Positioning loss detection method, system and equipment for intelligent operation and maintenance robot of computing power center and medium

Abstract

The invention provides a positioning loss detection method, system and device for an intelligent operation and maintenance robot of a computing power center and a medium, and belongs to the technical field of operation and maintenance robot positioning detection.The method comprises the steps that a laser radar, an IMU and an odometer of the robot are used for collecting real-time data to construct a likelihood domain model map for a computer room of the computing power center; performing scanning matching on the likelihood domain model map based on the predicted pose and the real-time data of the laser radar to generate an updated pose; calculating the difference between the updated pose and the predicted pose, and generating a pose difference score; a historical pose sequence and a historical speed sequence are collected, and a trajectory deviation degree score and a speed deviation degree score are calculated through Gaussian distribution modeling; calculating a matching confidence score of the point cloud in the real-time data of the laser radar and the likelihood domain model map; and performing weighted summation on the scores to obtain a comprehensive loss positioning risk score. According to the invention, the lost positioning condition of the robot in a similar environment can be accurately and timely detected, and the accuracy of lost positioning detection is improved.

Description

Technical Field

[0001] This application belongs to the technical field of positioning detection of operation and maintenance robots, and particularly relates to a method, system, device and medium for detecting lost positioning of intelligent operation and maintenance robots in a computing power center. Background Art

[0002] With the development of information technology, as the core facility for data processing and storage, the scale and complexity of the computing power center are constantly increasing. Intelligent operation and maintenance robots play an important role in patrol and maintenance in the computing power center, and the accuracy of their positioning is crucial for ensuring the stable operation of the computing power center. However, due to the neat arrangement of a large number of server cabinets and the single environmental characteristics, the internal environment of the computing power center has a high degree of similarity, making it easy for the robot to lose positioning during the patrol process, resulting in its inability to accurately complete the operation and maintenance tasks, and even possibly causing safety accidents.

[0003] At present, although there are some positioning detection methods, there are still many problems for the lost positioning problem in the similar environment of the computing power center. First, traditional data center cabinets mostly use grid doors for heat dissipation, and the maps constructed by Tof sensors such as lidar have many map burrs. When the positioning deviates, relying solely on the occupancy grid score to judge the positioning quality is slow and inaccurate. Second, the vision-based positioning method is greatly affected by light conditions and is difficult to work stably in the complex lighting environment inside the computing power center. Third, most of the existing positioning detection methods rely on single sensor data, lack the fusion analysis of multiple information, and are difficult to comprehensively and accurately judge the positioning state of the robot. In addition, the traditional detection method has a slow response and cannot meet the reliability requirements of intelligent operation and maintenance robots in the computing power center in practical applications. Summary of the Invention

[0004] In a first aspect, an embodiment of this application provides a method for detecting lost positioning of an intelligent operation and maintenance robot in a computing power center, including the following steps: S1. Use the lidar, IMU and odometer of the robot to collect real-time data to construct a likelihood domain model map for the computer room of the computing power center; S2. Calculate the predicted pose based on the initial pose of the robot and the real-time data of the IMU and odometer, and perform scan matching on the likelihood domain model map based on the predicted pose and the real-time lidar data to generate an updated pose; S3. Calculate the difference between the updated pose and the predicted pose to generate a pose difference score; S4. Collect the historical pose sequence and historical speed sequence, and calculate the trajectory deviation degree score of the updated pose relative to the historical pose sequence and the speed deviation degree score of the updated speed relative to the historical speed sequence through Gaussian distribution modeling; S5. Calculate the matching confidence score between the point cloud in the real-time lidar data and the likelihood domain model map. S6. Perform a weighted sum of the pose difference score, the trajectory deviation degree score, the speed deviation degree score, and the matching confidence score to obtain a comprehensive positioning loss risk score. By performing a weighted sum of multiple evaluation scores to obtain a comprehensive positioning loss risk score, different factors can be comprehensively considered, providing a comprehensive and quantitative positioning loss risk assessment index for subsequent decision-making and intervention.

[0005] Further, the specific steps of step S1 are as follows: S11. Control the robot to run in the computing center computer room, and control the robot to collect distance data from the environment boundary and the robot's pose data during the running process. S12. Perform ray projection mapping on the distance data from the environment boundary and the robot's pose data collected by the robot during the running process into a two-dimensional plane grid to obtain a grid map, and then use the likelihood domain model to construct a likelihood domain model map of the grid map. Limiting the data collection and the construction process of the likelihood domain model map improves the operability and accuracy of map construction, providing a reliable map basis for subsequent positioning detection.

[0006] Further, the specific steps of using the likelihood domain model to construct a likelihood domain model map of the grid map in step S12 are as follows: Calculate the distance between each grid and the nearest obstacle. According to the distance to the nearest obstacle, calculate the occupancy probability of each grid using the following sensor model:

[0007] where is the attenuation coefficient, and d is the distance from the grid to the nearest obstacle. Limiting the construction method of the likelihood domain model map can finely describe the uncertainty of the environment by calculating the distance between the grid and the obstacle and using the sensor model to calculate the probability value, improving the map's support ability for robot positioning.

[0008] Further, the specific steps of step S2 are as follows: S21. Calculate the predicted pose within the target time period based on the robot's initial pose and the real-time acceleration data collected by the IMU and the angular velocity and acceleration data collected in real time by the odometer. S22. Determine the search area on the likelihood domain model map according to the predicted pose. S23. In the search area, perform grid-by-grid matching between the real-time lidar data and the likelihood domain model map. S24. Calculate the consistency of the probability values between the lidar measurement points and the grids. S25. Search for the pose that maximizes the overall matching score within the search area through an optimization algorithm as the updated pose. The calculation of the updated pose is restricted, including predicted pose calculation, search area determination, grid-by-grid matching, consistency calculation, and application of the optimization algorithm, making the generation process of the updated pose clear and accurate, and improving the efficiency and accuracy of positioning updates.

[0009] Further, the specific steps of step S4 are as follows: S41. Collect the historical pose sequence of the robot within a set time period and calculate the historical pose mean; S42. Collect the historical average speed of the robot within a set time period; S43. Calculate the deviation degree between the updated pose and the historical pose mean through the Gaussian distribution function; S44. Obtain the updated speed and calculate the deviation degree between the updated speed and the historical average speed through the Gaussian distribution function. Through the specific steps of collecting historical data, calculating the mean, and calculating the deviation degree score using Gaussian distribution modeling, the deviation degree of the robot's motion state can be accurately evaluated, providing a reliable basis for lost positioning detection.

[0010] Further, the specific steps of step S5 are as follows: S51. Obtain the point cloud in the real-time data collected by the lidar at the current pose; S52. Transform the point cloud in the real-time data to the coordinate system used in the likelihood domain map based on the updated pose; S53. Calculate the grid position corresponding to each point cloud in the likelihood domain map and determine the occupied probability value of the corresponding grid position in the likelihood domain map; S54. Calculate the matching probability score of each point cloud with the occupied probability value, and then calculate the mean of the matching probability scores of all point clouds to obtain the matching score; S55. Use the covariance matrix during the matching process of each point cloud with the likelihood domain map as the scan matching information matrix; S56. Calculate the matching confidence score based on the matching scores of all point clouds and the mean of the scan matching information. The specific steps of calculating the matching confidence score are restricted, including point cloud acquisition, coordinate transformation, probability value matching, matching score calculation, and information matrix application, making the calculation process of the matching confidence score clear and accurate, and improving the reliability of the detection result.

[0011] Further, in step S6, the pose difference score, trajectory deviation degree score, speed deviation degree score, and matching confidence score are weighted and summed through the following formula:

[0012] Among them, is the pose difference score, is the trajectory deviation degree score, is the speed deviation degree score, is the matching score, is the scan matching information matrix, and and and are the weights of the corresponding evaluation items. The weighted summation formula for the comprehensive positioning loss risk score is defined to make the scoring process transparent and operable, facilitate the adjustment of weights according to actual needs, and improve the flexibility and adaptability of the detection method.

[0013] In a second aspect, an embodiment of the present application further provides a positioning loss detection system for an intelligent operation and maintenance robot in a computing power center, including: A likelihood domain model map construction module, configured to use the lidar, IMU, and odometer of the robot to collect real-time data to construct a likelihood domain model map for the computing power center computer room; An updated pose calculation module, configured to calculate a predicted pose based on the initial pose of the robot and the real-time data of the IMU and odometer, perform scan matching on the likelihood domain model map based on the predicted pose and the real-time lidar data, and generate an updated pose; A pose difference score calculation module, configured to calculate the difference between the updated pose and the predicted pose, and generate a pose difference score; A deviation degree score calculation module, configured to collect a historical pose sequence and a historical speed sequence, and calculate the trajectory deviation degree score of the updated pose relative to the historical pose sequence and the speed deviation degree score of the updated speed relative to the historical speed sequence through Gaussian distribution modeling; A matching confidence score calculation module, configured to calculate the matching confidence score between the point cloud in the real-time lidar data and the likelihood domain model map; A positioning loss comprehensive scoring module, configured to perform weighted summation on the pose difference score, the trajectory deviation degree score, the speed deviation degree score, and the matching confidence score to obtain a comprehensive positioning loss risk score. Through the interaction and collaboration of the likelihood domain model map construction module, the updated pose calculation module, the pose difference score calculation module, the deviation degree score calculation module, the matching confidence score calculation module, and the positioning loss comprehensive scoring module, the positioning loss situation of the robot in a similar environment can be accurately and timely detected, the accuracy and reliability of the positioning loss detection are improved, the stable operation of the computing power center is ensured, and the operation and maintenance cost and safety risk are reduced.

[0014] In a third aspect, an embodiment of the present application further provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the steps of the positioning loss detection method for the intelligent operation and maintenance robot in the computing power center as described in the first aspect are implemented.

[0015] In a fourth aspect, an embodiment of the present application further provides a storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method for detecting the loss of positioning of the intelligent operation and maintenance robot in the computing power center described in the first aspect are implemented.

[0016] As can be seen from the above technical solutions, the present application has the following advantages: In the method, system, device and medium for detecting the loss of positioning of the intelligent operation and maintenance robot in the computing power center provided by the present application, through multi-sensor data fusion, likelihood domain model map construction, and weighted summation of multiple evaluation methods, it is possible to accurately and timely detect the loss of positioning of the robot in a similar environment, improve the accuracy and reliability of the loss of positioning detection, ensure the stable operation of the computing power center, and reduce the operation and maintenance costs and safety risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the present application, the drawings required for description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0018] Figure 1 It is a schematic flowchart of the method for detecting the loss of positioning of the intelligent operation and maintenance robot in the computing power center of the present invention.

[0019] Figure 2 It is a schematic diagram of the system for detecting the loss of positioning of the intelligent operation and maintenance robot in the computing power center of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] In the following, the specific steps of the method for detecting the loss of positioning of the intelligent operation and maintenance robot in the computing power center will be described in detail, and various embodiments of the present disclosure will be described more comprehensively. The present disclosure may have various embodiments, and adjustments and changes may be made therein. However, it should be understood that there is no intention to limit the various embodiments of the present disclosure to the specific embodiments disclosed herein, but the present disclosure should be understood to cover all adjustments, equivalents and / or alternative solutions falling within the spirit and scope of the various embodiments of the present disclosure.

[0021] Exemplarily, with the rapid development of information technology, as the core infrastructure for data processing and storage, the scale and complexity of the computing power center continue to expand. Intelligent operation and maintenance robots play an important role in the inspection and maintenance tasks in the computing power center, and their positioning accuracy is extremely critical for ensuring the stable operation of the computing power center. However, the internal environment of the computing power center is highly similar, and the large number of neatly arranged server cabinets makes the environmental features single, which leads to the phenomenon of loss of positioning of the robot during the inspection process, and then it is impossible to accurately complete the operation and maintenance tasks, and even may cause safety accidents.

[0022] Although there are some positioning detection technologies at present, there are still many challenges in dealing with the problem of lost positioning in a similar environment of a computing power center. First, most cabinets in traditional data centers use grid doors to facilitate heat dissipation, which makes the maps constructed by ToF sensors such as lidar have many burrs. When the positioning deviates, relying solely on the occupancy grid score to judge the positioning quality is not only slow but also lacks accuracy. Second, vision-based positioning methods are extremely sensitive to light conditions and it is difficult to maintain stable operation in the complex lighting environment of a computing power center. In addition, most existing positioning detection methods rely on single-sensor data, lack the fusion analysis of multiple information, and it is difficult to comprehensively and accurately evaluate the positioning state of the robot. Finally, traditional detection methods have a slow response speed and cannot meet the high requirements for reliability in the actual application of intelligent operation and maintenance robots in a computing power center.

[0023] In view of the above problems, this embodiment provides a method for detecting lost positioning of an intelligent operation and maintenance robot in a computing power center. From the process of map construction to comprehensive scoring, through weighted summation of multiple evaluation methods, it can comprehensively consider the influence of different factors on lost positioning, improve the accuracy and reliability of detection, and provide a guarantee for the stable operation of the intelligent operation and maintenance robot in the computing power center.

[0024] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0025] Please refer to Figure 1 The following is a flowchart of a method for detecting lost positioning of an intelligent operation and maintenance robot in a computing power center in a specific embodiment. The method includes the following steps: S1. Use the lidar, IMU and odometer of the robot to collect real-time data to construct a likelihood domain model map for the computer room of the computing power center; It should be noted that by fusing multi-sensor data to construct a likelihood domain model map, the environment of the computer room of the computing power center can be accurately described, providing a basis for positioning detection, improving the stability and accuracy of the map, and reducing the positioning error caused by environmental similarity; S2. Calculate the predicted pose based on the initial pose of the robot and the real-time data of the IMU and odometer, and perform scan matching on the likelihood domain model map based on the predicted pose and the real-time data of the lidar to generate an updated pose; It should be noted that by calculating the predicted pose in combination with the initial pose and real-time sensor data and generating the updated pose through scan matching, the position information of the robot can be updated in real time, improving the dynamic accuracy of positioning and enhancing the adaptability of the robot in complex environments; S3. Calculate the difference between the updated pose and the predicted pose to generate a pose difference score; It should be noted that calculating the difference between the updated pose and the predicted pose and generating a pose difference score can directly reflect the rationality of the robot's pose change, timely detect potential positioning loss risks, and provide a basis for subsequent comprehensive evaluation; S4. Collect the historical pose sequence and historical speed sequence, and calculate the trajectory deviation degree score of the updated pose relative to the historical pose sequence and the speed deviation degree score of the updated speed relative to the historical speed sequence through Gaussian distribution modeling; It should be noted that by collecting the historical pose and speed sequences and calculating the deviation degree scores using Gaussian distribution modeling, the motion state of the robot can be evaluated from both the time and space dimensions, improving the comprehensiveness and accuracy of positioning loss detection; S5. Calculate the matching confidence score between the point cloud in the real-time lidar data and the likelihood domain model map; It should be noted that calculating the matching confidence score between the lidar point cloud and the likelihood domain model map can evaluate the positioning quality from the perspective of the matching between sensor data and the map, increasing the detection dimension and improving the reliability of the detection results; S6. Perform weighted summation on the pose difference score, trajectory deviation degree score, speed deviation degree score, and matching confidence score to obtain a comprehensive positioning loss risk score.

[0026] In this embodiment, by performing weighted summation on multiple evaluation scores to obtain a comprehensive positioning loss risk score, different factors can be comprehensively considered, providing a comprehensive and quantitative positioning loss risk assessment index, which is convenient for subsequent decision-making and intervention.

[0027] Furthermore, as a refinement and extension of the specific implementation manner of the above embodiment, in order to fully illustrate the specific implementation process in this embodiment, another method for detecting positioning loss of the intelligent operation and maintenance robot in the computing power center is provided. This method includes the following steps: S1. Use the lidar, IMU, and odometer of the robot to collect real-time data to construct a likelihood domain model map for the computing power center computer room; The specific steps of step S1 are as follows: S11. Control the robot to run in the computing power center computer room, and control the robot to collect distance data from the environmental boundary and the pose data of the robot during the running process; Specifically, the control robot collects the distance data from the boundary during operation through lidar, collects the angular velocity and acceleration information of the robot through IMU, and collects the distance and direction information of the robot's movement through the odometer; the IMU uses a 9-axis IMU; S12. Ray-project the distance data from the environmental boundary and the pose data of the robot collected during operation onto a two-dimensional plane grid to obtain a grid map, and then use the likelihood domain model to construct a likelihood domain model map of the grid map; Exemplarily, taking a 5 cm grid map accuracy as an example, for instance, in a certain 10x10 meter area, a 200x200 grid map is constructed, and each grid represents the occupancy situation of the area; It should be noted that the limitation of the data collection and the construction process of the likelihood domain model map improves the operability and accuracy of map construction, providing a reliable map basis for subsequent positioning detection; The specific steps of using the likelihood domain model to construct a likelihood domain model map of the grid map in step S12 are as follows: Calculate the distance from each grid to the nearest obstacle; Based on the distance to the nearest obstacle, calculate the occupancy probability of each grid using the following sensor model:

[0028] where is the attenuation coefficient, and d is the distance from the grid to the nearest obstacle; It should be noted that the limitation of the construction method of the likelihood domain model map can finely describe the uncertainty of the environment by calculating the distance between the grid and the obstacle and using the sensor model to calculate the probability value, improving the support ability of the map for robot positioning; S2. Calculate the predicted pose based on the initial pose of the robot and the real-time data of the IMU and the odometer, and perform scan matching on the likelihood domain model map based on the predicted pose and the real-time lidar data to generate an updated pose; the specific steps of step S2 are as follows: S21. Calculate the predicted pose within the target time period based on the initial pose of the robot and the real-time acceleration data collected by the IMU and the angular velocity and acceleration data collected in real time by the odometer; Exemplarily, the wheel odometer collects the linear velocity v and angular velocity ω of the robot, and the IMU as the axis inertial unit collects angular acceleration compensation and attitude correction;

[0029] Δt represents the time difference from the previous positioning update to the current moment; It should be noted that the displacement is calculated through the linear velocity v and time Δt of the wheel odometer, and the x and y increments are decomposed in combination with the robot's orientation θ; the rotational motion directly integrates the angular velocity ω of the IMU to obtain the orientation change; the IMU compensation corrects the errors of the wheel odometer during slipping or bumping, such as the correction of the angular velocity in the Z axis. S22. Determine the search area on the likelihood domain model map according to the predicted pose. S23. Within the search area, use the real-time data of the lidar to perform grid-by-grid matching with the likelihood domain model map. S24. Calculate the consistency of the probability values between the lidar measurement points and the grids. S25. Use the optimization algorithm to find the pose that maximizes the overall matching score within the search area as the updated pose. After giving the initial pose, the robot calculates a predicted pose of the robot based on the real-time IMU angular velocity information and the wheel odometer linear velocity information. Then, based on the pose the scan-to-map based scan matching calculation is carried out. The calculation formula is as follows:

[0030] In the formula is to obtain the probability value of a certain grid in the likelihood domain model, is the attitude adjustment variable, is the lidar point measurement information; the meaning of the whole function is to calculate the maximum score of the entire point cloud falling in the occupied grid. This uses Ceres to perform non-linear least squares calculation, defines the above formula as an optimization problem, parameterizes the pose of the robot into translation and rotation, constructs the corresponding residual block to calculate the matching error between the point cloud data and the map, and finally obtains the maximum score by adjusting to obtain the updated pose ; Exemplarily, taking the initial pose of the robot as (0, 0, 0), that is, the coordinate origin, and the orientation as the positive direction of the x axis as an example Predicted pose calculation: According to the real-time IMU angular velocity information (for example, 0.1 rad / s) and the wheel odometer linear velocity information (for example, 0.5 m / s), calculate a predicted pose of the robot. After 1 second, the predicted pose is (0.5, 0, 0.1).

[0031] Scan matching calculation: Based on the pose, perform scan matching calculation based on scan_to_map, use Ceres to perform non-linear least squares calculation, define the above formula as an optimization problem, parameterize the pose of the robot into translation and rotation, construct the corresponding residual block to calculate the matching error between the point cloud data and the map, and finally obtain the maximum score through adjustment to obtain the updated pose as (0.52, 0.05, 0.12); It should be noted that the calculation of the updated pose is limited, including predicted pose calculation, search area determination, grid-by-grid matching, consistency calculation, and application of optimization algorithms, which makes the generation process of the updated pose clear and accurate, and improves the efficiency and accuracy of positioning update; S3. Calculate the difference between the updated pose and the predicted pose to generate a pose difference score; Specifically, obtain the predicted pose of the intelligent operation and maintenance robot and the updated pose obtained through scan matching calculation in step S2 to calculate the gap between the predicted pose and the updated pose .

[0032]

[0033] Ideally, there should be no large deviation between the predicted pose and the updated pose. If is greater than the set threshold , it is considered that there is a large difference between the prior pose and the updated pose, and there may be a risk of lost positioning; however, in the actual environment, abnormal conditions such as bumpy roads and skidding may cause deviations, so the consideration weight should be decreased, but is much greater than the set threshold , directly throw out the lost positioning flag; S4. Collect the historical pose sequence and historical speed sequence, and calculate the trajectory deviation degree score of the updated pose relative to the historical pose sequence and the speed deviation degree score of the updated speed relative to the historical speed sequence through Gaussian distribution modeling; The specific steps of step S4 are as follows: S41. Collect the historical pose sequence of the robot within the set time period and calculate the historical pose mean; S42. Collect the historical average speed of the robot within the set time period; S43. Calculate the deviation degree between the updated pose and the historical pose mean through the Gaussian distribution function; S44. Obtain the updated speed and calculate the deviation degree between the updated speed and the historical average speed through the Gaussian distribution function; Exemplarily, collect the historical pose sequence of the intelligent operation and maintenance robot within 3s , to avoid the accumulation of historical poses in the sequence, only the valid information within 3 seconds from the current moment is retained, and the average value of historical poses is calculated , timeout information is thrown; at the same time, the average speed within 3 seconds is calculated ; Since the overall running speed of the intelligent operation and maintenance robot will not exceed 1 m / s, the effective distance within 3 seconds is used to calculate the deviation degree between the updated pose and the historical trajectory :

[0034] Among them, and are the standard deviations of the deviation degree of the historical trajectory, and both are 1; In addition, there will be no phenomenon of excessive speed change in the robot, because the robot mostly runs smoothly during the task process, and then the difference between the updated speed and the historical average speed is calculated :

[0035] Among them, and are the standard deviations of the speed deviation, and both are 1; It should be noted that through the specific steps of collecting historical data, calculating the average value, and calculating the deviation degree score using Gaussian distribution modeling, the deviation degree of the robot's motion state can be accurately evaluated, providing a reliable basis for lost positioning detection; S5. Calculate the matching confidence score between the point cloud in the real-time data of the lidar and the likelihood domain model map; the specific steps of step S5 are as follows: S51. Obtain the point cloud in the real-time data collected by the lidar at the current pose; S52. Transform the point cloud in the real-time data to the coordinate system used by the likelihood domain map based on the updated pose; S53. Calculate the grid position corresponding to each point cloud in the likelihood domain map, and determine the occupancy probability value of the corresponding grid position in the likelihood domain map; S54. Calculate the matching probability score between each point cloud and the occupancy probability value, and then calculate the average value of the matching probability scores of all point clouds to obtain the matching score; S55. Use the covariance matrix during the matching process of each point cloud and the likelihood domain map as the scan matching information matrix; S56. Calculate the matching confidence score according to the matching scores of all point clouds and the average value of the scan matching information; It should be noted that the specific steps of calculating the matching confidence score are defined, including point cloud acquisition, coordinate transformation, probability value matching, matching score calculation, and information matrix application, making the calculation process of the matching confidence score clear and accurate, and improving the reliability of the detection result; S6. Perform weighted summation on the posture difference score, trajectory deviation score, speed deviation score and matching confidence score to obtain a comprehensive risk score for lost positioning; in step S6, the posture difference score, trajectory deviation score, speed deviation score and matching confidence score are weighted summated by the following formula:

[0036] in, is the pose difference score, is the trajectory deviation score, Score the speed deviation degree. is the matching score, is the scan matching information matrix, , , , is the weight of the corresponding evaluation item; for example, combined with the experience value , , , The values are set to 0.3, 0.2, 0.2, and 0.5 respectively; It should be noted that the weighted summation formula of the comprehensive location loss risk score is limited, which makes the scoring process transparent and operational, facilitates the adjustment of weights according to actual needs, and improves the flexibility and adaptability of the detection method.

[0037] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0038] like Figure 2 As shown, the following is an embodiment of the lost positioning detection system of the intelligent operation and maintenance robot of the computing power center provided by the embodiment of the present disclosure. The system and the lost positioning detection method of the intelligent operation and maintenance robot of the computing power center in the above-mentioned embodiments belong to the same inventive concept. For details not fully described in the embodiment of the lost positioning detection system of the intelligent operation and maintenance robot of the computing power center, please refer to the embodiment of the lost positioning detection method of the intelligent operation and maintenance robot of the computing power center mentioned above.

[0039] The system includes: The likelihood domain model map construction module is used to use the robot's lidar, IMU and odometer to collect real-time data to build a likelihood domain model map for the computing center room; The updated posture calculation module is used to calculate the predicted posture based on the robot's initial posture and the real-time data of the IMU and odometer, and to scan and match the likelihood domain model map based on the predicted posture and the real-time data of the lidar to generate an updated posture; A pose difference score calculation module, configured to calculate the difference between the updated pose and the predicted pose, and generate a pose difference score; A deviation degree score calculation module, configured to collect a historical pose sequence and a historical speed sequence, and calculate a trajectory deviation degree score of the updated pose relative to the historical pose sequence and a speed deviation degree score of the updated speed relative to the historical speed sequence through Gaussian distribution modeling; A matching confidence score calculation module, configured to calculate a matching confidence score between the point cloud in the lidar real-time data and the likelihood domain model map; A positioning loss comprehensive scoring module, configured to perform weighted summation on the pose difference score, the trajectory deviation degree score, the speed deviation degree score, and the matching confidence score to obtain a comprehensive positioning loss risk score.

[0040] In this embodiment, through the interaction and cooperation of the likelihood domain model map construction module, the updated pose calculation module, the pose difference score calculation module, the deviation degree score calculation module, the matching confidence score calculation module, and the positioning loss comprehensive scoring module, the positioning loss situation of the robot in a similar environment can be accurately and timely detected, the accuracy and reliability of the positioning loss detection are improved, the stable operation of the computing power center is ensured, and the operation and maintenance costs and security risks are reduced.

[0041] The positioning loss detection method for the intelligent operation and maintenance robot of the computing power center provided by the embodiment of the present application can be applied to an electronic device. Those skilled in the art can understand that the structure of the electronic device involved in the embodiment of the present invention does not constitute a limitation on the electronic device. The electronic device may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements. In the embodiment of the present invention, the electronic device includes, but is not limited to, a laptop computer, a desktop computer, a workbench, a personal digital assistant, a server, a blade server, a mainframe computer, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as, a personal digital processor, a cellular phone, a smart phone, a wearable device, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present application described and / or claimed in this article.

[0042] The electronic device may include a processor, an external memory interface, an internal memory, a universal serial bus (USB) interface, a charging management module, a power management module, a battery, a wireless communication module, an audio module, a speaker, a microphone, a sensor module, a key, a camera, a display screen, and a SIM card interface, etc.

[0043] It can be understood that the structure illustrated in the embodiments of the present application does not constitute a specific limitation on the electronic device. In other embodiments of the present application, the electronic device may include more or fewer components than those illustrated, or combine certain components, or split certain components, or have different component arrangements. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0044] The processor may include one or more processing units. For example, the processor may include a central processing unit (CPU), an application processor (AP), a modem processor, a graphics processing unit (GPU), an image signal processor (ISP), a controller, a memory, a video codec, a digital signal processor (DSP), a baseband processor, and / or a neural-network processing unit (NPU), etc. Among them, different processing units may be independent devices or integrated in one or more processors.

[0045] Among them, the processor may be the nerve center and command center of the electronic device. The controller can generate operation control signals according to the instruction operation code and timing signal to complete the control of fetching and executing instructions.

[0046] A memory may also be provided in the processor for storing instructions and data. In some embodiments, the memory in the processor is a cache memory. This memory can store the instructions or data that the processor has just used or recycled. If the processor needs to use the instruction or data again, it can be directly called from this memory. This avoids repeated accesses, reduces the waiting time of the processor, and thus improves the system efficiency.

[0047] The above-mentioned electronic device implements the method for detecting the loss of positioning of the intelligent operation and maintenance robot in the computing power center of the present application. The robot uses lidar, IMU, and odometer to collect real-time data to construct a likelihood domain model map for the computer room of the computing power center; calculates the predicted pose based on the initial pose of the robot and the real-time data of the IMU and odometer, and performs scan matching on the likelihood domain model map based on the predicted pose and the real-time lidar data to generate an updated pose; calculates the difference between the updated pose and the predicted pose to generate a pose difference score; collects the historical pose sequence and the historical speed sequence, and calculates the trajectory deviation degree score of the updated pose relative to the historical pose sequence and the speed deviation degree score of the updated speed relative to the historical speed sequence through Gaussian distribution modeling; calculates the matching confidence score between the point cloud in the real-time lidar data and the likelihood domain model map; performs weighted summation on the pose difference score, the trajectory deviation degree score, the speed deviation degree score, and the matching confidence score to obtain a comprehensive loss of positioning risk score. The technical solution can accurately and timely detect the loss of positioning of the robot in a similar environment by means of multi-sensor data fusion, construction of a likelihood domain model map, and weighted summation of various evaluation methods, improving the accuracy and reliability of loss of positioning detection, ensuring the stable operation of the computing power center, and reducing the operation and maintenance costs and safety risks.

[0048] In the storage medium provided by the present application, there is a program product that can implement the method for detecting the loss of positioning of the intelligent operation and maintenance robot in the computing power center.

[0049] The method for detecting the loss of positioning of the intelligent operation and maintenance robot in the computing power center includes: using the lidar, IMU, and odometer of the robot to collect real-time data to construct a likelihood domain model map for the computer room of the computing power center; calculating the predicted pose based on the initial pose of the robot and the real-time data of the IMU and odometer, and performing scan matching on the likelihood domain model map based on the predicted pose and the real-time lidar data to generate an updated pose; calculating the difference between the updated pose and the predicted pose to generate a pose difference score; collecting the historical pose sequence and the historical speed sequence, and calculating the trajectory deviation degree score of the updated pose relative to the historical pose sequence and the speed deviation degree score of the updated speed relative to the historical speed sequence through Gaussian distribution modeling; calculating the matching confidence score between the point cloud in the real-time lidar data and the likelihood domain model map; performing weighted summation on the pose difference score, the trajectory deviation degree score, the speed deviation degree score, and the matching confidence score to obtain a comprehensive loss of positioning risk score.

[0050] In some possible implementation manners, the method for detecting the loss of positioning of the intelligent operation and maintenance robot in the computing power center of the present disclosure may be implemented in the form of a program product, which includes program code. When the program product runs on a terminal device, the program code is used to cause the terminal device to execute the steps according to various exemplary embodiments of the present disclosure described in the "Exemplary Method" section above of this specification.

[0051] The storage medium of the present disclosure may employ any combination of one or more readable media. The readable media may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples (a non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0052] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily 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 invention. Therefore, the present invention will not be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An intelligent operation and maintenance robot positioning loss detection method for a computing power center, characterized in that, It includes the following steps: S1. Use the lidar, IMU and odometer of the robot to collect real-time data to construct a likelihood domain model map for the computing power center computer room; S2. Calculate the predicted pose based on the initial pose of the robot and the real-time data of the IMU and odometer, and perform scan matching on the likelihood domain model map based on the predicted pose and the real-time lidar data to generate an updated pose; S3. Calculate the difference between the updated pose and the predicted pose to generate a pose difference score; S4. Collect the historical pose sequence and historical speed sequence, and calculate the trajectory deviation degree score of the updated pose relative to the historical pose sequence and the speed deviation degree score of the updated speed relative to the historical speed sequence through Gaussian distribution modeling; S5. Calculate the matching confidence score between the point cloud in the real-time lidar data and the likelihood domain model map; S6. Perform weighted summation on the pose difference score, trajectory deviation degree score, speed deviation degree score and matching confidence score to obtain a comprehensive positioning loss risk score.

2. The method for detecting loss of positioning of the intelligent operation and maintenance robot in the computing power center according to claim 1, characterized in that, The specific steps of step S1 are as follows: S11. Control the robot to run in the computing power center computer room, and control the robot to collect the distance data from the environmental boundary and the pose data of the robot during the running process; S12. Perform ray projection mapping on the distance data from the environmental boundary and the pose data of the robot collected by the robot during the running process into a two-dimensional plane grid to obtain a grid map, and then use the likelihood domain model to construct a likelihood domain model map of the grid map.

3. The method for detecting the loss of positioning of the intelligent operation and maintenance robot in the computing power center according to claim 2, wherein, The specific steps of using the likelihood domain model to construct a likelihood domain model map of the grid map in step S12 are as follows: Calculate the distance between each grid and the nearest obstacle; According to the distance to the nearest obstacle, calculate the occupancy probability of each grid using the following sensor model: wherein, is the attenuation coefficient, and d is the distance from the grid to the nearest obstacle.

4. The method for detecting the loss of positioning of the intelligent operation and maintenance robot in the computing power center according to claim 2, wherein, The specific steps of step S2 are as follows: S21. Calculate the predicted pose within the target time period based on the initial pose of the robot and the real-time acceleration data collected by the IMU and the angular velocity and acceleration data collected by the odometer in real time; S22. Determine the search area on the likelihood domain model map according to the predicted pose; S23. In the search area, perform grid-by-grid matching between the real-time lidar data and the likelihood domain model map; S24. Calculate the consistency of the probability values between the lidar measurement points and the grids; S25. Find the pose that maximizes the overall matching score in the search area through an optimization algorithm as the updated pose.

5. The method for detecting the loss of positioning of the intelligent operation and maintenance robot in the computing power center according to claim 4, characterized in that, The specific steps of step S4 are as follows: S41. Collect the historical pose sequence of the robot within the set time period and calculate the historical pose mean; S42. Collect the historical average speed of the robot within the set time period; S43. Calculate the deviation degree between the updated pose and the historical pose mean through the Gaussian distribution function; S44. Obtain the updated speed and calculate the deviation degree between the updated speed and the historical average speed through the Gaussian distribution function.

6. The method for detecting the loss of positioning of the intelligent operation and maintenance robot in the computing power center according to claim 5, wherein, The specific steps of step S5 are as follows: S51. Obtain the point cloud in the real-time data collected by the lidar at the current pose; S52. Convert the point cloud in the real-time data to the coordinate system used by the likelihood domain map based on the updated pose; S53. Calculate the grid position corresponding to each point cloud in the likelihood domain map and determine the occupancy probability value of the corresponding grid position in the likelihood domain map. S54. Match the probability score of each point cloud computing with the occupied probability value, and then calculate the average value of the matching probability scores of all point clouds to obtain the matching score; S55. Use the covariance matrix during the matching process of each point cloud with the likelihood domain map as the scan matching information matrix; S56. Calculate the pose confidence score based on the matching scores of all point clouds and the average value of the scan matching information.

7. The method for detecting the loss of positioning of the intelligent operation and maintenance robot in the computing power center according to claim 6, characterized in that, In step S6, the pose difference score, the trajectory deviation degree score, the speed deviation degree score, and the pose confidence score are weighted and summed through the following formula: Among them, is the pose difference score, is the trajectory deviation degree score, is the speed deviation degree score, is the matching score, is the scan matching information matrix, 、 、 、 are the weights of the corresponding evaluation items.

8. An intelligent operation and maintenance robot positioning loss detection system for a computing power center, characterized in that, Including: A likelihood domain model map construction module, which is used to use the lidar, IMU, and odometer of the robot to collect real-time data to construct a likelihood domain model map for the computer room of the computing power center; An updated pose calculation module, which is used to calculate the predicted pose based on the initial pose of the robot and the real-time data of the IMU and odometer, and perform scan matching on the likelihood domain model map based on the predicted pose and the real-time lidar data to generate an updated pose; A pose difference score calculation module, which is used to calculate the difference between the updated pose and the predicted pose to generate a pose difference score; A deviation degree score calculation module, which is used to collect the historical pose sequence and the historical speed sequence, and calculate the trajectory deviation degree score of the updated pose relative to the historical pose sequence and the speed deviation degree score of the updated speed relative to the historical speed sequence through Gaussian distribution modeling; A pose confidence score calculation module, which is used to calculate the pose confidence score of the point cloud in the real-time lidar data and the likelihood domain model map; A lost positioning comprehensive scoring module, which is used to perform weighted summation on the pose difference score, the trajectory deviation degree score, the speed deviation degree score, and the pose confidence score to obtain a comprehensive lost positioning risk score.

9. An electronic device, characterized in that, Including a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein when the processor executes the program, the steps of the lost positioning detection method for the intelligent operation and maintenance robot of the computing power center according to any one of claims 1 to 7 are implemented.

10. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the steps of the lost positioning detection method for the intelligent operation and maintenance robot of the computing power center according to any one of claims 1 to 7 are implemented.

Images