing a mapping task for each sub-scene, and configuring a monitoring sensor group according to the mapping task, and using the monitoring sensor group to perform real-time monitoring of the robot; The method comprises the following steps: measuring, obtaining a real-time data set of the robot; constructing an evaluation network, wherein the evaluation network comprises a first evaluation sub-network, a second evaluation sub-network and a third evaluation sub-network, wherein the first evaluation sub-network, the second evaluation sub-network and the third evaluation sub-network are respectively mapped to a ruin space sub-scene, a narrow space sub-scene and a smoke environment sub-scene, and the first evaluation sub-network, the second evaluation sub-network and the third evaluation sub-network are provided with different bias weights; sending the real-time data set, the basic rating identifier and the mapping task to the evaluation network, performing adaptability test and evaluation of the robot in multiple scenes through each sub-evaluation network in the evaluation network, and generating adaptability evaluation results; and controlling and managing the robot according to the adaptability evaluation results.

[0008] In the second aspect, the present application also provides a robot adaptability testing platform in multiple scenarios, which is used to execute the robot adaptability testing method in multiple scenarios as described in the first aspect, wherein the robot adaptability testing platform in multiple scenarios includes: a scene establishment module, which is used to establish a simulated fire scene and obtain scene information of the simulated fire scene, divide the sub-scenes according to the scene information, and generate a ruin space sub-scene, a narrow space sub-scene and a smoke environment sub-scene; a rating identifier generation module, which is used to establish a scene complexity rating for each sub-scene based on the scene information and generate a basic rating identifier; a real-time data set acquisition module, which is used to establish a mapping task for each sub-scene, and configure a monitoring sensor group according to the mapping task, and use the monitoring sensor group to perform real-time monitoring of the robot to obtain A real-time data set of the robot; a network construction module, used to construct an evaluation network, the evaluation network includes a first evaluation sub-network, a second evaluation sub-network and a third evaluation sub-network, the first evaluation sub-network, the second evaluation sub-network and the third evaluation sub-network are respectively mapped to the ruins space sub-scene, the narrow space sub-scene and the smoke environment sub-scene, and the first evaluation sub-network, the second evaluation sub-network and the third evaluation sub-network are set with different bias weights; an adaptability evaluation result generation module, used to send the real-time data set, basic rating identifier, and mapping task to the evaluation network, and perform adaptability test and evaluation of the robot in multiple scenarios through each sub-evaluation network in the evaluation network to generate adaptability evaluation results; a control management module, used to control and manage the robot according to the adaptability evaluation results.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0010] A simulated fire scene is established and scene information of the simulated fire scene is obtained, sub-scenes are divided according to the scene information to generate a ruin space sub-scene, a narrow space sub-scene and a smoke environment sub-scene; scene complexity ratings of each sub-scene are established based on the scene information to generate a basic rating identifier; mapping tasks of each sub-scene are established, and a monitoring sensor group is configured according to the mapping tasks, and the monitoring sensor group is used to perform real-time monitoring of the robot to obtain a real-time data set of the robot; an evaluation network is constructed, and the evaluation network includes a first evaluation sub-network, a second evaluation sub-network and a third evaluation sub-network, and the first evaluation sub-network, the second evaluation sub-network and the third evaluation sub-network are respectively associated with the ruin space sub-scene , narrow space sub-scene and smoke environment sub-scene mapping, and the first evaluation sub-network, the second evaluation sub-network and the third evaluation sub-network are set with different bias weights; the real-time data set, the basic rating identifier, and the mapping task are sent to the evaluation network, and the adaptability test and evaluation of the robot in multiple scenarios are performed through each sub-evaluation network in the evaluation network to generate an adaptability evaluation result; the robot is controlled and managed according to the adaptability evaluation result, which effectively solves the technical problem that the existing technology lacks consideration of the diversity of environments and tasks, resulting in inaccurate adaptability evaluation of the robot in complex and changeable scenarios, improves the comprehensiveness and accuracy of the robot evaluation, enhances the adaptability of the robot, and optimizes the control and management of the robot.

[0011] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented according to the contents of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically cited below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] 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 in the following description are only exemplary, and for ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0013] Figure 1 This is a flowchart of the robot adaptability testing method in multiple scenarios of this application.

[0014] Figure 2 This is a schematic diagram of the structure of the robot adaptability testing platform in multiple scenarios of this application.

[0015] Description of reference numerals:

[0016] Scenario establishment module 11, rating mark generation module 12, real-time data set acquisition module 13, network construction module 14, adaptability evaluation result generation module 15, control management module 16. DETAILED DESCRIPTION

[0017] This application solves the technical problem that the existing technology lacks consideration of environmental and task diversity, resulting in inaccurate robot adaptability evaluation in complex and changeable scenarios, by providing a robot adaptability testing method and platform in multiple scenarios. It improves the comprehensiveness and accuracy of robot evaluation, enhances the adaptability of the robot, and optimizes the control and management of the robot.

[0018] Below, the technical solutions in the present application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited to the example embodiments described herein. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application. It should also be noted that, for the convenience of description, only the parts related to the present application are shown in the accompanying drawings, rather than all of them.

[0019] For example, please refer to the attached Figure 1 The present application provides a robot adaptability testing method in multiple scenarios, wherein the robot adaptability testing method in multiple scenarios is applied to a robot adaptability testing platform in multiple scenarios, and the robot adaptability testing method in multiple scenarios specifically includes the following steps:

[0020] S1: Establish a simulated fire scene, obtain scene information of the simulated fire scene, divide the scene into sub-scenes according to the scene information, and generate a ruin space sub-scene, a narrow space sub-scene, and a smoke environment sub-scene.

[0021] Specifically, use simulation software to create a virtual fire environment and set various parameters of the fire scene, such as fire scale, fire source location, fire spread speed, smoke concentration, etc. Add obstacles, building structures, etc. to the scene to simulate a real fire scene. Through the simulation software interface, obtain various information in the scene, such as terrain, obstacle location, fire source location, etc. Divide the scene into different sub-scenes according to the characteristics of the fire scene and the test requirements of the robot. Each sub-scene should have unique characteristics and challenges to fully test the adaptability of the robot. The ruins space sub-scene simulates the ruins space formed after the collapse of the building, which contains a large number of obstacles and narrow passages. The narrow space sub-scene simulates the narrow space in the fire scene, such as corridors, stairwells, etc. The smoke environment sub-scene simulates the smoke environment in the fire scene. The smoke will interfere with the robot's vision and sensors to test its performance under low visibility conditions. Choose a suitable simulation software or game engine, such as Unity or Unreal Engine. Make sure that the selected tool can provide enough customization options and interfaces to meet the test requirements. Create a new scene in the simulation tool and add fire effects, smoke effects, etc. Adjust parameters such as fire scale and fire source location to simulate real fire conditions. Add obstacles such as buildings, walls, and furniture to the scene to simulate a real fire scene. Ensure that the layout and number of obstacles can form sub-scenes such as ruin space, narrow space, and smoke environment. Divide the scene into ruin space sub-scenes, narrow space sub-scenes, and smoke environment sub-scenes based on the obstacles and fire characteristics in the scene. Use the interface provided by the simulation tool to export information such as terrain, obstacle location, and fire source location in the scene.

[0022] S2: Establishing a scene complexity rating for each sub-scene based on the scene information, and generating a basic rating identifier.

[0023] Specifically, the information includes but is not limited to terrain undulations, obstacle distribution, fire source location and intensity, smoke concentration and diffusion, etc. Rating criteria may include the density of obstacles, the complexity of the terrain, the intensity of the fire, the degree of interference of smoke on sensors and vision, etc. According to the rating criteria, each sub-scenario is evaluated one by one and a corresponding complexity rating is given. This rating should be a quantitative indicator that can clearly reflect the relative complexity of the sub-scenario. For example, a 1-10 rating system can be used, where 1 represents the simplest and 10 represents the most complex. After giving the complexity rating, a basic rating logo needs to be generated for each sub-scenario. This logo should be concise and clear, and the complexity rating of the sub-scenario can be seen at a glance. For example, color coding or digital coding can be used to represent different complexity levels. Finally, the rating information and basic rating logo need to be recorded and applied to subsequent robot adaptability tests. During the test, the difficulty and focus of the test can be adjusted according to the complexity rating of the sub-scenario to ensure the comprehensiveness and effectiveness of the test.

[0024] S3: Establishing mapping tasks for each sub-scenario, and configuring a monitoring sensor group according to the mapping tasks, using the monitoring sensor group to perform real-time monitoring of the robot, and obtaining a real-time data set of the robot.

[0025] Specifically, for each sub-scenario, such as ruins, narrow spaces, and smoke environments, clearly define the mission objectives that the robot needs to complete in the scenario, such as exploration, rescue, and firefighting. Set specific parameters for each task, such as the conditions for starting and ending the task, the initial position of the robot, and the time limit for task execution. Plan the robot's movement path and action sequence in the scene based on the mission objectives and scene characteristics. Select appropriate sensor types, such as cameras, infrared sensors, smoke sensors, and temperature sensors, based on mission requirements and scene characteristics. Rationally arrange sensors on the robot to ensure that the robot's status and environmental information in the scene can be fully and accurately monitored. Set the sensor's sampling frequency, accuracy, range, and other parameters based on mission requirements and sensor characteristics. When the robot performs a task, start the configured monitoring sensor group and start real-time data collection. Preprocess the collected raw data, such as denoising, filtering, and calibration, to improve the accuracy and reliability of the data.

[0026] S4: constructing an evaluation network, wherein the evaluation network includes a first evaluation subnetwork, a second evaluation subnetwork and a third evaluation subnetwork, wherein the first evaluation subnetwork, the second evaluation subnetwork and the third evaluation subnetwork are respectively mapped to a ruin space subscene, a narrow space subscene and a smoke environment subscene, and the first evaluation subnetwork, the second evaluation subnetwork and the third evaluation subnetwork are set with different bias weights.

[0027] Specifically, the evaluation network is a composite neural network consisting of three subnetworks, the first evaluation subnetwork, the second evaluation subnetwork, and the third evaluation subnetwork. Each subnetwork is used to evaluate the robot's performance in a specific subscene. The first evaluation subnetwork is mapped to the ruin space subscene. It is responsible for evaluating the robot's navigation ability, obstacle recognition and handling ability, and stability in the ruin environment. The second evaluation subnetwork is mapped to the narrow space subscene. It focuses on evaluating the robot's mobility flexibility in narrow spaces, adaptability to narrow environments, and task execution capabilities in such environments. The third evaluation subnetwork is mapped to the smoke environment subscene. It mainly evaluates the robot's visual and sensor performance, adaptability to smoke, and task completion in smoke environments. Each evaluation subnetwork is set with different bias weights, which reflect the different requirements and challenges of different subscenes for the robot's adaptability. The bias of the ruin space subnetwork pays more attention to the robot's stability and obstacle handling ability, so a higher bias weight is set on the relevant nodes. The bias of the narrow space subnetwork pays more attention to the robot's mobility flexibility and adaptability to narrow environments, so a higher bias weight is set on the nodes related to these capabilities. The bias of the smoke environment subnetwork focuses more on the performance of the robot's vision and sensors in smoke, as well as the robot's adaptability to smoke environments, so a higher bias weight is set on the nodes in these aspects. Each evaluation subnetwork consists of an input layer, a hidden layer, and an output layer. The input layer receives the robot's real-time data set, including sensor data, task execution status, etc. The hidden layer processes and analyzes the data and extracts key features based on the bias weights and node configurations. The output layer generates an adaptability evaluation result, indicating the robot's performance level in a specific sub-scenario.

[0028] S5: Send the real-time data set, basic rating identifier, and mapping task to the evaluation network, and conduct adaptability test and evaluation of the robot in multiple scenarios through each sub-evaluation network in the evaluation network to generate adaptability evaluation results.

[0029] Specifically, these data are sent to the evaluation network for adaptability test and evaluation. After receiving the data, the evaluation network will assign the real-time data set and the basic rating identifier to the corresponding sub-evaluation network according to the mapping task. For example, the data related to the ruins space sub-scene will be sent to the first evaluation sub-network, the data related to the narrow space sub-scene will be sent to the second evaluation sub-network, and the data related to the smoke environment sub-scene will be sent to the third evaluation sub-network. In each sub-evaluation network, the real-time data set will be used to simulate the actual operation of the robot in the corresponding sub-scene. The basic rating identifier provides the complexity rating information of the sub-scene, which helps the sub-evaluation network to more accurately evaluate the performance of the robot in the scene. The sub-evaluation network will test and evaluate the adaptability of the robot in the sub-scene based on the real-time data set and the basic rating identifier, and generate a preliminary adaptability evaluation result. Finally, the evaluation network will integrate the preliminary adaptability evaluation results generated by each sub-evaluation network to form a comprehensive adaptability evaluation result. This result reflects the overall performance of the robot in all sub-scenes.

[0030] S6: Perform robot control management based on the adaptability evaluation results.

[0031] Specifically, analyze the adaptability evaluation results, including the robot's navigation ability, obstacle handling ability, mobility flexibility, task execution ability, and sensor performance in sub-scenarios such as ruins, narrow spaces, and smoke environments. Further identify the problems and advantages of the robot in different scenarios. Problems may include navigation difficulties of the robot in a specific environment, sensor performance degradation, or inaccurate task execution. Advantages may include the robot's outstanding performance in certain scenarios or outstanding performance of specific capabilities. Develop corresponding control management strategies for the identified problems and advantages. For problems, you can adjust the robot's parameter settings, optimize algorithms, or improve sensor performance. For advantages, you can further develop and utilize these advantages to improve the robot's overall performance.

[0032] Furthermore, the present application also includes:

[0033] The basic rating identification and mapping tasks are received through the preprocessing layer, and data analysis is performed to establish a data analysis set; task importance analysis of task path nodes is performed according to the data analysis set, and feature weight coefficients are established based on the task importance analysis results; sub-networks within the evaluation network are initialized through the feature weight coefficients, and the adaptability test and evaluation of the robot in multiple scenarios are performed based on the initialized sub-networks.

[0034] Specifically, first, the preprocessing layer receives the basic rating identification and mapping tasks. These data contain rating information about the complexity of different scenarios, as well as instructions on which data needs to be mapped to which subnetwork of the evaluation network. The preprocessing layer performs parsing operations on these data, extracts key information, and establishes a data parsing set. This data parsing set will serve as the basis for subsequent task importance analysis and subnetwork initialization. Next, based on the information in the data parsing set, task importance analysis of task path nodes is performed. Based on the results of task importance analysis, feature weight coefficients are established. These coefficients will be used to assign different weights to different features when initializing subnetworks in the evaluation network to reflect their different degrees of influence on the robot's adaptability. The subnetworks in the evaluation network are initialized using the feature weight coefficients. This includes setting the parameters, structure, and initial weights of the subnetworks to ensure that they can accurately evaluate the robot's performance based on the feature weight coefficients. Finally, based on the initialized subnetworks, the robot's adaptability test evaluation in multiple scenarios is performed. The adaptability evaluation results will reflect the robot's overall performance in different scenarios, including its performance in navigation, obstacle handling, mobility flexibility, and task execution capabilities. These results can be used for subsequent robot control management, optimization and improvement.

[0035] Furthermore, the evaluation features of the first evaluation subnetwork include terrain adaptability evaluation features, mobility evaluation features, load capacity evaluation features, autonomous decision-making evaluation features, and stability evaluation features. The first evaluation subnetwork is as follows:

[0036] ;in, Characterize the first evaluation subnetwork, Characterize the terrain adaptability evaluation characteristics, is the total number of detection points in the test, Characterize any detection point, Characterization The path deviation rate of each detection point is Characterize the mobility evaluation features, is the number of path segments, To represent any path, Characterization The ratio of the actual time of the segment path to the ideal time, is the number of turns, is the ideal number of turns, Characterize the load capacity evaluation characteristics, is the total number of load tests, Characterize any load test, For the The actual load of the load test, For the design load, Characterized in the Performance scores for load tests, Characterize the evaluation characteristics of autonomous decision-making, is the total number of decisions, Representing any decision, For the The correctness of the decision, For the The decision complexity of the decision Characterize the stability evaluation characteristics, is the total number of test points, Characterize any test point, Characterized in The instability event of each test point, Characterized in The importance of each test point, , , , , They are the decision coefficients of terrain adaptability evaluation feature, mobility evaluation feature, load capacity evaluation feature, autonomous decision-making evaluation feature, and stability evaluation feature after being initialized by feature weight coefficients.

[0037] Specifically, the terrain adaptability evaluation feature measures the robot's ability to adapt to terrain by the path deviation rate at the detection point. The path deviation rate reflects the degree of deviation between the robot's actual driving path and the preset path. The calculation of the terrain adaptability evaluation feature takes into account the path deviation rate of all detection points and weights it by the feature weight coefficient. The maneuverability evaluation feature evaluates the robot's maneuverability by the ratio of the actual time to the ideal time of the path segment. The closer the ratio is to 1, the better the robot's maneuverability. The maneuverability evaluation feature also considers the comparison between the number of turns and the ideal number of turns to comprehensively evaluate the robot's maneuverability. The load capacity evaluation feature measures the robot's load capacity by comparing the actual load of the load test with the designed load. The performance score reflects the robot's performance under different loads, and the higher the score, the stronger the load capacity. The calculation of the load capacity evaluation feature takes into account the performance scores of all load tests and weights it by the feature weight coefficient. The autonomous decision-making evaluation feature evaluates the robot's autonomous decision-making ability by the correctness and complexity of the decision. The correctness of the decision reflects the proportion of correct decisions made by the robot, while the complexity of the decision measures the difficulty of the decision-making process. The calculation of the autonomous decision evaluation feature considers the correctness and complexity of all decisions and weights them by the feature weight coefficient. The stability evaluation feature measures the stability of the robot by the instability events at the test points. The fewer instability events, the better the stability of the robot. The calculation of the stability evaluation feature also considers the importance of the test points to highlight the impact of key test points on the stability evaluation.

[0038] Furthermore, the evaluation features of the third evaluation subnetwork include visual clarity features, environmental perception features, and communication maintenance features. The third evaluation subnetwork is as follows: ;in, Characterize the third evaluation subnetwork, Characterize visual clarity characteristics, Characterize environmental perception features, Characterize communication maintenance characteristics, , , They respectively represent the decision coefficients of the visual clarity feature, environmental perception feature, and communication maintenance feature after being initialized by the feature weight coefficients.

[0039] Specifically, the visual clarity feature mainly measures the clarity of the robot's visual system in complex environments such as smoke. Visual clarity directly affects the robot's perception and navigation capabilities of the environment. The decision coefficient of the visual clarity feature initialized by the feature weight coefficient reflects the importance of visual clarity in the robot's adaptability evaluation. The environmental perception feature evaluates the robot's ability to perceive environmental information. In a smoke environment, the robot needs to accurately perceive obstacles, paths and other information in the surrounding environment. The decision coefficient of the environmental perception feature indicates the contribution of environmental perception to the robot's adaptability. The communication maintenance feature measures the robot's ability to maintain communication with the base station or other robots in a complex environment. In an interference environment such as smoke, maintaining stable communication is crucial for the robot's task execution. The decision coefficient of the communication maintenance feature reflects the weight of the communication capability in the robot's adaptability evaluation. In the third evaluation subnetwork, these three evaluation features are integrated and weighted by the feature weight coefficient to form a comprehensive decision coefficient for the robot's adaptability evaluation in a specific scenario. These decision coefficients reflect the contribution of different evaluation features to the robot's adaptability and can be used for subsequent robot control management and optimization improvements. By comprehensively considering visual clarity, environmental perception, and communication maintenance capabilities, the robot's adaptive performance in complex environments can be more comprehensively evaluated.

[0040] Furthermore, the present application also includes:

[0041] Acquire setting information of the robot, and configure an error attenuation factor according to the setting information; evaluate navigation accuracy through the real-time data set, and perform self-correction compensation through the error attenuation factor, and establish a navigation accuracy evaluation result according to the compensation result; and correct the adaptability evaluation result through the navigation accuracy evaluation result.

[0042] Specifically, the robot's setting information is obtained from the robot's configuration file or through the communication interface with the robot. The setting information includes the robot model, the type of navigation system, the sensor configuration, and any parameters related to navigation performance. According to the obtained robot setting information, an error attenuation factor is configured. This factor is used to attenuate the error in the real-time data set during the navigation accuracy evaluation process to improve the accuracy and reliability of the evaluation. The specific value of the error attenuation factor may depend on the robot model and the characteristics of the navigation system. The navigation accuracy is evaluated using the real-time data set. This evaluation process involves evaluating the deviation of the robot's navigation path, the accuracy of reaching the target point, and the stability during the navigation process. The evaluation results will reflect the navigation performance of the robot in the current scenario. According to the results of the navigation accuracy evaluation, self-correction compensation is performed through the error attenuation factor. Reduce the error in the evaluation and make a more accurate evaluation of the robot's navigation performance. The compensation results will be used to establish the final navigation accuracy evaluation results. Finally, the navigation accuracy evaluation results are used to correct the previous adaptability evaluation results. This correction process may involve adjusting certain parameters in the adaptability evaluation features to more accurately reflect the adaptability of the robot in the current scenario. The revised adaptability evaluation results will more comprehensively reflect the overall performance of the robot and can be used for subsequent robot control management and optimization improvements.

[0043] Furthermore, the present application also includes:

[0044] An abnormal triggering threshold is configured according to the mapping task; when the adaptability test evaluation is performed through the evaluation network, the collaborative authentication is triggered through the abnormal triggering threshold; and the adaptability evaluation result is compensated by triggering the collaborative authentication result.

[0045] Specifically, according to the specific requirements of the mapping task, an abnormal trigger threshold is configured. This threshold is used to determine whether the robot's performance has reached the standard required to trigger collaborative certification during the adaptability test evaluation process. If some evaluation features or the overall evaluation score of the robot are lower than this threshold, collaborative certification will be triggered. Next, the robot is evaluated for adaptability through the evaluation network. This evaluation process takes into account various performance parameters and real-time data sets of the robot to evaluate the adaptability of the robot in a specific scenario. During the adaptability test evaluation process, if some evaluation features or the overall evaluation score of the robot are lower than the previously configured abnormal trigger threshold, collaborative certification will be triggered. Collaborative certification involves additional testing, data verification, or interaction with other systems to more comprehensively evaluate the performance of the robot. After performing collaborative certification, a collaborative certification result is obtained. This result is a new evaluation score, a set of additional performance parameters, or other information about the adaptability of the robot. Finally, the previous adaptability evaluation result is compensated according to the result of collaborative certification. If collaborative certification shows that the robot's performance is actually better than the initial evaluation, the robot's adaptability evaluation score may be increased. Conversely, if collaborative certification shows that the robot's performance is worse, the adaptability evaluation score may be reduced.

[0046] Furthermore, the present application also includes:

[0047] Determine whether the triggering collaborative authentication result satisfies the abnormal warning; if the triggering collaborative authentication result satisfies the abnormal warning, stop the robot test and simultaneously issue an abnormal warning.

[0048] Specifically, according to the specific requirements of the mapping task, an abnormal trigger threshold is set. This threshold is used to determine whether the robot's performance has reached the standard that needs to trigger collaborative certification. The robot is evaluated for adaptability through the evaluation network, considering various performance parameters and real-time data sets of the robot to evaluate its adaptability in a specific scenario. If some evaluation features or the overall evaluation score of the robot is lower than the abnormal trigger threshold, collaborative certification will be triggered. Collaborative certification additional tests, data verification or interaction with other systems. After performing collaborative certification, a collaborative certification result will be obtained, which is a new evaluation score, a set of additional performance parameters or other information about the adaptability of the robot. The collaborative certification result is analyzed to determine whether it meets the preset abnormal warning conditions. These conditions include specific performance parameter thresholds, the decline in evaluation scores, etc. If the collaborative certification result meets the abnormal warning conditions, the current robot test is stopped immediately, and the abnormal warning alarm is issued synchronously. Including sending warning notifications to relevant personnel, recording abnormal information, etc.

[0049] In summary, the robot adaptability testing method in multiple scenarios provided by this application has the following technical effects:

[0050] A simulated fire scene is established and scene information of the simulated fire scene is obtained, sub-scenes are divided according to the scene information to generate a ruin space sub-scene, a narrow space sub-scene and a smoke environment sub-scene; scene complexity ratings of each sub-scene are established based on the scene information to generate a basic rating identifier; mapping tasks of each sub-scene are established, and a monitoring sensor group is configured according to the mapping tasks, and the monitoring sensor group is used to perform real-time monitoring of the robot to obtain a real-time data set of the robot; an evaluation network is constructed, and the evaluation network includes a first evaluation sub-network, a second evaluation sub-network and a third evaluation sub-network, and the first evaluation sub-network, the second evaluation sub-network and the third evaluation sub-network are respectively associated with the ruin space sub-scene , narrow space sub-scene and smoke environment sub-scene mapping, and the first evaluation sub-network, the second evaluation sub-network and the third evaluation sub-network are set with different bias weights; the real-time data set, the basic rating identifier, and the mapping task are sent to the evaluation network, and the adaptability test and evaluation of the robot in multiple scenarios are performed through each sub-evaluation network in the evaluation network to generate an adaptability evaluation result; the robot is controlled and managed according to the adaptability evaluation result, which effectively solves the technical problem that the existing technology lacks consideration of the diversity of environments and tasks, resulting in inaccurate adaptability evaluation of the robot in complex and changeable scenarios, improves the comprehensiveness and accuracy of the robot evaluation, enhances the adaptability of the robot, and optimizes the control and management of the robot.

[0051] Embodiment 2: Based on the robot adaptability testing method in multiple scenarios in the previous embodiment, the present application also provides a robot adaptability testing platform in multiple scenarios, please refer to the attached Figure 2 , the robot adaptability testing platform under multiple scenarios includes:

[0052] The scene establishment module 11 is used to establish a simulated fire scene, obtain scene information of the simulated fire scene, divide the scene into sub-scenes according to the scene information, and generate a ruin space sub-scene, a narrow space sub-scene and a smoke environment sub-scene.

[0053] The rating mark generation module 12 is used to establish the scene complexity rating of each sub-scene based on the scene information and generate a basic rating mark.

[0054] The real-time data set acquisition module 13 is used to establish mapping tasks for each sub-scene, and configure a monitoring sensor group according to the mapping tasks, perform real-time monitoring of the robot with the monitoring sensor group, and acquire the real-time data set of the robot.

[0055] The network construction module 14 is used to construct an evaluation network, which includes a first evaluation sub-network, a second evaluation sub-network and a third evaluation sub-network. The first evaluation sub-network, the second evaluation sub-network and the third evaluation sub-network are mapped to the ruins space sub-scene, the narrow space sub-scene and the smoke environment sub-scene respectively, and the first evaluation sub-network, the second evaluation sub-network and the third evaluation sub-network are set with different bias weights.

[0056] The adaptability evaluation result generation module 15 is used to send the real-time data set, basic rating identifier, and mapping task to the evaluation network, and perform adaptability test evaluation of the robot in multiple scenarios through each sub-evaluation network in the evaluation network to generate adaptability evaluation results.

[0057] The control management module 16 is used to perform robot control management according to the adaptability evaluation results.

[0058] Furthermore, the robot adaptability testing platform under multiple scenarios also includes an adaptive evaluation module for:

[0059] The basic rating identification and mapping tasks are received through the preprocessing layer, and data analysis is performed to establish a data analysis set; task importance analysis of task path nodes is performed according to the data analysis set, and feature weight coefficients are established based on the task importance analysis results; sub-networks within the evaluation network are initialized through the feature weight coefficients, and the adaptability test and evaluation of the robot in multiple scenarios are performed based on the initialized sub-networks.

[0060] Furthermore, the robot adaptability testing platform under multiple scenarios also includes a first evaluation sub-network construction module, which is used to:

[0061] The evaluation features of the first evaluation subnetwork include terrain adaptability evaluation features, mobility evaluation features, load capacity evaluation features, autonomous decision-making evaluation features, and stability evaluation features. The first evaluation subnetwork is as follows:

[0062] ;in, Characterize the first evaluation subnetwork, Characterize the terrain adaptability evaluation characteristics, is the total number of detection points in the test, Characterize any detection point, Characterization The path deviation rate of each detection point is Characterize the mobility evaluation features, is the number of path segments, To represent any path, Characterization The ratio of the actual time of the segment path to the ideal time, is the number of turns, is the ideal number of turns, Characterize the load capacity evaluation characteristics, is the total number of load tests, Characterize any load test, For the The actual load of the load test, For the design load, Characterized in the Performance scores for load tests, Characterize the evaluation characteristics of autonomous decision-making, is the total number of decisions, Representing any decision, For the The correctness of the decision, For the The decision complexity of the decision Characterize the stability evaluation characteristics, is the total number of test points, Characterize any test point, Characterized in Instability events at each test point, Characterized in The importance of each test point, , , , , They are the decision coefficients of terrain adaptability evaluation feature, mobility evaluation feature, load capacity evaluation feature, autonomous decision-making evaluation feature, and stability evaluation feature after being initialized by feature weight coefficients.

[0063] Furthermore, the robot adaptability testing platform under multiple scenarios also includes a third evaluation sub-network construction module, which is used to:

[0064] The evaluation features of the third evaluation sub-network include visual clarity features, environmental perception features, and communication maintenance features. The third evaluation sub-network is as follows: ;in, Characterize the third evaluation subnetwork, Characterize visual clarity characteristics, Characterize environmental perception features, Characterize communication maintenance characteristics, , , They respectively represent the decision coefficients of the visual clarity feature, environmental perception feature, and communication maintenance feature after being initialized by the feature weight coefficients.

[0065] Furthermore, the robot adaptability testing platform under multiple scenarios also includes a result correction module, which is used to:

[0066] Acquire setting information of the robot, and configure an error attenuation factor according to the setting information; evaluate navigation accuracy through the real-time data set, and perform self-correction compensation through the error attenuation factor, and establish a navigation accuracy evaluation result according to the compensation result; and correct the adaptability evaluation result through the navigation accuracy evaluation result.

[0067] Furthermore, the robot adaptability testing platform under multiple scenarios also includes a result compensation module, which is used to:

[0068] An abnormal triggering threshold is configured according to the mapping task; when the adaptability test evaluation is performed through the evaluation network, the collaborative authentication is triggered through the abnormal triggering threshold; and the adaptability evaluation result is compensated by triggering the collaborative authentication result.

[0069] Furthermore, the robot adaptability testing platform under multiple scenarios also includes a pre-alarm output module, which is used to:

[0070] Determine whether the triggering collaborative authentication result satisfies the abnormal warning; if the triggering collaborative authentication result satisfies the abnormal warning, stop the robot test and simultaneously issue an abnormal warning.

[0071] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. Figure 1 The robot adaptability test method and specific examples in multiple scenarios in the first embodiment are also applicable to the robot adaptability test platform in multiple scenarios of the present embodiment. Through the above detailed description of the robot adaptability test method in multiple scenarios, those skilled in the art can clearly know the robot adaptability test platform in multiple scenarios of the present embodiment, so for the sake of brevity of the specification, it will not be described in detail here. For the platform 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 description.

[0072] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those 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 will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

[0073] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalent technology, the present application is also intended to include these modifications and variations.

Claims

1. A robot adaptability testing method in multiple scenarios, characterized in that: include: Establishing a simulated fire scene, obtaining scene information of the simulated fire scene, dividing the scene into sub-scenes according to the scene information, and generating a ruin space sub-scene, a narrow space sub-scene, and a smoke environment sub-scene; Establishing a scene complexity rating for each sub-scene based on the scene information, and generating a basic rating identifier; Establishing mapping tasks for each sub-scenario, and configuring a monitoring sensor group according to the mapping tasks, using the monitoring sensor group to perform real-time monitoring of the robot, and obtaining a real-time data set of the robot; Constructing an evaluation network, the evaluation network comprising a first evaluation subnetwork, a second evaluation subnetwork and a third evaluation subnetwork, the first evaluation subnetwork, the second evaluation subnetwork and the third evaluation subnetwork being mapped to a ruin space subscene, a narrow space subscene and a smoke environment subscene respectively, and the first evaluation subnetwork, the second evaluation subnetwork and the third evaluation subnetwork being set with different bias weights; The real-time data set, basic rating identifier, and mapping task are sent to the evaluation network, and the adaptability test and evaluation of the robot in multiple scenarios are performed through each sub-evaluation network in the evaluation network to generate an adaptability evaluation result; Carry out robot control management based on adaptability evaluation results.

2. The robot adaptability testing method in multiple scenarios as claimed in claim 1, characterized in that: After sending the real-time data set, basic rating identifier, and mapping task to the evaluation network, the method further includes: Receiving the basic rating identification and mapping task through the preprocessing layer, and performing data parsing to establish a data parsing set; Performing task importance analysis on task path nodes according to the data analysis set, and establishing feature weight coefficients based on the task importance analysis results; The sub-networks in the evaluation network are initialized by using the characteristic weight coefficients, and the adaptability test and evaluation of the robot in multiple scenarios are performed based on the initialized sub-networks.

3. The robot adaptability testing method in multiple scenarios as claimed in claim 2, characterized in that: The evaluation features of the first evaluation subnetwork include terrain adaptability evaluation features, mobility evaluation features, load capacity evaluation features, autonomous decision-making evaluation features, and stability evaluation features. The first evaluation subnetwork is as follows: ； in, Characterize the first evaluation subnetwork, Characterize the terrain adaptability evaluation characteristics, is the total number of detection points in the test, Characterize any detection point, Characterization The path deviation rate of each detection point is Characterize the mobility evaluation features, is the number of path segments, is the weight coefficient of the i-th detection point, To represent any path, Characterization The ratio of the actual time of the segment path to the ideal time, is the number of turns, is the ideal number of turns, Characterize the load capacity evaluation characteristics, is the total number of load tests, Characterize any load test, For the The actual load of the load test, For the design load, Characterized in the Performance scores for load tests, Characterize the evaluation characteristics of autonomous decision-making, is the total number of decisions, Representing any decision, For the The correctness of the decision, For the The decision complexity of the decision Characterize the stability evaluation characteristics, is the total number of test points, Characterize any test point, Characterized in The instability event of each test point, Characterized in The importance of each test point, , , , , They are the decision coefficients of terrain adaptability evaluation feature, mobility evaluation feature, load capacity evaluation feature, autonomous decision-making evaluation feature, and stability evaluation feature after being initialized by feature weight coefficients.

4. The robot adaptability testing method in multiple scenarios as claimed in claim 2, characterized in that: The evaluation features of the third evaluation sub-network include visual clarity features, environmental perception features, and communication maintenance features. The third evaluation sub-network is as follows: ； in, Characterize the third evaluation subnetwork, Characterize visual clarity characteristics, Characterize environmental perception features, Characterize communication maintenance characteristics, , , They respectively represent the decision coefficients of the visual clarity feature, environmental perception feature, and communication maintenance feature after being initialized by the feature weight coefficients.

5. The robot adaptability testing method in multiple scenarios as claimed in claim 1, characterized in that: Also includes: Acquire setting information of the robot, and configure an error attenuation factor according to the setting information; Performing navigation accuracy evaluation through the real-time data set, performing self-correction compensation through the error attenuation factor, and establishing a navigation accuracy evaluation result according to the compensation result; The adaptability evaluation result is corrected by the navigation accuracy evaluation result.

6. The robot adaptability testing method in multiple scenarios as claimed in claim 1, characterized in that: Also includes: Configure an exception trigger threshold according to the mapping task; When the adaptability test evaluation is performed through the evaluation network, triggering collaborative authentication is performed through the abnormal trigger threshold; The adaptive evaluation results are compensated by triggering the collaborative authentication results.

7. The robot adaptability testing method in multiple scenarios as claimed in claim 6, characterized in that: Also includes: Determine whether the triggering collaborative authentication result satisfies the abnormal warning; If the triggered collaborative authentication result meets the abnormal warning, the robot test is stopped and the abnormal warning is issued simultaneously.

8. The robot adaptability testing platform in multiple scenarios is characterized by: The steps for implementing the robot adaptability testing method in multiple scenarios as described in any one of claims 1 to 7, wherein the robot adaptability testing platform in multiple scenarios comprises: A scene establishment module is used to establish a simulated fire scene, obtain scene information of the simulated fire scene, divide the scene into sub-scenes according to the scene information, and generate a ruin space sub-scene, a narrow space sub-scene, and a smoke environment sub-scene; A rating mark generation module, used to establish a scene complexity rating for each sub-scene based on the scene information and generate a basic rating mark; A real-time data set acquisition module is used to establish mapping tasks for each sub-scene, and configure a monitoring sensor group according to the mapping tasks, perform real-time monitoring of the robot with the monitoring sensor group, and acquire the real-time data set of the robot; A network construction module is used to construct an evaluation network, wherein the evaluation network includes a first evaluation subnetwork, a second evaluation subnetwork and a third evaluation subnetwork, wherein the first evaluation subnetwork, the second evaluation subnetwork and the third evaluation subnetwork are respectively mapped to a ruin space subscene, a narrow space subscene and a smoke environment subscene, and the first evaluation subnetwork, the second evaluation subnetwork and the third evaluation subnetwork are set with different bias weights; An adaptability evaluation result generation module is used to send the real-time data set, basic rating identifier, and mapping task to the evaluation network, and to perform adaptability test evaluation of the robot in multiple scenarios through each sub-evaluation network in the evaluation network to generate an adaptability evaluation result; The control management module is used to perform robot control management based on the adaptability evaluation results.

Images