An intelligent control system for an underbody inspection robot

Abstract

The present invention discloses an intelligent control system for a vehicle bottom detection robot, which relates to the technical field of vehicle bottom detection. The data acquisition module is configured to collect omnidirectional data information of the target vehicle bottom, and the data information at least includes the position information of the vehicle bottom targets, the spacing information between different vehicle bottom targets, and the contour information of the vehicle bottom targets. The data analysis module is connected to the data acquisition module and is configured to perform: deeply analyze the collected data to identify the types, positions, shapes, and size characteristics of each target on the vehicle bottom. Through the optimized movement path generated by the path planning module in combination with the distribution and parameter conditions of the vehicle bottom targets to be detected, the present invention can reduce unnecessary acceleration, deceleration, and turning operations of the robot compared with the traditional path planning, reduce energy consumption, and improve the detection efficiency at the same time.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle bottom detection, and more specifically, to an intelligent regulation system for a vehicle bottom detection robot. Background Art

[0002] A vehicle bottom detection robot is an intelligent device used to detect the condition of the vehicle bottom and is widely applied in fields such as subways, high-speed rails, public security, and customs;

[0003] In the vehicle bottom target detection technology, the vehicle bottom detection robot usually performs detection along a preset fixed route. This detection method requires the robot to frequently adjust its parameters during the detection process to adapt to different detection tasks and environments. However, this frequent parameter adjustment not only increases the complexity of the operation but also significantly reduces the grasping efficiency of the robot. Summary of the Invention

[0004] To solve the above problems, the present invention provides an intelligent regulation system for a vehicle bottom detection robot.

[0005] The present invention provides an intelligent regulation system for a vehicle bottom detection robot, including:

[0006] A data acquisition module configured to acquire omnidirectional data information of the target vehicle bottom, where the data information at least includes the position information of the vehicle bottom targets, the spacing information between different vehicle bottom targets, and the contour information of the vehicle bottom targets;

[0007] A data analysis module connected to the data acquisition module and configured to perform: deeply analyze the acquired data to identify the types, positions, shapes, and size characteristics of each target on the vehicle bottom;

[0008] Use a data analysis algorithm to statistically analyze the target distribution law and determine the detection parameters corresponding to each target;

[0009] A path planning module connected to the data analysis module and configured to: generate a movement path based on the target characteristics and distribution law information and in combination with the kinematic parameters of the robot.

[0010] Preferably, the specific steps of generating a movement path based on the vehicle bottom target characteristics and target distribution information given by the data analysis module and in combination with the movement parameters of the vehicle bottom detection robot are as follows:

[0011] A1: Divide the vehicle bottom space into multiple detection regions, and each region correspondingly contains the first type of target to be detected or the second type of target to be detected;

[0012] A2: Perform the following operations on the region where the first type of target to be detected is located:

[0013] Obtain the detection parameters corresponding to each target in this region;

[0014] Generate a movement path in the order of monotonically increasing or decreasing adjustment amplitude according to the detection parameters;

[0015] A3: Execute the following for the area where the second type of target to be detected is located:

[0016] Calculate the distance between each target in this area;

[0017] Generate a movement path according to the principle of minimizing the difference in movement distance between adjacent targets.

[0018] Preferably, the method for dividing the detection area specifically includes:

[0019] Determine the accuracy requirements of the detection parameters for each target to be detected;

[0020] Classify multiple targets with detection parameter accuracy requirements within a preset fixed range into the first type of target area to be detected;

[0021] Classify multiple targets with detection parameter accuracy allowing dynamic adjustment into the second type of target area to be detected.

[0022] Preferably, the specific manner in which the path planning module selects a movement path is:

[0023] When the detection path needs to extend to an adjacent area:

[0024] If the next area contains the first type of target to be detected:

[0025] Obtain the detection parameters of the target at the end of the current path;

[0026] Calculate the adjustment amplitude of the detection parameters of the targets in the adjacent area;

[0027] Select the target with the smallest adjustment amplitude of the detection parameters as the next path node;

[0028] If the next area contains the second type of target to be detected:

[0029] Calculate the distance between the end position of the current path and each target in the adjacent area;

[0030] Select the target with the smallest distance as the next path node.

[0031] Preferably, the path planning module is further configured to re-evaluate the path according to the precise parameters required by the current target, specifically:

[0032] Obtain an initial path, where the initial path is the current position of the detection robot , and use this as the starting point and the target position as the end point to plan a shortest straight-line path , the length of which ;

[0033] According to the precise parameters required for the current target provided by the data analysis module, through the formula , obtain the time required for parameter adjustment , where represents the current detection angle of the robot, represents the optimal detection angle required to complete the high-precision detection of the current target, refers to the adjustment speed of the robot's detection angle, that is, the number of degrees by which the robot can change the detection angle per unit time;

[0034] Calculate the initial path movement time;

[0035] Compare the times to obtain the results of whether to re-plan the route and whether to re-plan the route;

[0036] For the need to re-plan the route, introduce a new heuristic function , on the basis of the original heuristic function that only considers distance, add a detour penalty term to obtain , where is used to measure the degree of path detour, and is used to measure the degree of detour of the path from the starting point to the node this path, is the weight coefficient, used to control the degree of penalty for detours, represents a node in the path search process, where i is the index of this node, is the newly designed heuristic function, used to evaluate the estimated path cost from the node to the target position ; is the original heuristic function that only considers distance, is the node to the target position distance;

[0037] Taking the current position of the robot as the starting point, use the improved A* algorithm, and based on the new heuristic function to find the intermediate point , , , , ;

[0038] Calculate the time to move along the adjusted path to the target to see if it can meet the time required for parameter adjustment;

[0039] If it meets the requirement, this path is the path that meets the parameter adjustment requirements; if it does not meet the requirement, adjust the weight coefficient in the heuristic function and re-search for the path;

[0040] The degree of detour of the path is evaluated by comparing the lengths of the adjusted path and the initial path.

[0041] Preferably, the specific steps for calculating the movement time of the initial path are as follows:

[0042] Obtain the provided speed v of the robot;

[0043] Through the formula , calculate the time to move along the initial path to the target.

[0044] Preferably, the specific result of comparing the times to obtain the results of needing to re-plan the route and not needing to re-plan the route is as follows:

[0045] Compare with , and judge whether holds;

[0046] If it holds, it means that the initial path cannot meet the time required for parameter adjustment and the path needs to be re-planned. Otherwise, the initial path meets the requirements and there is no need to re-plan.

[0047] Preferably, the specific method for evaluating the degree of detour of the path is as follows:

[0048] Calculate the detour factor through the formula to obtain, where the detour factor is an index used to quantify the degree of detour of the path. By comparing the proportional relationship between the length of the adjusted path and the length of the initial shortest straight path, it reflects the detour situation of the path relative to the shortest path. refers to the length of the adjusted path, which is the total length of the path obtained after re-planning considering that the robot needs to meet the parameter adjustment requirements. As a reference length, it is used to compare with the length of the adjusted path to evaluate whether the path has a detour due to meeting the parameter adjustment requirements.

[0049] Preferably, the path planning module further includes:

[0050] Execute the serialized detection operation along the planned movement path. When an abnormality of the current target component to be detected is detected, trigger the correlation analysis instruction;

[0051] According to the preset fault propagation model and component dependency graph, it is determined that all components functionally associated with the to-be-detected target component with current problems need to be synchronously overhauled. At the same time, based on the fault history data, real-time operating condition parameters, and system operating status of all components functionally associated with the to-be-detected target component with current problems, a detection exemption decision for all components functionally associated with the to-be-detected target component with current problems is generated.

[0052] In response to the detection exemption decision, an updated path is generated through an online replanning algorithm.

[0053] Preferably, the generation of the updated path through the online replanning algorithm is specifically as follows:

[0054] Remove the positions of all components functionally associated with the to-be-detected target component with current problems from the initial path node sequence.

[0055] Recalculate the optimal traversal order of the remaining nodes.

[0056] Beneficial effects: Through the path generated by the path planning module, the underbody robot can move methodically under the vehicle, avoiding missing detection areas, improving detection efficiency, saving detection time, and generating an optimized movement path in combination with the distribution and parameter conditions of the underbody targets to be detected. Compared with traditional path planning, it can reduce unnecessary acceleration, deceleration, and turning operations of the robot, reduce energy consumption, and improve detection efficiency at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 is a flowchart of the regulation system of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0058] As Figure 1 shown: An intelligent regulation system for an underbody inspection robot includes:

[0059] A data acquisition module configured to collect all-round data information of the target underbody. The data information at least includes the position information of the underbody targets, the spacing information between different underbody targets, and the contour information of the underbody targets. Specifically, this module comprehensively uses a variety of sensors. For example, a lidar emits laser beams and receives reflected light, which can accurately measure the distance between the underbody target and the robot, and then obtain the position information of the underbody target. When measuring the position, its accuracy can reach the millimeter level. An industrial camera is equipped with a special wide-angle lens and takes pictures of the underbody from different angles. Through image processing algorithms, it can extract the contour information of the underbody targets. For example, the exhaust pipe and suspension components of the underbody can clearly show their contours. The ultrasonic sensor array is used to measure the spacing information between different underbody targets. It uses the reflection characteristics of ultrasonic waves to quickly and accurately obtain the target spacing data.

[0060] Actual effect: Taking the inspection of the car underbody as an example, through the data acquisition module, we can clearly obtain the precise positions of various components under the car, such as the engine oil pan, transmission, etc., as well as the distances between them. At the same time, we can also get the contours of these components, providing a rich and accurate data basis for subsequent analysis.

[0061] The data analysis module, connected to the data acquisition module, is configured to perform: conduct in-depth analysis on the acquired data to identify the types, positions, shapes, and size characteristics of each target under the car;

[0062] Use data analysis algorithms to statistically analyze the target distribution law and determine the detection parameters corresponding to each target;

[0063] Specifically, first, use advanced image recognition algorithms and point cloud analysis techniques to conduct in-depth analysis on the acquired data. For example, by using a deep learning model to identify the images captured by an industrial camera, it is possible to accurately determine the types of each target under the car, whether it is a tire, a shock absorber, or other components. At the same time, combining the data from lidar and ultrasonic sensors, accurately determine the position, shape, and size characteristics of the target. Then, use data analysis algorithms to statistically analyze a large amount of detection data to find the distribution law of the targets under the car. For example, statistically analyze the common distribution positions and occurrence frequencies of components under the car of different models. Finally, according to the characteristics of the target, such as type, position, etc., determine the detection parameters corresponding to each target, such as for different types of components under the car, decide which resolution camera to use, what detection frequency, etc.;

[0064] Actual effect: Through the operation of this module, we can clearly understand the specific situation of each component under the car, and formulate a targeted detection plan according to the characteristics of the components, greatly improving the accuracy and efficiency of detection.

[0065] The path planning module, connected to the data analysis module, is configured to: generate a movement path based on the target characteristics and distribution law information, and combine the kinematic parameters of the robot.

[0066] Specifically, based on the target characteristics and distribution law information obtained by the data analysis module, and at the same time considering the kinematic parameters of the robot itself, such as the turning radius and moving speed of the robot. Use path planning algorithms, such as the A* algorithm, etc., to calculate the optimal movement path of the robot under the car. For example, if the target distribution is dense in certain areas under the car, the path planning module will plan a path that can efficiently cover these areas; if there are narrow spaces under the car where the robot's turning is restricted, it will also plan a suitable turning path according to its kinematic parameters;

[0067] Actual effect: Through the path generated by the path planning module, the underbody robot can move methodically under the vehicle, avoiding missing detection areas, reducing unnecessary movement, improving detection efficiency, and saving detection time.

[0068] As a further embodiment, based on the underbody target features and target distribution information given by the data analysis module, the specific steps for generating a movement path in combination with the movement parameters of the underbody detection robot are as follows:

[0069] A1: Divide the underbody space into multiple detection areas, and each area correspondingly contains the first type of target to be detected or the second type of target to be detected;

[0070] A2: Execute for the area where the first type of target to be detected is located:

[0071] Obtain the detection parameters corresponding to each target in this area;

[0072] Generate a movement path in the order of monotonically increasing or decreasing adjustment amplitude according to the detection parameters;

[0073] Specifically, the underbody space structure is complex and contains many different types of components. In order to perform detection more efficiently, the underbody space needs to be divided into multiple detection areas. The division basis is the type of underbody target, that is, each area correspondingly contains the first type of target to be detected or the second type of target to be detected. For example, the area of the underbody containing suspension system components (such as shock absorbers, springs, etc.) can be divided into the area of the first type of target to be detected, because the detection focus of such components is on structural integrity, wear degree, etc., and the detection parameters are relatively unified. And the area of the underbody containing electrical circuits, fuel pipelines, etc. is divided into the area of the second type of target to be detected. The detection parameters of such components vary greatly. For example, electrical circuits focus on insulation performance detection, and fuel pipelines focus on sealing performance detection. Through this division method, the underbody detection robot can adopt more targeted detection strategies and movement path planning in different areas;

[0074] In the area where the first type of target to be detected is located, the data analysis module has determined the detection parameters corresponding to each target according to the type, position, shape and size characteristics of the target. For example, for the shock absorber in the suspension system, the detection parameters may include that the resolution of the camera needs to reach 1080P to clearly capture the fine cracks on the surface, and the detection frequency is once every 5 seconds; for the spring, it may be required to use light of a specific wavelength to detect the stress concentration area on its surface, etc.;

[0075] Generate a movement path in the order of monotonically increasing or decreasing adjustment amplitude according to the detection parameters. Assume that the detection parameter is the camera resolution. In this area, due to different degrees of importance or potential failure risks of shock absorbers at different positions, cameras with different resolutions may be required for detection. Starting from the shock absorber with the lowest resolution requirement, the robot moves along a path to the positions of shock absorbers with gradually increasing resolution requirements in turn. This can avoid the robot frequently making large adjustments to the detection parameters and improve the detection efficiency. For example, move from a shock absorber at the bottom edge of the vehicle body, which is relatively less critical and requires a resolution of 720P, to a shock absorber near the axle, which requires a resolution of 1080P, and then to a shock absorber near the engine, which requires a resolution of 4K, to form a monotonically increasing detection path.

[0076] A3: Execute the following for the area where the second type of target to be detected is located:

[0077] Calculate the distance between each target in this area;

[0078] Generate a movement path according to the principle of minimizing the difference in movement distance between adjacent targets.

[0079] Specifically, in the area where the second type of target to be detected is located, due to the diverse types of targets and large differences in detection parameters, it is necessary to first calculate the distance between each target in this area. For example, in an area where electrical lines and fuel pipes coexist, it is necessary to accurately measure the distance between each section of the electrical line and the adjacent fuel pipe. Through the data obtained by lidar and ultrasonic sensors and combined with spatial coordinate calculations, accurate distance information can be obtained. Assume that the closest distance between a section of electrical line and the adjacent fuel pipe is 20 centimeters;

[0080] Generate a movement path according to the principle of minimizing the difference in movement distance between adjacent targets. This means that when the underbody inspection robot moves from one target to the next, it tries to keep the change in movement distance to a minimum. For example, in the above area, the robot moves from an electrical connection point to the fuel pipe interface closest to it, and then to another electrical connection point adjacent to it with the smallest change in distance. This can reduce unnecessary acceleration, deceleration, and turning operations of the robot, reduce energy consumption, and improve the detection efficiency at the same time. If this principle is not followed, the robot may make large jumping movements in the area, not only wasting time and energy, but also potentially missing some potential detection areas.

[0081] As a further embodiment, the method for dividing the detection area specifically includes:

[0082] Determine the accuracy requirements for the detection parameters of each target to be detected; it should be understood that during the underbody detection process, different targets to be detected have different requirements for the accuracy of detection parameters due to their functions, structures, and importance during vehicle operation. For example, for the brake pipelines on the underbody, since they are related to the safety of the vehicle braking system, any minor leakage or damage may lead to serious consequences, so the accuracy requirements for their detection parameters are extremely high. When detecting the wall thickness of the brake pipeline, it may be necessary to be accurate to 0.01 mm, and the pressure detection accuracy requirement reaches ±0.05 MPa. For some decorative components on the underbody, such as plastic splash guards, whose main function is to protect the underbody components from mud and water splashing, the accuracy requirements for detection parameters are relatively low. When detecting the surface flatness, an accuracy of 1 mm can meet the requirements. By comprehensively considering factors such as the function, quality standard, and potential risks of each target to be detected on the underbody, determine its specific accuracy requirements for detection parameters.

[0083] Classify multiple targets whose accuracy requirements for detection parameters are within a preset fixed range into the first type of target area to be detected; it should be understood that the preset fixed range is formulated based on long-term underbody detection experience and industry standards. For some underbody targets with similar functions and quality requirements, their accuracy requirements for detection parameters often concentrate within a relatively stable range. For example, the components of the suspension system on the underbody, including shock absorbers, springs, lower control arms, etc., are mainly responsible for the driving stability and comfort of the vehicle. When detecting the dimensions, shapes, and mechanical properties of these components, the accuracy requirements for detection parameters are relatively fixed. For example, the detection accuracy requirement for the diameter of the shock absorber piston rod is within ±0.05 mm, and the detection accuracy requirement for the elastic coefficient of the spring is within ±5%. Classify multiple targets such as these suspension system components whose accuracy requirements for detection parameters are within the preset fixed range into the first type of target area to be detected. The advantage of such classification is that when the underbody detection robot conducts detection within this area, relatively unified detection equipment parameters and detection processes can be adopted. For example, industrial cameras with the same resolution and measurement sensors with the same accuracy level can be used to detect the targets within the area at a fixed detection frequency, improving the detection efficiency and accuracy, and at the same time facilitating the management and analysis of detection data.

[0084] Classify multiple targets whose accuracy of detection parameters allows dynamic adjustment into the second type of target area to be detected.

[0085] It should be understood that there are some targets under the vehicle chassis, and the accuracy requirements of their detection parameters vary depending on the usage conditions of the vehicle, the operating environment, and the state changes of the targets themselves, that is, the accuracy of the detection parameters allows dynamic adjustment. For example, for the electrical circuit under the vehicle chassis, in the new vehicle state, the detection accuracy requirement for its insulation resistance may be relatively low. However, as the vehicle usage time increases, the electrical circuit may be affected by factors such as aging and wear. At this time, it is necessary to improve the detection accuracy of the insulation resistance. Another example is the fuel pipeline under the vehicle chassis. During normal operation, the pressure accuracy requirement for its sealing detection is ±0.1 MPa. However, when the vehicle is driving in special environments such as high temperature and high altitude, the detection pressure accuracy may need to be increased to ±0.05 MPa. Electrical circuits, fuel pipelines, and other targets whose detection parameter accuracy allows dynamic adjustment are classified as the second type of target area to be detected. In this area, the vehicle chassis detection robot needs to have the ability to dynamically adjust the detection parameters according to real-time situations. For example, when it detects that the temperature of the electrical circuit rises, it automatically improves the detection accuracy of the insulation resistance; when the vehicle is in a special operating environment, it timely adjusts the pressure accuracy of the fuel pipeline sealing detection according to the information fed back by the sensors. Through this dynamic adjustment mechanism, potential problems of these targets in different states can be detected more accurately, ensuring the safe operation of the vehicle.

[0086] As a further embodiment, the specific way for the path planning module to select the movement path is as follows:

[0087] When the detection path needs to extend to an adjacent area:

[0088] If the next area contains the first type of target to be detected:

[0089] Obtain the detection parameters of the target at the end of the current path;

[0090] Calculate the adjustment range of the detection parameters of the target in the adjacent area;

[0091] Select the target with the smallest adjustment range of the detection parameters as the next path node;

[0092] It should be understood that during the vehicle chassis detection process, the decision-making of the path planning module is crucial for the robot to complete the detection task efficiently and accurately. When the detection path needs to extend to an adjacent area, different path selection strategies will be adopted according to the different types of targets in the adjacent area. When the vehicle chassis detection robot moves along the planned path and reaches the end of the current path, it needs to record the detection parameters of the target being detected at this time (assumed to be target A). For example, if target A is a shock absorber in the vehicle chassis suspension system, its detection parameters may include a camera resolution of 1080P for detecting surface cracks, a detection frequency of once every 3 seconds, and a detection light intensity of 500 Lux, etc. These parameters reflect the current detection state and the detection requirements for this target;

[0093] There are multiple first - type targets to be detected in the adjacent area. For each target, the magnitude of the adjustment of the detection parameters from the detection parameters of the target at the end of the current path to the detection parameters required for this target needs to be calculated. Suppose there are target B and target C in the adjacent area. Target B is also a shock absorber. However, due to the different importance and potential risks of its location, a camera with a higher resolution may be required for detection. For example, the resolution needs to be increased to 4K, the detection frequency becomes once every 2 seconds, and the light intensity is adjusted to 600 Lux. Then the magnitude of the adjustment of the detection parameters from target A to target B is calculated as follows: Magnitude of camera resolution adjustment = |(4K - 1080P) / 1080P|×100% (the resolution needs to be unified in units before calculation), Magnitude of detection frequency adjustment = |(3 - 2) / 3|×100%, Magnitude of light intensity adjustment = |(600 - 500) / 500|×100%. Combining the adjustment magnitudes of these parameters gives an overall adjustment magnitude value. Similarly, calculate the magnitude of the adjustment of the detection parameters from target A to target C;

[0094] By comparing the adjustment magnitudes of the detection parameters of each target in the adjacent area, select the target with the smallest adjustment magnitude as the next path node. The reason for this is that a small adjustment magnitude of the detection parameters means that when the robot switches the detection target, there is no need to make large - scale parameter changes to the detection equipment, reducing the equipment adjustment time and the error probability, and enabling the detection task to be completed more efficiently. For example, after calculation, it is found that the adjustment magnitude of the detection parameters of target C is the smallest. Then the robot will move to the position where target C is located and use target C as the next path node to continue the detection operation;

[0095] If the next area contains second - type targets to be detected:

[0096] Calculate the distances between the end position of the current path and each target in the adjacent area;

[0097] Select the target with the smallest distance as the next path node.

[0098] It should be understood that in the under - vehicle space, when the robot reaches the end of the current path, devices such as lidar and ultrasonic sensors are used to measure the distances between the current position and each second - type target to be detected in the adjacent area. For example, the end of the current path is near the under - vehicle electrical circuit, and the adjacent area contains multiple targets such as fuel pipe joints. The distance from the current position to fuel pipe joint 1 is measured to be 30 cm, the distance to fuel pipe joint 2 is 45 cm, the distance to electrical connection point 3 is 20 cm, etc.;

[0099] From the calculated distances, select the target with the smallest distance from the end position of the current path as the next path node. This is because selecting the target with the smallest distance can reduce the unnecessary movement distance of the robot during movement, reduce energy consumption, and at the same time enable the robot to reach the next detection target faster, improving the detection efficiency. For example, in the above example, the distance between the current position and the electrical connection point 3 is the smallest, so the robot will move to the electrical connection point 3 and use it as the next path node to carry out the detection work on this target.

[0100] As a further embodiment, the path planning module is further configured to re-evaluate the path according to the precise parameters required by the current target, specifically:

[0101] Obtain the initial path, which is the current position of the inspection robot , and use this as the starting point and the target position as the end point to plan a shortest straight-line path , and its length ;

[0102] According to the precise parameters required by the current target provided by the data analysis module, through the formula , obtain the time required for parameter adjustment , where represents the current detection angle of the robot, represents the optimal detection angle required to complete the high-precision detection of the current target, refers to the adjustment speed of the robot's detection angle, that is, the number of degrees by which the robot can change the detection angle per unit time;

[0103] Calculate the movement time of the initial path;

[0104] Compare the times to obtain the results of whether to re-plan the route or not;

[0105] For the need to re-plan the route, introduce a new heuristic function , on the basis of the original heuristic function that only considers distance, add a detour penalty term to obtain , where is used to measure the degree of path detour, and is used to measure the degree of detour of the path from the starting point to the node this path, is the weight coefficient, used to control the penalty degree for detour, represents a node in the path search process, where i is the index of this node, is the newly designed heuristic function, used to evaluate from the node to the target position The path cost estimate value, is the original heuristic function that only considers distance, is the node to the target position The distance; it should be noted that Through the formula , where represents the total actual path length from the starting point passing through , , , , to the current node That is, the sum of the distances between adjacent nodes on the path, This ratio reflects the length multiple of the current path relative to the initial straight-line shortest path. When this ratio is 1, it means that the current path has no detours and is the straight-line shortest path; when the ratio is greater than 1, the larger the part greater than 1, the higher the degree of detour of the path, The degree of detour of the path can be quantified and added to the heuristic function as a detour penalty term to punish the detoured path. Through the combined action of these parameters on the heuristic function, the improved A* algorithm can minimize the detour of the path while meeting the parameter adjustment requirements and find a path with better comprehensive performance.

[0106] It should be further noted that in the current formula, i represents the index of the node currently being evaluated in the path node sequence. The path planning process starts from the starting point and gradually searches and expands nodes. When a specific node is searched, the degree of detour of the path from the starting point to the current node is calculated through this expression. As the search algorithm traverses different nodes, the value of i will change accordingly, so as to dynamically evaluate the detour situation of the path at different nodes;

[0107] Example: When i = 3, it means that the degree of detour of the path from passing through , to is being evaluated.

[0108] In the current formula, j represents the loop variable used for summation calculation. When calculating the total path length from the starting point to the current node , j traverses the node indices from 0 to i - 1 to accumulate the distance between each adjacent node .

[0109] It should be further noted that The value determination method of

[0110] 1. Set the weight coefficient by virtue of in-depth understanding of the robot's operating environment and task requirements and past experience. If there is rich practical experience in the task and path planning of this type of robot, and the relative importance of path detours and parameter adjustments in specific scenarios is understood, then the value of

[0111] 2. Through multiple experiments in a simulated environment or an actual scenario, try different values, observe the path planning results, and determine the optimal value according to the preset evaluation indicators, etc.

[0112] When re-planning the path, design a new heuristic function H, and add a detour penalty term on the basis of considering the distance factor. This makes the path planning not only focus on the direct distance from the current point to the target point, but also take into account the degree of path detour. Through the weight coefficient can flexibly control the penalty strength for detours, guide the search algorithm to find a path that can meet the time requirements for parameter adjustment and is not overly detoured, balancing the path efficiency and parameter adjustment requirements;

[0113] By reasonably planning the path, reducing unnecessary detours, shortening the total path length for the robot to reach the target, thus saving running time and energy consumption. At the same time, it avoids frequent parameter adjustments or insufficient parameter adjustments caused by unreasonable paths, and further improves the overall efficiency of the robot to complete tasks.

[0114] Taking the current position of the robot as the starting point, using the improved A* algorithm, based on the new heuristic function to find intermediate points , , , , ; It should be noted that when searching, consider the cost of expanding from the current point to adjacent points. This cost includes the actual moving distance and the cost estimated according to the heuristic function to the target point (including the detour penalty term). Through continuous expansion and evaluation of nodes, an adjusted path is formed.

[0115] Calculate the time to move to the target along the adjusted path to see if it can meet the time required for parameter adjustment;

[0116] If it meets the requirement, this path is the path that meets the parameter adjustment requirements; if it does not meet the requirement, adjust the weight coefficient in the heuristic function and re-search for the path; It should be noted that the adjusted path Length , is a series of nodes on the re - planned path. i is the index of these nodes. When i = 0, the corresponding is the starting position of the robot. As i increases from 0 to n - 1, successively represents different intermediate position points on the path. represents the next adjacent node of . By calculating the distance and between each pair of adjacent nodes , the length of each small segment on the path can be obtained; represents the distance between node and its next adjacent node . Usually, this distance can be calculated using the corresponding distance formula according to the coordinate system of the space where the robot is located.

[0117] Furthermore, calculate the time to move along the adjusted path to the target to see if it can meet the time required for parameter adjustment. Specifically, calculate the time to move along the adjusted path to the target, and judge again whether it holds.

[0118] By comparing the length of the adjusted path with the initial path, evaluate the degree of detour of the path. It should be noted that if R is too large, it indicates excessive detour. Adjust the heuristic function or reconsider the parameter adjustment method to reduce the path detour and improve the detection efficiency while meeting the parameter adjustment requirements.

[0119] It should be understood that traditional path planning usually focuses on basic goals such as the shortest distance and obstacle avoidance. The present invention constructs a new path planning logic by introducing the factor of time and space required to meet parameter adjustment. For example, by analyzing the time required for parameter adjustment such as the detection angle and comparing it with the time to move along the path, to decide whether to re - plan the path. This way of associating parameter adjustment with the path movement time helps to plan the path more reasonably;

[0120] Through the path planning module, it can reasonably re - evaluate the path according to the precise parameters required by the target, ensuring that the robot has enough time and space to adjust the parameters to the optimal state when approaching the target, meeting the high - precision detection requirements.

[0121] As a further embodiment, the specific steps to calculate the initial path movement time are as follows:

[0122] Obtain the provided speed v of the robot;

[0123] Calculate the time to move to the target along the initial path through the formula .

[0124] It should be understood that the time to move to the target along the initial path is obtained through the above formula, which is convenient for subsequent comparative calculations and is beneficial for path planning.

[0125] As a further embodiment, the specific results of comparing the time to obtain the results of needing to re-plan the route and not needing to re-plan the route are as follows:

[0126] Compare with , and judge whether it holds;

[0127] If it holds, it means that the initial path cannot meet the time required for parameter adjustment and the path needs to be re-planned. Otherwise, the initial path meets the requirements and does not need to be re-planned.

[0128] It should be understood that through the judgment formula, it can be obtained whether further route planning is needed and the optimal route can be planned.

[0129] As a further embodiment, the specific method for evaluating the degree of detour of the path is as follows:

[0130] Calculate the detour factor through the formula to obtain, where the detour factor is an index used to quantify the degree of detour of the path. By comparing the ratio relationship between the length of the adjusted path and the length of the initial straight shortest path, it reflects the detour situation of the path relative to the shortest path. refers to the length of the adjusted path, which is the total length of the path obtained after re-planning considering that the robot needs to meet the parameter adjustment requirements. As a reference length, it is used to compare with the length of the adjusted path to evaluate whether the path detours due to meeting the parameter adjustment requirements.

[0131] It should be noted that the closer the value is to 1, the closer the path is to the straight shortest path and the lower the degree of detour; the larger the value, the longer the path is relative to the initial shortest path and the higher the degree of detour.

[0132] It should be understood that in the above description, by analyzing the parameter adjustment requirements, such as the adjustment of the detection angle, the time required for parameter adjustment is accurately calculated . Compare it with the time to move along the initial path ​​Compare and determine whether the initial path meets the parameter adjustment requirements. If not, re-plan the path, which ensures that the robot has enough time to adjust the parameters to the optimal state before reaching the target, thus meeting the strict requirements of high-precision detection for parameters and improving the accuracy and reliability of the detection results;

[0133] In actual detection tasks, different targets may require different detection parameters, and high precision is required for the parameters. This method can dynamically plan the path according to the specific parameter requirements of each target, enabling the robot to flexibly handle various complex detection task scenarios and enhancing the adaptability of the robot in diverse working environments.

[0134] As a further embodiment, the path planning module further includes:

[0135] Execute a serialized detection operation along the planned movement path. When an abnormality is detected in the current target component to be detected, trigger an association analysis instruction;

[0136] According to the preset fault propagation model and component dependency graph, determine that all components functionally associated with the current target component to be detected with problems need to be synchronously repaired. At the same time, based on the fault history data, real-time working condition parameters, and system operation status of all components functionally associated with the current target component to be detected with problems, generate a detection exemption decision for all components functionally associated with the current target component to be detected with problems;

[0137] Respond to the detection exemption decision and generate an updated path through an online re-planning algorithm.

[0138] Specifically, the underbody inspection robot performs serialized inspections on each target component of the underbody according to the pre-planned movement path. During the inspection process, the robot uses various sensors and inspection equipment to collect data of the target components in real time. For example, when inspecting the shock absorber of the underbody suspension system, the working state is judged by measuring parameters such as the stroke and damping force of the shock absorber. Once an abnormality is detected in the current target component to be detected, such as the damping force of the shock absorber exceeding the normal range, the robot will immediately trigger an association analysis instruction. This triggering mechanism is like an "alarm switch", which is turned on when the detection data deviates from the preset normal range, triggering a series of subsequent operations;

[0139] When the correlation analysis instruction is triggered, the system will analyze based on the preset fault propagation model and the component dependency graph. The fault propagation model is constructed based on a large amount of vehicle fault data and principles of mechanics, electricity, etc., and it describes the propagation law between faults of different components. The component dependency graph intuitively shows the functional association relationships between components under the vehicle bottom. For example, the shock absorber, spring, lower control arm and other components together form the suspension system, and they are functionally interdependent. Taking the abnormal shock absorber as an example, through the fault propagation model, it can be judged that the shock absorber fault may cause uneven force on the spring, which in turn affects the working state of the lower control arm. According to the component dependency graph, all components that have a functional association with the shock absorber can be identified, such as the spring, lower control arm, steering knuckle, etc. The system determines that these components need to be repaired synchronously because the fault of one component often affects other associated components. Timely repair of the associated components can avoid the occurrence of potential faults and ensure the overall performance of the vehicle;

[0140] After determining the associated components that need to be repaired synchronously, the system will generate a detection exemption decision based on the fault history data, real-time operating condition parameters and system operating status of these components. The fault history data records information such as the type, time, frequency of past faults of each associated component. By analyzing these data, the fault tendency of the component can be understood. The real-time operating condition parameters reflect the current operating state of the vehicle, such as vehicle speed, engine speed, vehicle load, etc. These parameters will affect the working conditions of the components. The system operating status includes the working status of the detection system itself, such as the accuracy of sensors, the stability of data transmission, etc. For example, for the spring associated with the shock absorber, if its fault history data shows that it rarely fails, the current real-time operating condition parameters indicate that the vehicle is driving smoothly, the spring is under normal force, and the detection system is in good operating condition. After comprehensive evaluation, the system may generate a detection exemption decision for this spring, that is, it is considered that a detailed detection can be temporarily not carried out under the current circumstances to improve the detection efficiency. However, if the fault history data of a certain associated component shows frequent failures and the current real-time operating condition parameters indicate that its working conditions are relatively harsh, then the system will not give a detection exemption;

[0141] After the detection exemption decision is generated, the path planning module responds to this decision and generates an updated path through an online replanning algorithm. The online replanning algorithm takes into account factors such as the positions of the components with detection exemptions, the distribution of the remaining components to be detected, and the kinematic parameters of the robot. Suppose a certain associated component (such as a spring) is given a detection exemption, and the originally planned path contains a detection point for this spring. Then the online replanning algorithm will automatically avoid this detection point and recalculate a more efficient path so that the robot can continue to detect other components that have not been given detection exemptions. For example, the algorithm may plan a path directly from the current position to the next unexempt and closest component to be detected, while ensuring that the path meets the motion limitations of the robot, such as turning radius, moving speed, etc. By generating an updated path, the underbody detection robot can quickly adjust the detection strategy after discovering component abnormalities and making a detection exemption decision, and continue to efficiently complete the underbody detection task, improving the flexibility and adaptability of the entire detection process.

[0142] As a further embodiment, the updated path is generated through an online replanning algorithm, specifically:

[0143] Remove all component positions that are functionally associated with the current problematic component to be detected from the initial path node sequence;

[0144] Recalculate the optimal traversal order of the remaining nodes.

[0145] It should be understood that during the detection process of the underbody detection robot, once all the components that are functionally associated with the current problematic component to be detected are determined, the first step of path replanning is to remove the positions of these associated components from the initial path node sequence. For example, the initial path node sequence is planned in sequence according to the layout of the underbody components. Suppose an abnormality is detected in the shock absorber of the underbody suspension system currently. According to the fault propagation model and the component dependency graph, the components functionally associated with it are determined to be the spring, lower control arm, steering knuckle, etc. These components each correspond to their respective detection nodes in the initial path. When performing path replanning, the system will delete the position information of the associated components such as the spring, lower control arm, and steering knuckle in the initial path node sequence. This operation is like erasing the marks of specific locations on the map, so that subsequent path planning no longer involves the positions of these components that have been determined to require synchronous maintenance and may be given detection exemptions, avoiding the robot from doing useless work and improving the detection efficiency;

[0146] After removing the positions of the associated components, the remaining nodes form a new set of components to be detected. At this time, the online replanning algorithm needs to recalculate the optimal traversal order of these remaining nodes, replan the route based on the above planning method, and finally determine the optimal traversal order of the remaining nodes, so as to plan a new and efficient detection path for the underbody inspection robot, enabling it to continue to complete the underbody inspection task while ensuring the comprehensiveness and accuracy of the inspection.

[0147] The above is only the preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions within the idea of the present invention belong to the protection scope of the present invention. It should be noted that for those of ordinary skill in the art, several improvements and refinements made without departing from the principle of the present invention should also be regarded as within the protection scope of this template.

Claims

1. An intelligent control system for an underbody inspection robot, characterized in that: include: A data acquisition module is configured to collect all-round data information of the target vehicle bottom, wherein the data information at least includes position information of the target under the vehicle bottom, spacing information between different targets under the vehicle bottom, and contour information of the target under the vehicle bottom; A data analysis module, connected to the data acquisition module, configured to perform: performing in-depth analysis on the acquired data to identify the type, location, shape and size characteristics of each target under the vehicle; Use data analysis algorithms to statistically analyze target distribution patterns and determine the detection parameters corresponding to each target; A path planning module, connected to the data analysis module, configured to generate a moving path based on target characteristics and distribution law information and in combination with the kinematic parameters of the robot; The specific steps of generating a moving path based on target features and distribution law information and combining the kinematic parameters of the robot are as follows: A1: Divide the space under the vehicle into multiple detection areas, each area corresponding to the first type of target to be detected or the second type of target to be detected; A2: Execute the following operations on the area where the first type of target to be detected is located: Obtain the detection parameters corresponding to each target in the area; Generate a moving path in the order of monotonically increasing or monotonically decreasing adjustment amplitudes of the detection parameters; A3: Execute the following operations on the area where the second type of target to be detected is located: Calculate the distance between each target in the area; Generate a moving path based on the principle of minimizing the difference in moving distances between adjacent targets; The detection area division method specifically includes: Determine the detection parameter accuracy requirements for each target to be detected; Classifying multiple targets whose detection parameter accuracy requirements are within a preset fixed range as a first type of target area to be detected; A plurality of targets whose detection parameter accuracy allows for dynamic adjustment are classified as the second type of target area to be detected.

2. The intelligent control system of the vehicle underbody inspection robot according to claim 1, characterized in that: The specific method of selecting the moving path by the path planning module is: When the detection path needs to be extended to adjacent areas: If the next area contains the first type of target to be detected: Get the detection parameters of the target at the end of the current path; Calculate the adjustment range of the target detection parameters in adjacent areas; Select the target with the smallest adjustment amplitude of detection parameters as the next path node; If the next area contains the second type of target to be detected: Calculate the distance between the end position of the current path and each target in the adjacent area; Select the target with the smallest distance as the next path node.

3. The intelligent control system of the vehicle underbody inspection robot according to claim 1, characterized in that: The path planning module is also used to re-evaluate the path according to the precise parameters required by the current goal, specifically: Get the initial path, which is the current position of the detection robot , and use this as the starting point and the target location As the end point, plan a straight line shortest path , its length ; According to the precise parameters required for the current goal provided by the data analysis module, the formula , get the time required for parameter adjustment ,in, Indicates the current detection angle of the robot. Represents the optimal detection angle required to complete high-precision detection of the current target. Refers to the adjustment speed of the robot's detection angle, that is, the degree to which the robot can change the detection angle per unit time; Calculate the initial path moving time; Compare the times to determine whether rerouting is necessary or not; Introducing new heuristic functions for re-planning routes , in the original heuristic function that only considers distance On the basis of, add the detour penalty term, get ,in, It is used to measure the degree of circuitousness of the path, and to measure the To Node The circuitousness of this path, is the weight coefficient, which is used to control the degree of penalty for detour. Represents a node in the path search process, where i is the index of the node, A newly designed heuristic function for evaluating slave nodes To the target location The estimated path cost is is the original heuristic function that only considers distance, Is a node To the target location distance; The robot's current position As a starting point, using the improved A* algorithm, according to the new heuristic function Finding the middle point , , , , ; Calculate the time to move to the target along the adjusted path to see if it can meet the time required for parameter adjustment; If satisfied, this path is the path that meets the parameter adjustment requirements; if not satisfied, adjust the weight coefficient in the heuristic function and search the path again; The circuitousness of the path was assessed by comparing the length of the adjusted path with the length of the initial path.

4. The intelligent control system of the vehicle underbody inspection robot according to claim 3, characterized in that: The specific steps of calculating the initial path moving time are: Get the robot's provided speed v; By formula , calculate along the initial path Time to move to the target.

5. The intelligent control system of the vehicle underbody inspection robot according to claim 4, characterized in that: The comparison of the time to obtain the result of whether the route needs to be re-planned or not is specifically: contrast and ,judge whether it is established; If true, it means that the initial path cannot meet the time required for parameter adjustment and the path needs to be replanned. Otherwise, the initial path meets the requirements and no replanning is required.

6. The intelligent control system of the vehicle underbody inspection robot according to claim 5, characterized in that: The specific method of evaluating the circuitousness of the path is: Calculate the detour factor by the formula It is obtained that, among which, the circuitous factor It is an indicator used to quantify the degree of path tortuosity. By comparing the proportional relationship between the adjusted path length and the initial straight shortest path length, it reflects the tortuosity of the path relative to the shortest path. Refers to the adjusted path length, which is the total length of the path obtained after replanning, taking into account the robot's need to meet parameter adjustment requirements. As a benchmark length, it is used to compare with the adjusted path length to evaluate whether the path is detour due to meeting parameter adjustment requirements.

7. The intelligent control system of the vehicle underbody inspection robot according to claim 6, characterized in that: The path planning module also includes: Execute serialized inspection operations along the planned moving path, and trigger the correlation analysis instruction when an abnormality is detected in the current target component to be inspected; According to the preset fault propagation model and component dependency graph, it is determined that all components that are functionally associated with the target component to be detected that currently has problems need to be repaired simultaneously. At the same time, based on the fault history data, real-time operating parameters and system operating status of all components that are functionally associated with the target component to be detected that currently has problems, a decision on exemption from inspection of all components that are functionally associated with the target component to be detected that currently has problems is generated; In response to the detection exemption decision, an updated path is generated by an online replanning algorithm.

8. The intelligent control system of the vehicle underbody inspection robot according to claim 7, characterized in that: The update path is generated by the online replanning algorithm, specifically: Remove all component positions that are functionally associated with the target component to be inspected that has a problem, from the initial path node sequence; Recalculate the optimal traversal order of the remaining nodes.

Images