Robot positioning and navigation system based on three-dimensional laser radar and cross photoelectric correction

Abstract

The application discloses a kind of robot positioning navigation system based on three-dimensional laser radar and cross photoelectric correction, the system includes navigation task starting module, real-time data acquisition module, positioning deviation calculation module and collaborative obstacle avoidance module, in navigation task starting module, inspection robot planning navigation path, auxiliary positioning robot constructs station track cooperative positioning reference;Real-time data acquisition module in robot real-time data acquisition and synchronization;Positioning deviation calculation module is based on the difference of cross photoelectric data and standard database, the fitting degree of auxiliary robot image and the positioning of inspection robot, calculates positioning trust, adjusts sensor parameter and navigation path;Collaborative obstacle avoidance module real-time monitoring station obstacle, ensure double robot synchronous safety start-stop, realize high-precision positioning and dynamic correction by double robot cooperation, reduce positioning error, adapt to subway night short time window inspection demand, improve inspection precision and efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of rail train inspection robot positioning, in particular to a robot positioning and navigation system based on three-dimensional laser radar and cross photoelectric correction. BACKGROUND

[0002] With the rapid development of urban rail transit, as the core public transportation mode, the operation safety and reliability of the rail system directly relate to the safety of passengers and the operation efficiency. The subway rail is subjected to train load, environmental erosion and vibration impact for a long time, and is prone to faults such as fastener loosening, track deformation and sleeper damage, which requires regular inspection and maintenance. The high-density characteristics of subway operation determine that the inspection operation is mostly carried out during the night shutdown period, which puts forward high requirements for inspection efficiency, positioning accuracy and automation degree.

[0003] Currently, subway rail inspection mainly relies on two types of technical solutions: one is manual inspection combined with handheld detection equipment, which is low in efficiency, high in labor intensity, and is easily affected by human experience, prone to missed inspection and false inspection, and difficult to meet the inspection needs of large-scale rail network; the other is an automatic inspection scheme based on a single robot, which realizes positioning and navigation by carrying laser radar, inertial navigation or visual sensor on the rail robot.

[0004] However, the existing automatic inspection scheme has obvious technical bottlenecks: the positioning accuracy is greatly affected by environmental interference, the light in the subway tunnel is dim, the dust concentration is high, and there is strong electromagnetic interference, and when the positioning is solely dependent on the three-dimensional laser radar of the inspection robot itself, the positioning drift is easily caused by the superposition of track environment point cloud data noise and feature matching deviation; although the cross photoelectric sensor can collect landmark features such as rail fasteners and sleepers, it only relies on the comparison of its own data with the standard database, lacks external independent benchmark verification, and when the local features of the track change due to wear and dirt, the positioning reliability is easily misjudged, causing the inspection robot to deviate from the target task point. In the existing scheme, if the positioning deviation of the inspection robot occurs, it is mostly dependent on the backtracking calibration of the preset path or the manual intervention of the background, which has slow correction response and limited accuracy; although some schemes introduce auxiliary positioning devices, the signboards are easily damaged due to track construction and environmental erosion, and cannot follow the dynamic movement of the inspection robot, making it difficult to realize continuous positioning verification and real-time correction of the entire inspection path.

[0005] Therefore, in order to solve the problems existing in the prior art, the present application proposes a robot positioning and navigation system based on three-dimensional laser radar and cross photoelectric correction. SUMMARY

[0006] In view of the deficiencies in the prior art, the purpose of the present application is to provide a robot positioning and navigation system based on three-dimensional laser radar and cross photoelectric correction.

[0007] To achieve the above object, the present application provides the following technical solutions:

[0008] The robot positioning and navigation system based on three-dimensional laser radar and cross photoelectric correction comprises:

[0009] The navigation task starting module receives a patrol task by the patrol robot, plans a navigation path based on a preset track path and three-dimensional laser radar scanning data, and sends a following instruction to the auxiliary positioning robot, and the auxiliary positioning robot completes self-positioning by a platform preset positioning mark and establishes a platform coordinate system.

[0010] The real-time data acquisition module collects track environment point cloud data by the three-dimensional laser radar of the patrol robot, and collects track landmark feature data by the cross photoelectric sensor; the auxiliary positioning robot synchronously follows the movement, collects appearance features and surrounding track feature data of the patrol robot, and collects real-time straight line distance data of the patrol robot.

[0011] The positioning deviation calculation module calculates the positioning trust degree of the patrol robot based on the difference between the cross photoelectric sensor data uploaded by the patrol robot and the preset track standard feature database and the fitting degree between the image data uploaded by the auxiliary positioning robot and the self-positioning data of the patrol robot, and if the positioning trust degree is higher than a preset threshold, the patrol robot continues to move along the current navigation path; otherwise, a deviation compensation value is calculated and the navigation system is corrected.

[0012] As a further improvement of the present application, the navigation task starting module comprises an encoding label containing mileage information, the auxiliary positioning robot obtains its coordinates in the platform coordinate system by recognizing the encoding label, and simultaneously establishes a mapping relationship between the encoding label and a track side preset retroreflective mark to complete the construction of a cooperative positioning reference of the platform and the track; the track side preset retroreflective mark is arranged at intervals along the outside of the track, the encoding label is arranged along the edge of the platform corresponding to the positions of the retroreflective marks, the auxiliary positioning robot photographs the retroreflective marks by an image acquisition module, and combines a distance measurement module to measure the fixed distance between the encoding label and the retroreflective marks to complete the calibration of the cooperative positioning reference.

[0013] As a further improvement of the present application, the real-time data acquisition module comprises that the auxiliary positioning robot dynamically adjusts its moving speed and direction based on the real-time positioning data of the patrol robot, ensures that the patrol robot is always within the effective field of view of the image acquisition module of the auxiliary positioning robot, and keeps the distance between the auxiliary positioning robot and the patrol robot within a preset following interval.

[0014] As a further improvement of the present application, the track signature features include track fasteners, sleepers and track mileage markers, the cross photoelectric sensor collects image data of the signature features at a preset frequency, while recording the time stamp of the collection time, the auxiliary positioning robot synchronously records the same time stamp when collecting data and completes the time synchronization of the inspection robot and the auxiliary positioning robot.

[0015] As a further improvement of the present application, the positioning deviation calculation module includes, the photoelectric data difference rate is obtained by calculating the ratio of the deviation value of the data collected by the cross photoelectric sensor and the preset standard deviation threshold, and the image fitting degree is obtained by calculating the matching degree of the feature image of the inspection robot collected by the auxiliary positioning robot and the image feature converted from the positioning data of the inspection robot itself, and the positioning trust degree is calculated according to the photoelectric data difference rate and the image fitting degree.

[0016] As a further improvement of the present application, the calculation of the image fitting degree includes, the pixel coordinates of the appearance features of the inspection robot in the image collected by the auxiliary positioning robot are converted into actual coordinates in the station coordinate system combined with the positioning data of the auxiliary positioning robot; the positioning data of the inspection robot itself is converted into coordinates in the station coordinate system; the Euclidean distance of the two groups of coordinates is calculated, and the image fitting degree is determined based on the ratio of the Euclidean distance and the preset distance threshold.

[0017] As a further improvement of the present application, the deviation compensation value calculation includes, the auxiliary positioning robot continuously collects multiple frames of image data and distance data, and obtains average image feature coordinates and average distance value after removing abnormal data; the displacement deviation and angle deviation of the positioning of the inspection robot itself and the auxiliary verification are calculated based on the average image feature coordinates, and the displacement deviation is corrected combined with the average distance value to obtain the final deviation compensation value.

[0018] As a further improvement of the present application, it further includes a cooperative obstacle avoidance module, which is communicatively connected with the inspection robot and the auxiliary positioning robot, obtains the image of the subway station and judges the distance between the auxiliary positioning robot and the obstacle, if the distance between the auxiliary positioning robot and the obstacle is less than a safety threshold, the auxiliary positioning robot stops moving and feeds back the position information, and the inspection robot synchronously stops moving.

[0019] As a further improvement of the present application, the navigation system correction comprises that after the inspection robot receives the deviation compensation value, the compensation value is decomposed into displacement correction and angle correction by the self-positioning module; the current positioning coordinates in the navigation system are updated based on the displacement correction, and the scanning reference angle of the three-dimensional laser radar and the collection field angle of the cross photoelectric sensor are adjusted in combination with the angle correction; the navigation system calls the corrected positioning coordinates and sensor parameters to regenerate the navigation path to the target task point, and controls the driving module to move according to the new path, and in the moving process, the corrected data are fed back to the positioning trust degree calculation link in real time, the correction effect of the navigation system is continuously verified until the moving state is stable and the positioning trust degree is maintained in the preset standard interval.

[0020] The beneficial effects of the present application are: improving positioning accuracy and reliability, through the double-robot collaborative architecture, the problem that the single robot positioning is easily affected by tunnel dust and electromagnetic interference is solved, through double-robot trust degree verification, the positioning accuracy is improved, and then the accuracy of track train inspection is improved. Through the cooperation of the inspection robot and the auxiliary positioning robot, the positioning trust degree is calculated by fitting the cross photoelectric sensor data combined with the auxiliary robot image, and the reference is constructed by the platform coding label and the track reflective marker, the positioning drift problem caused by tunnel dust and electromagnetic interference is solved, the long-distance inspection error accumulation is avoided, and the positioning reliability is greatly improved. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 is the robot positioning navigation system block diagram of the present application based on three-dimensional laser radar and cross photoelectric correction;

[0022] Figure 2 is the navigation task starting and collaborative positioning reference construction flow chart of the present application;

[0023] Figure 3 is the real-time data acquisition and time synchronization flow chart of the double robot of the present application;

[0024] Figure 4 is the positioning trust degree calculation and deviation judgment flow chart of the present application;

[0025] Figure 5 is the deviation compensation value calculation and correction flow chart of the present application. DETAILED DESCRIPTION

[0026] The present application will be further described in detail below in combination with the drawings and embodiments. Wherein the same parts are denoted by the same reference numerals. It should be noted that the words "front", "back", "left", "right", "up" and "down" used in the following description refer to the directions in the drawings, and the words "bottom surface" and "top surface", "inner" and "outer" refer to the directions towards or away from the geometric center of a particular part.

[0027] This invention proposes a robot positioning and navigation system based on three-dimensional lidar and cross-photoelectric correction, comprising:

[0028] The navigation task initiation module allows the inspection robot to receive the inspection task, plan a navigation path based on the preset track path and three-dimensional lidar scanning data, and send a follow command to the auxiliary positioning robot. The auxiliary positioning robot completes its own positioning and establishes the platform coordinate system through the preset positioning marks on the platform.

[0029] The navigation task initiation module is the initial trigger and benchmark establishment unit for the system to carry out inspection operations. The inspection task received by the inspection robot not only includes the scope of the target inspection section, but also specifies the specific types of track components to be inspected, such as track fasteners, concrete sleepers, and rail joints, as well as the corresponding inspection accuracy requirements, ensuring that the inspection robot clearly understands the operation objectives and quality standards. The preset track path used by the inspection robot to plan the navigation path is a digital path model constructed in advance based on subway line design drawings, actual track laying parameters, and historical inspection data. This model contains key geometric information such as track mileage coordinates, curve curvature radius, and track gradient variation, providing a basic framework for initial path planning.

[0030] During path planning, the 3D LiDAR first scans the initial track environment around the inspection robot. By continuously emitting laser beams and receiving reflected signals from the track and surrounding structures, such as tunnel walls and track support devices, it generates 3D point cloud data containing spatial coordinates and reflection intensity information. The inspection robot's path planning unit compares this point cloud data with a preset track path model, identifies subtle deviations between the initial path and the actual track, and dynamically corrects the initial path based on the comparison results. This ensures that the planned navigation path closely matches the actual track direction, avoiding navigation deviations caused by differences between the preset model and the actual environment.

[0031] Simultaneously, the inspection robot sends a follow command to the auxiliary positioning robot. This command includes the inspection robot's initial positioning coordinates, expected speed, and the platform area corresponding to the current inspection route, ensuring the auxiliary positioning robot clearly understands the spatial boundaries and speed matching benchmarks for following. The auxiliary positioning robot completes its own positioning using pre-set positioning markers on the platform. These markers are visual identifiers spaced along the platform edge, each integrating a unique mileage code and made of wear-resistant and oil-resistant polymer material, maintaining clear visibility over a long period in complex environments such as dust and humidity on subway platforms.

[0032] The image acquisition unit of the auxiliary positioning robot will take pictures of these positioning marks, extract the mileage information in the marks through image decoding algorithms, and then combine the movement distance data recorded by the odometer carried by itself to calibrate the preliminary obtained position information, and finally determine the accurate coordinates of itself in the platform plane. Based on the accurate coordinates, the auxiliary positioning robot will establish a platform coordinate system, which takes the fixed reference point at the starting end of the platform as the origin, takes the horizontal direction parallel to the edge of the platform as the X-axis, and takes the horizontal direction perpendicular to the edge of the platform and towards the track as the Y-axis, forming a unified spatial reference system for the subsequent positioning verification process, and providing a reference for the coordinate conversion between the inspection robot and the auxiliary positioning robot.

[0033] The real-time data acquisition module is responsible for acquiring multi-source data required for cooperative positioning of the two robots. Through the collaborative work of the sensors of the two types of robots, the key information such as the track environment and the robot position is comprehensively collected.

[0034] The real-time data acquisition module is responsible for acquiring multi-source data required for cooperative positioning of the two robots. Through the collaborative work of the sensors of the two types of robots, the key information such as the track environment and the robot position is comprehensively collected.

[0035] The three-dimensional laser radar carried by the inspection robot will continuously collect track environment point cloud data during movement, and its scanning range covers the track surface, the fastener system on both sides of the track, the sleeper, and the lower area of the tunnel inner wall. The scanning frequency of the laser radar will be dynamically adjusted according to the moving speed of the inspection robot, ensuring that sufficient density of point cloud data can be obtained at different moving speeds. These data not only clearly present the overall contour shape of the track, but also capture abnormal features such as slight protrusions, depressions or fastener missing on the track surface, providing environmental reference for subsequent track fault detection and positioning correction.

[0036] The cross photoelectric sensor of the inspection robot is dedicated to collecting track landmark feature data. The sensor is composed of a light signal emitting end and a receiving end. The emitting end emits near-infrared light of a specific wavelength, which has strong penetrating power in a dusty environment and can reduce environmental interference. The receiving end receives the light signal reflected by the track landmark features. The track landmark features include the shape contour of the track fastener, the cross-sectional size and spacing of the sleeper, and the mileage markings on the side of the steel rail. These features have stable geometric shape and position distribution rules, and become the core elements of positioning reference. The cross photoelectric sensor extracts the geometric parameters and spatial position information of these features by analyzing the intensity change and phase shift of the received light signal, ensuring that the collected data can accurately reflect the actual state of the track features.

[0037] The auxiliary positioning robot carries out data collection in a synchronous following manner, and the following process adopts dynamic position closed-loop control logic: the auxiliary positioning robot receives the positioning data sent by the inspection robot in real time, calculates the relative position relationship of the two based on the position information of the auxiliary positioning robot, and if it is found that the moving speed of the inspection robot changes, the drive control unit of the auxiliary positioning robot will immediately adjust the speed of the auxiliary positioning robot to ensure that the relative distance of the two in the extension direction of the track remains stable; at the same time, the vision monitoring unit of the auxiliary positioning robot continuously tracks the position of the inspection robot, and if it is found that the inspection robot has a tendency to deviate from the image acquisition field of view, it will timely adjust the moving direction to ensure that the inspection robot is always within the effective range of image acquisition, avoiding the interruption of positioning verification due to loss of vision.

[0038] In terms of data collection content, the image acquisition unit of the auxiliary positioning robot will collect two types of data at the same time: one is the appearance features of the inspection robot, including the pre-set high-contrast geometric patterns on the body of the inspection robot, and the relative position relationship of the high-contrast geometric patterns has been calibrated in advance, which has unique identification, and through the collection of these patterns, the specific position of the inspection robot in the image can be determined; the other is the track features around the inspection robot, including the reflective markers on the side of the track and the positions of the track joints, etc., which can complement the track features collected by the inspection robot and reduce the limitations of single sensor collection. In addition, the laser ranging device carried by the auxiliary positioning robot continuously collects the real-time straight-line distance from the inspection robot. The device calculates the straight-line distance between the two by emitting laser signals to specific reflective areas of the inspection robot, recording the time interval of the signal round trip, and combining the propagation speed of laser in air. The distance data will be used for subsequent correction calculation of positioning deviation to improve the positioning accuracy.

[0039] The positioning deviation calculation module calculates the positioning trust degree of the inspection robot based on the difference between the cross photoelectric sensor data uploaded by the inspection robot and the preset track standard feature database and the fitting degree of the image data uploaded by the auxiliary positioning robot and the positioning data of the inspection robot. If it is higher than the preset threshold, the inspection robot continues to move along the current navigation path; otherwise, a deviation compensation value is calculated and the navigation system is corrected.

[0040] The positioning deviation calculation module is the core unit for judging the positioning reliability of the inspection robot and realizing dynamic correction, and its operation logic is carried out around data comparison, trust quantification and deviation correction.

[0041] The module first processes two types of key data differences: one is the difference between the cross photoelectric sensor data uploaded by the inspection robot and the preset track standard feature database. The preset track standard feature database is a digital storage system based on a large amount of track feature data in normal state, which includes standard geometric parameters of different types of track fasteners, sleepers, and mileage markers, as well as standard coordinate information of these features at different mileage positions. The database is updated regularly according to the parameter changes after track maintenance to ensure the timeliness of the standard data. The module compares the feature parameters collected by the cross photoelectric sensor with the standard parameters at the corresponding mileage position in the database dimension by dimension, analyzes the geometric size deviation and spatial position offset of the features, and the comprehensive results of these indicators reflect the deviation of the data collected by the inspection robot from the standard state, which is the basis for judging the positioning accuracy.

[0042] The other type is the fitting degree calculation of the image data uploaded by the auxiliary positioning robot and the positioning data of the inspection robot itself. The process first preprocesses the image collected by the auxiliary positioning robot, improves the influence of dust and light changes on image quality through image enhancement algorithm, and then extracts the pixel coordinates of the appearance features of the inspection robot through feature extraction algorithm. Combined with the accurate position of the auxiliary positioning robot in the station coordinate system and the internal parameters of the image acquisition unit, the pixel coordinates are converted to actual spatial coordinates in the station coordinate system through coordinate conversion algorithm. At the same time, the module also converts the positioning coordinates calculated by the inspection robot itself to the same station coordinate system, and calculates the spatial position coincidence degree between the two sets of coordinates to get the fitting degree. The fitting degree directly reflects the consistency of the positioning results of the two types of robots, and the higher the fitting degree, the more reliable the self-positioning of the inspection robot.

[0043] Based on the analysis results of the above two types of data, the module calculates the positioning trust degree of the inspection robot. The positioning trust degree is a quantitative index that comprehensively reflects the positioning reliability. Its calculation process does not rely on a single type of data, but combines the difference between the cross photoelectric sensor data and the standard database, and the fitting degree of the image data according to the preset weight. The weight is set based on the reliability verification results of the two types of data in different environments to ensure that the trust degree can objectively reflect the positioning state. The final trust degree value ranges from 0 to 1, and the higher the value, the more reliable the current positioning.

[0044] The preset positioning threshold in the module is determined based on the accuracy requirement of metro track inspection, error statistics of historical positioning data, and environmental characteristics of different inspection scenes. The threshold in different scenes can be flexibly adjusted according to actual needs to balance the positioning accuracy and the response sensitivity of the system. If the calculated positioning trust degree is higher than the threshold, it indicates that the current positioning of the inspection robot is reliable, and the navigation execution unit will control the inspection robot to continue moving along the currently planned navigation path. If the positioning trust degree is lower than the threshold, it indicates that there is a positioning deviation of the inspection robot, and the module will start the deviation compensation value calculation process.

[0045] The calculation of the deviation compensation value calls multiple sets of image data and distance data collected by the auxiliary positioning robot in recent period, and filters out effective data samples through an outlier elimination algorithm. Based on these effective samples, the displacement deviation and angle deviation between the self-positioning of the inspection robot and the auxiliary positioning result are calculated. In combination with the real-time straight-line distance data collected by the auxiliary positioning robot, the displacement deviation is further corrected, and finally the deviation compensation value accurately reflecting the positioning error of the inspection robot is obtained.

[0046] After obtaining the deviation compensation value, the module triggers the navigation system correction process: the positioning correction unit of the inspection robot decomposes the deviation compensation value into displacement correction and angle correction, the displacement correction is used to adjust the current positioning coordinates recorded in the navigation system to ensure that the coordinates are consistent with the actual position of the inspection robot, and the angle correction is used to adjust the scanning reference angle of the three-dimensional laser radar to make the scanning range of the laser radar more accurately cover the track area, and adjust the collection field angle of the cross photoelectric sensor to ensure that it can accurately capture the track landmark features and avoid data deviation caused by the angle deviation of the sensor.

[0047] After the correction is completed, the navigation system calls the updated positioning coordinates and adjusted sensor parameters to regenerate the navigation path from the current position to the target task point. The path is optimized according to the corrected position and the actual trend of the track to ensure the accuracy of the path. The drive module controls the inspection robot to move according to the new path parameters. In the moving process, the system continuously collects corrected sensor data and positioning data and feeds back to the positioning trust degree calculation link to monitor the positioning state in real time until the positioning trust degree is stably maintained within the preset standard range, ensuring that the positioning accuracy always meets the work requirements in the subsequent inspection process.

[0048] Specifically, as Figures 1 to 5As shown, the navigation task starting module includes that the platform preset positioning mark is an encoding label containing mileage information, the auxiliary positioning robot obtains its coordinates in the platform coordinate system by recognizing the encoding label, and at the same time, the mapping relationship between the encoding label and the track side preset reflective mark is established to complete the construction of the cooperative positioning reference of the platform and the track. The track side preset reflective mark is arranged at intervals along the outside of the track, the encoding label is arranged along the edge of the platform corresponding to the position of the reflective mark, and the auxiliary positioning robot photographs the reflective mark through the image acquisition module, and combines the distance measurement module to measure the fixed distance between the encoding label and the reflective mark to complete the calibration of the cooperative positioning reference.

[0049] The platform preset positioning mark adopts the form of an encoding label, and the recognition process of the auxiliary positioning robot on the encoding label needs to go through multiple image processing and data analysis. First, the image acquisition module of the auxiliary positioning robot will continuously photograph the edge area of the platform. During the photographing process, the exposure parameters will be automatically adjusted according to the intensity of the ambient light. In the dark platform end area, the module will increase the exposure time and enhance the image gain to ensure that the label image is clear. In the middle of the platform where the light is directly incident, the exposure intensity will be reduced to avoid image overexposure caused by the reflection of the label surface. After obtaining the label image, the image preprocessing unit will first perform noise filtering, eliminate the dust particles and light interference points in the image through the Gaussian filtering algorithm, and then extract the contour boundary of the label through the edge detection algorithm to separate the label area from the complex platform background and avoid the interference of background elements on the encoding recognition.

[0050] The encoding analysis unit will perform grayscale processing on the separated label image, convert the color image into a grayscale image to simplify data operation, and then convert the grayscale image into a black and white binary image through a binarization algorithm to make the black and white contrast of the encoding pattern unit more distinct. Based on the binary image, the analysis unit will identify the binary information of the pattern unit row by row and column by column according to the preset decoding rule, convert the pattern signal into a digital signal, and then extract the mileage data contained in the label. At the same time, the odometer of the auxiliary positioning robot will record the cumulative displacement distance in the movement process, compare the displacement distance with the mileage data extracted from the encoding label, and if the deviation is within the range set based on the accuracy level of the odometer, the mileage data of the encoding label will be directly used to determine the coordinates of the auxiliary positioning robot in the platform coordinate system. If the deviation exceeds the range, the two types of data will be combined through a weighted fusion algorithm, with the encoding label data as the main and the odometer data as the auxiliary, to correct the coordinates of the auxiliary positioning robot and ensure the accuracy of the positioning result.

[0051] When establishing the mapping relationship, the auxiliary positioning robot first moves to the position of a certain coded label, and captures the track side reflective marker corresponding to the coded label through the image acquisition module. During the capturing process, the auxiliary positioning robot adjusts its own posture to make the lens optical axis of the image acquisition module approximately perpendicular to the center of the reflective marker, thereby reducing the deformation of the marker image caused by too large a shooting angle. At the same time, the auxiliary positioning robot turns on the light supplementing device to emit near-infrared light to the reflective marker, so that the reflective marker appears as a bright spot in the image, facilitating the identification by the image processing unit. The image processing unit extracts the spot from the image containing the reflective marker, determines the pixel coordinates of the spot in the image, and then, in combination with the intrinsic parameters of the image acquisition module and the coordinates of the auxiliary positioning robot itself, calculates the coordinates of the reflective marker in the platform coordinate system through a perspective projection transformation algorithm.

[0052] The background system calls the track coordinate system coordinates of the reflective marker in the database, associates the track coordinate system coordinates with the calculated platform coordinate system coordinates, and establishes a one-to-one correspondence between the reflective marker and the corresponding coded label, that is, the platform coordinates of the coded label can be directly associated with the track coordinates of the reflective marker, and vice versa. To ensure the stability of the mapping relationship, the auxiliary positioning robot performs multiple capturing and coordinate calculation on the same set of coded labels and reflective markers, takes the average value after removing abnormal data, and uses the average value as the final mapping coordinates to reduce the influence of single measurement error on the mapping relationship. Through the one-by-one association of all coded labels and corresponding reflective markers, a platform and track cooperative positioning reference covering the entire inspection section is finally formed, providing a unified spatial reference framework for the subsequent positioning data conversion and verification of the double robots.

[0053] Specifically, as shown in Figures 1 to 5 , the real-time data acquisition module includes that the auxiliary positioning robot dynamically adjusts its moving speed and direction based on the real-time positioning data of the inspection robot, to ensure that the inspection robot is always within the effective field of view of the image acquisition module of the auxiliary positioning robot, and the distance between the auxiliary positioning robot and the inspection robot is kept within a preset following interval.

[0054] Specifically, as shown in Figures 1 to 5 , the track landmark features include track fasteners, sleepers, and track mileage markers, the cross photoelectric sensor acquires image data of the landmark features at a preset frequency, and records the time stamp of the acquisition time, and the auxiliary positioning robot records the same time stamp and completes the time synchronization between the inspection robot and the auxiliary positioning robot when acquiring data.

[0055] Specifically, as shown in Figures 1 to 5As shown, the positioning deviation calculation module includes, through the cross photoelectric sensor data acquisition and the deviation value of the track standard feature database and the preset standard deviation threshold value, the photoelectric data difference rate is obtained, and the matching degree of the image feature converted from the inspection robot feature image collected by the auxiliary positioning robot and the positioning data of the inspection robot itself is calculated to obtain the image fitting degree, and the positioning trust degree is calculated according to the photoelectric data difference rate and the image fitting degree.

[0056] Specifically, as shown in the figure, Figures 1 to 5 The calculation of the image fitting degree includes converting the pixel coordinates of the appearance features of the inspection robot in the image collected by the auxiliary positioning robot into actual coordinates in the station coordinate system combined with the positioning data of the auxiliary positioning robot; the positioning data of the inspection robot itself is converted into coordinates in the station coordinate system; the Euclidean distance of the two groups of coordinates is calculated, and the image fitting degree is determined based on the ratio of the Euclidean distance to the preset distance threshold.

[0057] Specifically, as shown in the figure, Figures 1 to 5 The deviation compensation value calculation includes that the auxiliary positioning robot continuously collects multiple image data and distance data, and removes abnormal data to obtain average image feature coordinates and average distance value; based on the average image feature coordinates, the displacement deviation and angle deviation of the positioning of the inspection robot itself and the auxiliary verification positioning are calculated, and the displacement deviation is corrected combined with the average distance value to obtain the final deviation compensation value.

[0058] The raw data collected by the cross photoelectric sensor contains a large amount of environmental interference signals, so before comparison with the standard database, the data needs to be preprocessed. The preprocessing link is completed by the signal purification unit and the feature reconstruction unit: the signal purification unit adopts an adaptive median filtering algorithm, dynamically adjusts the filtering window size according to the distribution density of noise points in the image, can eliminate isolated dust noise points, and can also retain the edge details of the track landmark features, avoiding excessive filtering leading to feature blur; the feature reconstruction unit is aimed at the features blocked by local stains, through the neighborhood feature interpolation algorithm, based on the feature parameters of the unblocked area, the feature data of the blocked area is completed, ensuring the integrity of the data to be compared, and avoiding deviation misjudgment caused by local shielding.

[0059] The track standard feature database is not a static image set, but a dynamic feature model library constructed based on track component industry standards, actual installation parameters and full life cycle operation and maintenance data. In the database, each track landmark feature contains multi-dimensional standard parameters: taking track fasteners as an example, the parameters cover the length, width and height of the fastener base, the bending radius of the elastic component, the center distance of the bolts and the diameter of the bolt head, and even the normal wear parameter range under different service life. The database will store these parameters by subway line mileage segments, and each mileage interval corresponds to a unique set of feature parameters to ensure that the inspection robot can call the standard data of the corresponding position for comparison.

[0060] In the multi-dimensional deviation extraction process, the calculation module will compare the pre-processed collected data with the standard parameters of the corresponding mileage in the database one by one to extract the deviation value of each dimension. Taking the track sleeper as an example, the actual cross-sectional width deviation, actual spacing deviation, and actual height deviation of the sleeper in the collected data are calculated respectively; for track fasteners, the actual center distance deviation of the bolts and the actual length deviation of the fastener base are calculated. Then, the multi-dimensional deviation values are fused by weighted average algorithm, and the influence weight of different dimensional parameters on positioning accuracy is calculated, such as the reference value of the fastener bolt spacing on positioning is higher than that of the fastener base height, so a higher weight is given, and the comprehensive deviation value is calculated, which can fully reflect the overall deviation degree of the collected data from the standard state.

[0061] The preset standard deviation threshold is a dynamic threshold determined based on the accuracy requirements of subway track inspection and historical data statistics. For different track landmark features, the threshold setting is different: for example, the track mileage marker is directly related to the mileage coordinate, and has the greatest influence on positioning accuracy, so the threshold setting is more strict; the sleeper spacing has a relatively weak auxiliary reference effect on positioning, so the threshold can be appropriately relaxed. At the same time, the threshold will be dynamically adjusted according to the service life of the track: the feature parameters of newly laid track are stable, and the threshold is small; the feature parameters of the track with long service life fluctuate greatly due to natural wear and tear, and the threshold will be appropriately increased to avoid misjudgment of positioning deviation due to normal wear and tear.

[0062] The conversion formula of photoelectric data difference rate is difference rate = comprehensive deviation value / preset standard deviation threshold, and the value range of difference rate is 0 to 1. When the comprehensive deviation value is 0, the difference rate is 0, indicating that the collected data is completely consistent with the standard data; when the comprehensive deviation value is equal to the preset standard deviation threshold, the difference rate is 1, indicating that the collected data has reached the upper limit of the allowed deviation; if the comprehensive deviation value exceeds the threshold, the difference rate will be greater than 1, at which point it is directly determined that the current photoelectric data cannot support reliable positioning, and the deviation correction process needs to be triggered first.

[0063] Image fitting degree is a core index for measuring the consistency of the image collected by the auxiliary positioning robot and the positioning data of the inspection robot, ensuring that the results accurately reflect the coordination of the positioning of the two robots.

[0064] The feature image of the inspection robot collected by the auxiliary positioning robot needs to be processed by feature enhancement and target segmentation before effective matching features can be extracted. The feature enhancement part uses the Retinex image enhancement algorithm to separate the illumination component and the reflection component of the image, eliminate the impact of uneven light in the tunnel on image quality, and significantly improve the contrast of the appearance features of the inspection robot. The target segmentation part is based on a semantic segmentation algorithm, which separates the inspection robot region in the image from the track and tunnel background, avoiding interference from background elements in feature extraction. After segmentation, the feature extraction unit uses the ORB feature detection algorithm to extract key points of the appearance features of the inspection robot, such as the vertices of geometric markers and edge turning points, and calculates the descriptor of each key point, which contains information such as the gray distribution and neighborhood pixel gradient direction of the key point, which can be used for subsequent feature matching.

[0065] The conversion of the positioning data of the inspection robot itself relies on the previously established station and track collaborative positioning reference. The positioning data of the inspection robot is generated based on the track coordinate system and includes longitudinal coordinates along the track mileage direction, transverse coordinates perpendicular to the track direction, and attitude angles. In the conversion process, the calculation module first calls the mapping relationship between the track coordinates and the station coordinates in the collaborative positioning reference to convert the track coordinate system coordinates of the inspection robot into theoretical coordinates in the station coordinate system. At the same time, combined with the attitude angle of the inspection robot, the theoretical position of the key points of its appearance features in the station coordinate system is calculated. For example, if the inspection robot has an attitude offset, the actual position of its body geometric markers will be different from that without offset, and the theoretical coordinates need to be corrected through the attitude angle to ensure that the theoretical position matches the actual attitude of the inspection robot.

[0066] The feature matching quantization part is not simply a judgment of whether the features coincide or not, but a quantitative evaluation through key point matching, coordinate deviation calculation, and fitting degree conversion. First, the feature key point descriptors extracted by the auxiliary positioning robot are matched with the theoretical key point descriptors converted from the positioning data of the inspection robot. By calculating the Hamming distance between the descriptors, key point pairs with high matching degree are selected, and the RANSAC algorithm is used to remove false matching key point pairs to ensure the reliability of the matching results.

[0067] Subsequently, the Euclidean distance of each pair of matched key points in the station coordinate system is calculated, which can reflect the spatial deviation of the actual collected key point position and the theoretical key point position. In order to avoid the fitting degree misjudgment caused by the deviation of a single key point, the average Euclidean distance of all effective matching key point pairs is calculated, which can comprehensively reflect the overall deviation of the positioning data of the dual robots. The preset distance threshold is set based on the accuracy of the image acquisition module of the auxiliary positioning robot and the positioning verification requirements.

[0068] The conversion formula of image fitting degree is fitting degree = 1 - (average Euclidean distance / preset distance threshold), and the value range of fitting degree is 0 to 1. When the average Euclidean distance is 0, the fitting degree is 1, indicating that the feature position collected by the auxiliary positioning robot is completely consistent with the positioning of the inspection robot itself; when the average Euclidean distance is equal to the preset distance threshold, the fitting degree is 0, indicating that the positioning deviation has reached the upper limit allowed; if the average Euclidean distance exceeds the threshold, the fitting degree will be less than 0, at which time it is directly judged that there is a significant deviation in the positioning data of the dual robots, and the correction process needs to be started.

[0069] The positioning trust degree is not a simple addition of the photoelectric data difference rate and the image fitting degree, but a weighted fusion based on the reliability weight of the two, and its calculation process needs to combine environmental adaptability analysis and dynamic weight adjustment to ensure that the trust degree can objectively reflect the positioning state in different environments.

[0070] After the positioning trust degree is calculated, the module compares it with the preset trust degree threshold. The preset trust degree threshold also needs to be set in combination with the accuracy requirements of the inspection scene: in the inspection scene of track fasteners, rail joints and other scenes that require high-precision positioning, the threshold is set higher to ensure that only when the positioning is extremely reliable can the inspection continue; in the scene of overall track contour detection and other scenes with lower precision requirements, the threshold can be appropriately reduced to balance the positioning accuracy and the inspection efficiency. If the trust degree is higher than the threshold, it indicates that the current positioning state meets the inspection requirements, and the calculation module will send a continue navigation instruction to the navigation execution unit of the inspection robot; if the trust degree is lower than the threshold, the calculation module will immediately start the deviation compensation value calculation process and send a pause navigation instruction to the navigation execution unit to avoid invalidation of the inspection data or deviation of the robot from the track due to positioning deviation.

[0071] In addition, the calculation module also performs continuous trend analysis on the positioning trust degree. If the trust degree is higher than the threshold but shows a continuous downward trend, it indicates that the positioning state of the inspection robot is gradually deteriorating and there may be a potential deviation risk. At this time, the module will send a positioning early warning signal to the background system in advance, and the background system can actively adjust the acquisition frequency of the inspection robot or the following strategy of the auxiliary positioning robot to prevent the positioning trust degree from further falling below the threshold, realize the early control of the positioning risk, and improve the stability of the system operation.

[0072] Specifically, as Figures 1 to 5As shown, the cooperative obstacle avoidance module is in communication connection with the inspection robot and the auxiliary positioning robot, acquires images of the subway platform and judges the distance between the auxiliary positioning robot and the obstacle, and if the distance between the auxiliary positioning robot and the obstacle is less than a safety threshold, the auxiliary positioning robot stops moving and feeds back position information, and the inspection robot synchronously stops moving.

[0073] Specifically, as shown in the figure, Figures 1 to 5 As shown, the navigation system correction includes that after the inspection robot receives the deviation compensation value, the compensation value is decomposed into displacement correction and angle correction by the self-positioning module; the current positioning coordinates in the navigation system are updated based on the displacement correction, and the scanning reference angle of the three-dimensional laser radar and the collection field of view angle of the cross photoelectric sensor are adjusted in combination with the angle correction; the navigation system calls the corrected positioning coordinates and sensor parameters to regenerate the navigation path to the target task point, and controls the driving module to move according to the new path, and in the moving process, the corrected data is fed back to the positioning trust degree calculation link in real time, the correction effect of the navigation system is continuously verified until the moving state is stable and the positioning trust degree is maintained in the preset qualified interval.

[0074] The above shows and describes the basic features, principles and advantages of the present application. It should be noted that the present application is not limited by the above-mentioned embodiments, and only some embodiments are provided. Without departing from the spirit and scope of the present application, several improvements and supplements are considered as the protection scope of the present application.

Claims

1. A robot positioning and navigation system based on three-dimensional laser radar and cross photoelectric correction, characterized in that, The application relates to a navigation task starting module, a real-time data acquisition module and a positioning deviation calculation module. The navigation task starting module comprises a station preset positioning mark which is a coded label containing mileage information, and the auxiliary positioning robot obtains its coordinates in the station coordinate system by recognizing the coded label, and simultaneously establishes a mapping relationship between the coded label and track side preset reflective markers to complete the construction of a cooperative positioning reference of the station and the track; the track side preset reflective markers are arranged at intervals along the outside of the track, the coded label is arranged along the edge of the station corresponding to the positions of the reflective markers, the auxiliary positioning robot photographs the reflective markers through an image acquisition module, and the fixed distance between the coded label and the reflective markers is measured through a distance measurement module to complete the calibration of the cooperative positioning reference. The real-time data acquisition module comprises that the auxiliary positioning robot dynamically adjusts its moving speed and direction based on the real-time positioning data of the inspection robot, so that the inspection robot is always within the effective field of view of the image acquisition module of the auxiliary positioning robot, and the distance between the auxiliary positioning robot and the inspection robot is kept within a preset following interval. The track landmark features comprise track fasteners, sleepers and track mileage markers, the cross photoelectric sensor collects image data of the landmark features at a preset frequency, simultaneously records the time stamp of the collection time, and the auxiliary positioning robot synchronously records the same time stamp when collecting data and completes time synchronization between the inspection robot and the auxiliary positioning robot.

2. The robot positioning navigation system based on three-dimensional laser radar and cross photoelectric correction according to claim 1, characterized in that, The positioning deviation calculation module comprises that a photoelectric data difference rate is obtained by calculating the ratio of the deviation value of the data collected by the cross photoelectric sensor and the preset track standard feature database and a preset standard deviation threshold, an image fitting degree is obtained by calculating the matching degree of the inspection robot feature image collected by the auxiliary positioning robot and the image feature converted from the self-positioning data of the inspection robot, and the positioning trust degree is calculated according to the photoelectric data difference rate and the image fitting degree.

3. The robot positioning navigation system based on three-dimensional laser radar and cross photoelectric correction according to claim 1, characterized in that, ​ 4. The robot positioning navigation system based on three-dimensional laser radar and cross photoelectric correction according to claim 1, characterized in that, ​ 5. The robot positioning navigation system based on three-dimensional lidar and cross photoelectric correction of claim 1, wherein, ​ 6. The robot positioning navigation system based on three-dimensional laser radar and cross photoelectric correction according to claim 5, characterized in that, The image fitting degree calculation includes converting pixel coordinates of appearance features of the inspection robot in images collected by the auxiliary positioning robot into actual coordinates in the station coordinate system in combination with positioning data of the auxiliary positioning robot; converting self-positioning data of the inspection robot into coordinates in the station coordinate system; calculating Euclidean distances of the two groups of coordinates, and determining the image fitting degree based on a ratio of the Euclidean distances to a preset distance threshold.

7. The robot positioning navigation system based on three-dimensional lidar and cross photoelectric correction of claim 1, wherein, The deviation compensation value calculation includes that the auxiliary positioning robot continuously collects multiple frames of image data and distance data, and obtains average image feature coordinates and an average distance value after removing abnormal data; based on the average image feature coordinates, displacement deviation and angle deviation of self-positioning of the inspection robot and auxiliary verification positioning are calculated, the displacement deviation is corrected in combination with the average distance value, and a final deviation compensation value is obtained.

8. The robot positioning navigation system based on three-dimensional lidar and cross photoelectric correction of claim 1, wherein, The cooperative obstacle avoidance module is further included, which is communicatively connected with the inspection robot and the auxiliary positioning robot, acquires images of the subway station, and judges a distance between the auxiliary positioning robot and an obstacle; if the distance between the auxiliary positioning robot and the obstacle is less than a safety threshold, the auxiliary positioning robot stops moving and feeds back position information, and the inspection robot synchronously stops moving.

9. The robot positioning navigation system based on three-dimensional lidar and cross photoelectric correction of claim 1, wherein, The navigation system correction includes that after the inspection robot receives the deviation compensation value, the compensation value is decomposed into displacement correction and angle correction by the self-positioning module; the current positioning coordinates in the navigation system are updated based on the displacement correction, and the scanning reference angle of the three-dimensional laser radar and the collection field of view angle of the cross photoelectric sensor are adjusted in combination with the angle correction; The navigation system calls the corrected positioning coordinates and sensor parameters, regenerates a navigation path to a target task point, and controls the driving module to move according to the new path; in the moving process, corrected data is fed back to the positioning trust degree calculation link in real time, the correction effect of the navigation system is continuously verified, and the moving state is stable and the positioning trust degree is maintained in a preset qualified interval.

Images