A four-wheel encoder-based wall-climbing vehicle ultrasonic detection full-coverage scanning method

Abstract

The application discloses a wall-climbing trolley ultrasonic detection full-coverage scanning method based on four-wheel encoders, which establishes a space mapping relationship between the center of a trolley body and the center of an ultrasonic detection module, takes the center of the ultrasonic detection module as a tracking target of path planning and motion control, uses four-wheel independent encoders to solve the trolley body pose in real time, adjusts the rotating speed of the four wheels through PID closed-loop control, and makes the center of the ultrasonic detection module accurately track the target track; meanwhile, the method is based on the standard deviation of the four-wheel speed to determine and compensate the skidding. The application solves the edge missing detection problem caused by the misplacement of the control target of the existing wall-climbing robot, does not need expensive sensors such as visual cameras, and is especially suitable for ultrasonic nondestructive detection operation in a narrow space.

Description

Technical Field

[0001] This invention relates to the field of ultrasonic testing technology, and in particular to a method for full-coverage ultrasonic scanning of a wall-climbing vehicle based on a four-wheel encoder. Background Technology

[0002] In non-destructive testing of corrosion in industrial components such as storage tanks, pipes, and large metal plates, wall-climbing robots equipped with ultrasonic probes perform row-by-row full-coverage scanning of the metal wall, which is the mainstream automated inspection method. However, current ultrasonic scanning control technology for wall-climbing robots still has significant shortcomings, severely impacting the completeness of the inspection and positioning accuracy.

[0003] The primary problem with existing technologies is the misalignment of the target being scanned and tracked. Conventional robot path planning and motion control are based on the vehicle's center or outline, without considering the spatial offset between the vehicle and the front-mounted, cantilevered ultrasonic testing module. When the robot turns or travels along curves, the actual scanning trajectory of the ultrasonic module deviates from the preset grid, easily causing missed detections in component edge areas, thus failing to meet the requirement of full coverage in non-destructive testing.

[0004] Secondly, the four-wheel alignment accuracy is insufficient, and the accumulation of errors is serious. Traditional equipment often uses single-axis or dual-axis encoders to estimate mileage. Industrial walls generally have uneven adhesion and rough, corroded surfaces, which can easily cause one-sided wheel slippage. The system only calculates the body displacement by averaging wheel speed, which will produce a large positioning deviation, causing the ultrasonic detection module to continuously deviate from the preset scanning path.

[0005] Existing related technologies are mainly divided into two categories: environment-aware full-area coverage planning and pure path planning. Both take the robot itself as the core control target, ensuring only that the vehicle reaches the designated coordinates, without focusing on the coverage accuracy of the working tool. Most solutions rely on sensors such as LiDAR, vision cameras, IMU, and GPS, which incur high computing power and cost. Furthermore, sensors such as GPS are not suitable for indoor wall-mounted work scenarios. A few solutions that use the tool's end effector as the tracking target rely on the kinematics of the robotic arm to eliminate bias, but these are only suitable for flexible robotic arm structures and cannot be adapted to rigidly fixed ultrasonic detection modules. At the same time, existing technologies are mostly applied to scenarios such as floor cleaning and disinfection, and are not optimized for special working conditions such as wall adhesion and wheel slippage.

[0006] In summary, there is currently a lack of dedicated scanning control schemes adapted to ultrasonic nondestructive testing of walls, making it difficult to solve the problems of missed detections and positioning errors. There is an urgent need to propose new control methods to meet the requirements of high-precision full-coverage testing. Summary of the Invention

[0007] The purpose of this invention is to provide a method for ultrasonic inspection of a wall-climbing vehicle based on a four-wheel encoder, which achieves full coverage scanning. By calculating the spatial geometric relationship between the vehicle center and the inspection module, and using the odometer information from the four independent encoders, the vehicle is driven to perform compound motion (translation and steering), ensuring that the ultrasonic inspection module achieves full coverage scanning of the target area.

[0008] To achieve the above objectives, this invention provides a method for full-coverage ultrasonic testing of a wall-climbing vehicle based on a four-wheel encoder. The four-wheel wall-climbing vehicle is fixedly equipped with an ultrasonic testing module, each wheel is equipped with an independent encoder, and a main control unit is built into the vehicle body. The center of the ultrasonic testing module is offset from the center of the vehicle body. The scanning method specifically includes the following steps:

[0009] S1. Multi-coordinate system establishment and parameter calibration: Construct a matching world coordinate system, vehicle coordinate system and ultrasonic module coordinate system, calibrate and store the two-dimensional center offset vector of the ultrasonic module center relative to the vehicle center, and calibrate the vehicle structure parameters, wheel running parameters, ultrasonic effective detection width and preset detection overlap rate to provide unified reference parameters for subsequent steps.

[0010] S2. Real-time differential pose calculation and offset correction for four wheels: In each control cycle, the pulse increments of the four-wheel encoders are collected. Based on the pre-calibrated wheel running parameters and vehicle structure parameters, the real-time linear velocity of each wheel is calculated. Then, the linear velocity and angular velocity of the vehicle body are calculated, and the pose of the vehicle center is iteratively updated.

[0011] Based on the rigid body transformation relationship of three sets of coordinate systems, combined with the vehicle body attitude angle and two-dimensional offset vector, the real-time detection pose of the ultrasonic module center in the world coordinate system is obtained, and the ultrasonic module center is used as the reference point for the whole trajectory tracking.

[0012] S3, Four-way partition blind spot compensation scanning: Based on the calibrated ultrasonic parameters, a basic scanning trajectory is generated. For the blind spot caused by the eccentric installation of the detection module and the obstruction of the vehicle structure, a primary and secondary partition complementary scanning mechanism is adopted. The blind spot is closed by four sets of orthogonal angle partitions working independently and the trajectories are spliced ​​together. With the help of multi-posture collaborative motion, the eccentric detection dead zone is filled, and the wall surface is fully covered by scanning.

[0013] S4. Multi-level slippage recognition and trajectory compensation: Real-time monitoring of the motion status of the four wheels, identification of wheel slippage faults, and correction of trajectory deviations through a multi-level method of wheel torque adjustment to eliminate slippage and historical normal posture retraction correction, and then continue to complete the scanning operation.

[0014] S5. Dual-loop PID closed-loop precision control: Employing a dual-loop PID control architecture with an outer position loop and an inner speed loop, the outer position loop outputs the vehicle's target motion parameters based on the deviation between the actual position of the ultrasonic detection module's center and the target position.

[0015] The inner speed loop adjusts the speed of each wheel drive motor based on the deviation between the actual linear speed of the four wheels and the target linear speed, thereby achieving closed-loop precise tracking of the scanning trajectory.

[0016] Preferably, step S2 employs a split-type velocity calculation and planar coordinate matching method to achieve pose correction, specifically including:

[0017] The system collects independent pulse increments from all four wheels, combines these with vehicle structure parameters and wheel operating parameters to calculate the linear velocity of each wheel, and then uses the average velocity of the left and right wheels to calculate the vehicle's motion parameters.

[0018] Based on the rigid body translation transformation relationship of three sets of coordinate systems, the center pose transformation of the ultrasonic module is completed by combining the two-dimensional plane offset and the vehicle body attitude angle.

[0019] Preferably, the trajectory generation task in step S3 is configured with dedicated parameter arrangement logic, specifically including:

[0020] The pre-calibrated effective ultrasound detection width and preset detection overlap rate parameters are retrieved. Based on the ultrasound detection coverage parameters, multiple parallel scanning trajectories are planned, and adjacent scanning trajectories are arranged according to the preset overlap amount.

[0021] Preferably, in step S3, the primary and secondary partition complementary scanning mechanism operates sequentially for four sets of orthogonal angle partitions, performing multi-pose cooperative motion according to a fixed time sequence of 90° forward primary scanning, 0° right-side supplementary scanning, 270° reverse scanning, and 180° left-side supplementary scanning. The trajectories of each partition are generated independently and spliced ​​together.

[0022] In the initial state of operation, the wall-climbing trolley is positioned in the lower left corner of the detection area. Based on the dimensions of the trolley body, a fixed offset vector between the center of the trolley body and the center of the ultrasonic module is calibrated, so that the center of the ultrasonic module is positioned on the first scanning trajectory line segment.

[0023] The 90° forward main scan adopts a zigzag cyclic motion logic, which sequentially completes the straight-line scanning, backward turning with the center of the ultrasound module as the rotation reference, linear displacement compensation after rotation, straight backward, forward turning and secondary linear compensation actions, thus completing the main area trajectory traversal.

[0024] After the 90° forward main scan is completed, the vehicle attitude and working position are switched by rotating around the center of the ultrasonic module and axially translating. The vehicle then enters the 0° right-side supplementary scan condition and completes the lateral blind zone trajectory coverage by using a straight constant speed scan and a fixed-point arc retraction and line change method.

[0025] Then, the vehicle's attitude and position were adjusted, and a 270° reverse scan was performed using a zigzag trajectory. The forward and reverse scans were coordinated to cover the same set of trajectory segments.

[0026] Finally, switch the initial posture and position of the operation, and perform a 180° left-side supplementary scan according to the 90° scan motion logic.

[0027] Preferably, in step S4, the slippage fault identification adopts a multi-parameter joint judgment mechanism, which simultaneously collects wheel motion data and vehicle posture data, compares real-time operating parameters with preset normal operating parameter ranges, locates the slippage abnormal wheel based on parameter deviation state, and outputs slippage judgment signal.

[0028] Preferably, based on the multi-parameter joint judgment mechanism, the specific judgment logic for wheel slippage is as follows: calculate the standard deviation of the four wheel speeds in real time, combine the adaptive threshold positively correlated with the vehicle speed and the preset duration to complete the slippage judgment, and judge the wheel with the largest speed deviation as the abnormal slipping wheel.

[0029] Preferably, in step S4, the wheel torque adjustment anti-slip method matches the vehicle's motion state and outputs the corresponding wheel set torque. The lateral slip generated during wheel operation is offset by dynamic wheel difference torque distribution, thereby achieving active suppression of lateral slip force.

[0030] Preferably, in step S4, the specific compensation method for the historical pose rollback is as follows: pre-caching the historical pose data of the vehicle body and the ultrasonic detection module under normal operating conditions, and after determining the wheel slippage fault, resetting the current pose of the vehicle body and the ultrasonic detection module to the last normal pose before slippage, and continuing to execute the preset scanning trajectory based on the reset pose.

[0031] Preferably, in step S5, the dual-loop PID control adopts an independent control architecture with inner and outer loops nested in a layered manner. The outer loop of position is used to achieve trajectory tracking control, and the inner loop of speed is used to achieve motor power control. The inner and outer loops work together to complete the closed-loop dynamic tracking of the scanning trajectory.

[0032] Preferably, the outer position loop calculates and outputs the target motion parameters of the vehicle body based on the position deviation of the center of the ultrasonic module, while the inner speed loop independently adjusts the speed of the corresponding drive motor based on the speed deviation of each wheel. The position loop and speed loop are configured with independent sampling periods and are adapted to the dynamic working conditions of wall scanning to complete the dual-loop PID parameter calibration.

[0033] The beneficial effects of this invention are:

[0034] 1. This invention constructs multiple matching coordinate systems to accurately calibrate the eccentric offset vector between the ultrasonic module and the vehicle body, achieving precise conversion from the vehicle body pose to the ultrasonic detection center pose. It abandons the traditional tracking method based on the vehicle body center and uses the ultrasonic detection center as the entire trajectory tracking reference, effectively improving the trajectory deviation and missed detection problems caused by eccentric installation and misalignment of control targets. At the same time, this solution relies on encoder differential calculation combined with eccentricity correction algorithm to complete high-precision trajectory tracking. It does not require high-cost auxiliary sensors such as vision cameras, resulting in low equipment integration cost, compact structure, and no need to reserve space for visual recognition installation. It can significantly improve the accuracy of detection points and is adaptable to various ultrasonic non-destructive testing operation scenarios in confined spaces, making it more adaptable to different working conditions.

[0035] 2. This invention adopts a four-directional partitioned primary and secondary complementary scanning mechanism. Through the time-sequential collaborative scanning of four sets of orthogonal angle partitions, it specifically fills the blind spots caused by the eccentric installation of the ultrasonic module on the vehicle body. The multi-posture trajectories are spliced ​​together to close the detection dead angles, realizing full coverage scanning of the entire wall surface without blind spots, and solving the technical problem of missed detection in traditional single-direction scanning.

[0036] 3. This invention adopts a multi-parameter joint slippage determination mechanism, which combines the standard deviation of four-wheel speed with an adaptive threshold to achieve accurate slippage identification. Compared with the single-parameter determination method, it has stronger anti-interference ability and higher determination accuracy. At the same time, it is equipped with a multi-level trajectory compensation strategy of dynamic torque deslippage and historical posture reversal, which can quickly repair the trajectory deviation caused by slippage and ensure the continuity and stability of the scanning operation.

[0037] 4. This invention adopts a layered nested dual-loop PID control architecture. The outer position loop focuses on precise tracking of the trajectory position, while the inner speed loop independently regulates the speed of each wheel. The dual loops provide differentiated and coordinated control, which is suitable for the multi-wheel independent drive characteristics of the wall-climbing vehicle and the dynamic working conditions of the wall surface. This effectively improves the trajectory tracking accuracy and vehicle motion stability, and avoids the defects of poor dynamic response and low fault tolerance of single closed-loop control. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.

[0039] Figure 1 This is the system architecture diagram of the present invention.

[0040] Figure 2 This is a flowchart of the detection process of the present invention.

[0041] Figure 3 This is a schematic diagram of the four-way partition scanning detection points of the present invention.

[0042] Figure 4a This is a schematic diagram of the path of parallel line segments in the width direction of the present invention.

[0043] Figure 4b This is a schematic diagram of the path of parallel line segments along the length direction of the present invention.

[0044] Figure 4c This is a schematic diagram of the initial coordinates of the vehicle body of the present invention.

[0045] Figure 5a This is a schematic diagram of the linear scanning stage in the 90° scanning of this invention.

[0046] Figure 5b This is a schematic diagram of the backward turning phase during the 90° scan of this invention.

[0047] Figure 5c This is a schematic diagram of the linear compensation stage in the 90° scan of the present invention.

[0048] Figure 5d This is a schematic diagram of the linear backward movement phase during the 90° scan of this invention.

[0049] Figure 5e This is a schematic diagram of the forward turning phase during the 90° scan of this invention.

[0050] Figure 5f This is a schematic diagram of compensation during the 90° scan of the present invention.

[0051] Figure 5g This is a schematic diagram of the 90° scan completion state of the present invention.

[0052] Figure 6a This is a schematic diagram of the 0° scanning start state of the present invention.

[0053] Figure 6b This is a schematic diagram of the 0° scan of the present invention.

[0054] Figure 7a This is a schematic diagram of the starting state of the 270° scan of the present invention.

[0055] Figure 7b This is a schematic diagram of the 270° scanning of the present invention.

[0056] Figure 8 This is a schematic diagram of the 180° scanning of the present invention.

[0057] In the diagram: vehicle body 1; wheels 2; ultrasonic module 3; main control unit 4. Detailed Implementation

[0058] The embodiments of the present invention are described in detail below. Examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, but should not be construed as limiting the present invention.

[0059] In existing ultrasonic automated inspection technologies for wall-climbing robots, the industry generally uses the vehicle's center coordinates as the reference for trajectory tracking and closed-loop control. However, this idealized assumption only applies to experimental conditions where the probe is centrally mounted and there are no structural obstructions. In actual industrial assembly, due to hardware layout constraints such as probe size, fixed brackets, negative pressure adsorption structures, and internal wiring, ultrasonic inspection modules typically adopt an eccentric mounting method, creating a fixed two-dimensional offset between the actual inspection point and the vehicle's control center. If the control logic based on the traditional vehicle center reference is used, when the vehicle moves, yaws, or switches attitudes during the scanning process, the eccentric point will experience a nonlinear spatial offset as the vehicle rotates. This error accumulates continuously with the inspection stroke, eventually leading to problems such as misalignment between the preset trajectory and the actual inspection point, missed detection of material edges, and inaccurate defect localization, directly reducing the accuracy and completeness of non-destructive testing.

[0060] Therefore, as Figures 1-2 As shown, the four-wheeled wall-climbing robot of the present invention is equipped with an ultrasonic detection module, each wheel is equipped with an independent encoder, the main control unit is built into the vehicle body, and the center of the ultrasonic module is offset from the center of the vehicle body. By constructing multiple sets of matching coordinate systems, the offset vector between the ultrasonic module and the vehicle body is accurately calibrated, so as to realize the accurate conversion of the vehicle body pose to the ultrasonic detection center pose.

[0061] Specifically, this invention first rigorously completes the establishment of multiple coordinate systems and the calibration of all-dimensional parameters, constructing three sets of mutually matching reference coordinate systems, unifying the calculation rules and spatial definitions throughout the entire process, and providing a standardized calculation benchmark for subsequent pose calculation, eccentricity correction, trajectory planning, and motion control. The specific coordinate system establishment method is strictly adapted to the equipment assembly structure and sheet metal inspection conditions:

[0062] Establish the world coordinate system X W O W Y W O is the lower left corner of the area of ​​the board to be tested. W X is the horizontal extension direction along the board. W The axis follows the rule of negative values ​​on the left and positive values ​​on the right; the vertical direction extending up and down along the plate is Y. W The axes follow the rule of negative values ​​at the bottom and positive values ​​at the top. The world coordinate system serves as the absolute reference coordinate system, and the poses of all vehicle bodies and detection modules are ultimately mapped to this coordinate system.

[0063] Establish the vehicle coordinate system X C O C Y C : Define X as the origin, with the geometric center C of the wall-climbing car as the origin, and along the positive direction of the car's movement. c The axis, perpendicular to the vehicle's forward direction and pointing to the left side of the vehicle, is defined as y. C The spatial pose of the vehicle body in the world coordinate system is uniformly denoted as (x). C,y C ,θ), where θ is the vehicle body X C Axis and world coordinate system X W The included angle of the axis is taken as the positive direction of counterclockwise rotation, and the angle value is constrained to the range of [0, 2π), which fully represents the vehicle's planar position and heading attitude.

[0064] Establish the coordinate system X of the ultrasonic testing module P O P Y P With the geometric center P of the ultrasonic probe array as the origin, Y is defined along the direction of the ultrasonic array element arrangement. P The axis is defined as X along the direction of equipment scanning travel. P Since the ultrasonic testing module is rigidly fixed to the front of the vehicle body by a bracket, there is no relative rotation or wobbling between it and the vehicle body. Therefore, the coordinate system of the ultrasonic module and the coordinate system of the vehicle body only have a fixed translation relationship. The spatial orientation of the two coordinate systems always remains completely consistent, with no angular deviation.

[0065] After the coordinate system was established, the key parameters of the entire machine were calibrated and fixed to the main control unit, including the ultrasonic module eccentricity parameters, detection process parameters, and vehicle wheel structure parameters. Among these, the ultrasonic module center P has a fixed two-dimensional offset vector (d) relative to the vehicle center C. x ,d y Due to limitations imposed by the equipment's structural layout, for example, under normal assembly conditions, it may not meet the requirements of d. y =0、d x >0, positioning the ultrasonic testing center directly in front of the vehicle body, forming a front-mounted eccentric mounting structure. The real-time spatial coordinates of the ultrasonic module center in the world coordinate system are uniformly denoted as (x... P y P Simultaneously calibrate the effective detection width W of the ultrasound module. probe User-preset detection coverage percentage P overlap This is used for subsequent calculations of trajectory overlap and step spacing. The main calibration parameters for the vehicle body and wheel structures are the effective length L of the vehicle body. car Effective width of vehicle body W car Wheelbase W of left and right wheels 轮距 Fixed dimensional parameters provide a precise structural basis for solving the differential motion of four wheels.

[0066] After completing the coordinate system and parameter calibration, the system enters the real-time four-wheel encoder pose calculation stage. It completes the entire motion data acquisition and parameter calculation process within a fixed control cycle Δt, accurately reproducing the vehicle's real-time motion state. Within each control cycle Δt, the main control unit synchronously acquires the encoder pulse increment N corresponding to the four wheels: left front, right front, left rear, and right rear. iThese correspond to the real-time motion travel of the four wheels. Based on the relationship between encoder pulses and wheel travel distance, and using the wheel radius R and the total number of encoder pulses per revolution N, the motion is calculated. total The real-time independent linear velocities v of the four wheels are calculated. i denoted as v fl v fr v rl v rr It fully preserves the independent motion differences of each wheel, avoiding the problem of traditional overall average speed measurement smoothing out small posture deviations.

[0067] The formula for converting encoder pulses to linear velocity is: ;

[0068] Where: N i : The encoder pulse increment of the i-th wheel during time Δt. Specifically, in this paper, i refers to fl, fr, rl, and rr, which are used to represent the left front, right front, left rear, and right rear wheels, respectively.

[0069] It should be noted that the four-wheel mileage acquisition hardware of this invention is replaceable and is not limited to a single encoder type. Different sensor types such as photoelectric encoders and magnetic encoders can be used. At the same time, it is compatible with incremental encoders and absolute encoders. All types of encoders can realize independent pulse acquisition and speed and pose calculation of four wheels. It can be flexibly selected according to the actual detection scenario, cost budget and accuracy requirements.

[0070] Based on the independent linear velocities of the four wheels, the overall motion parameters of the vehicle body are further calculated: First, the average velocities of the left and right wheels are calculated separately, and the average linear velocity v of the right wheel is calculated separately. r Take the average speed of the right front and right rear wheels, and the average linear velocity v of the left wheel. l Take the average speed of the left front and left rear wheels; then calculate the overall linear velocity v of the vehicle body using the average speed of all four wheels. The calculation formula is: ;

[0071] The real-time yaw rate ω of the vehicle body is calculated by using the speed difference between the left and right wheels and the wheelbase parameter. The calculation formula is as follows: ;in:

[0072] v r It is the average linear velocity of the right wheel. ,

[0073] v l It is the average linear velocity of the left wheel. ,

[0074] W 轮距 The distance between the left and right wheels (wheel track).

[0075] Furthermore, by using linear velocity v and angular velocity ω, it can accurately distinguish between the translational motion of the vehicle body and the attitude deflection motion, adapting to complex dynamic conditions such as turning, changing direction, and attitude fine-tuning during the scanning process.

[0076] Based on the solved linear velocity, yaw rate, and real-time heading angle of the vehicle body, the system iteratively updates the absolute pose of the vehicle center in the world coordinate system every control cycle:

[0077] ,

[0078] ,

[0079] ,

[0080] Simultaneously, based on the rigid transformation relationship between the world coordinate system, vehicle coordinate system, and ultrasonic module coordinate system, combined with the pre-calibrated two-dimensional eccentricity offset vector (d... x ,d y The system uses the vehicle's real-time heading angle θ to accurately convert the vehicle's center pose to the ultrasonic detection center pose, and calculates the true coordinates (x, y, y) of the ultrasonic module center P in the world coordinate system in real time. P y P ):

[0081] ,

[0082] ,

[0083] Where, d x and d y It is the offset vector of the ultrasonic module center P relative to the vehicle body center C.

[0084] This spatial transformation method strictly matches the rigid installation characteristics of the equipment and can dynamically compensate for the nonlinear eccentricity offset error generated during vehicle rotation and attitude deflection, thus solving the defect that traditional fixed compensation methods cannot adapt to dynamic attitude changes.

[0085] This invention uses the center P of the ultrasonic testing module as the sole reference for trajectory tracking, position closure, and trajectory planning throughout the entire process. This ensures complete consistency between the control reference and the actual testing operation reference, fundamentally solving the problems of detection point offset, partial missed detection, and inaccurate defect location caused by eccentric installation. The entire pose calculation and eccentricity correction process described above relies entirely on the device's built-in encoder and calibration structural parameters, eliminating the need for external sensing devices such as vision or lidar. The hardware structure is simple, highly adaptable to various environments, and suitable for complex industrial testing scenarios such as high altitudes, low light, and confined spaces.

[0086] In addition to the aforementioned front-mounted eccentric mounting configuration, the ultrasonic module eccentric mounting structure of this invention can also be flexibly mounted at any position, such as the front of the vehicle body, the side edge, or the extended cantilever; corresponding to the offset vector (d x ,d y ( ) can be any combination of positive and negative values, not limited to conventional operating conditions d y =0、d x The front-mounted eccentricity configuration with a value >0 can adapt to various assembly structures such as centering, side offset, and rear placement. The core correction algorithm does not require modification, resulting in better equipment compatibility and operational scalability.

[0087] This invention abandons the traditional manual fixed-value trajectory arrangement method and adopts a hardware parameter driven adaptive trajectory generation mechanism. All trajectory spacing, overlap width and trajectory quantity parameters are automatically derived from ultrasonic testing hardware parameters and preset testing standards, eliminating the need for manual setting of trajectory parameters one by one, and greatly improving the degree of automation and adaptability to working conditions.

[0088] This invention employs a rounding-plus-one calculation logic to precisely adapt to boundary margins, effectively avoiding issues such as the end trajectory exceeding the detection area or undetected blind spots remaining at the edges of the material. The result is a uniformly arranged, gapless, low-redundancy, and fully covered parallel trajectory array, providing a precise and standardized trajectory foundation for subsequent full-area, blind-spot-free complementary scanning. This adaptive planning method can dynamically adapt to different probe specifications, material sizes, and overlap accuracy requirements, exhibiting extremely high versatility.

[0089] Before the actual operation, the system retrieves the pre-calibrated effective detection width W of the ultrasonic module. probe Percentage of overlap with preset detection P overlap First, the safe overlap width W of the trajectory is obtained through parameter calculation. overlap This ensures that there are no blind spots in the area where adjacent trajectories connect; then, by subtracting the overlap margin from the effective detection width, the standard effective step spacing D for a single scan is obtained. This maximizes the coverage of a single scan while avoiding missed detections, balancing detection integrity and operational efficiency.

[0090] like Figures 4a-4c As shown, after determining the standard step spacing D, the system combines the length L of the area to be detected. area Width W area Matching vehicle body length L car Vehicle width W car The structural footprint, adaptive calculation of the number of trajectories N in the length direction. L Number of trajectories N in the width direction W ,

[0091] The formula is as follows:

[0092] Trajectory overlap: ,

[0093] Effective step spacing: ,

[0094] Number of trajectories in the width direction: ,

[0095] Number of trajectories in the length direction: .

[0096] Based on the above parameters, the system automatically generates multiple sets of equally spaced parallel scanning line segments, forming a regular, full-coverage trajectory array. Combined with the equipment's assembly structure characteristics, the effective detection width W of the ultrasonic module... probe Typically smaller than the vehicle body width W car Furthermore, the ultrasonic module adopts an eccentric front-mounted structure at the front of the vehicle body (d x >0), a single linear scan in one direction cannot cover the area directly below the vehicle body and the lateral obstruction area. Therefore, this invention adopts a four-way orthogonal partition scanning method to eliminate structural blind spots through multi-posture and multi-directional partition operations.

[0097] To intuitively demonstrate the quantitative implementation effect of the parameter adaptive trajectory planning and initial pose calibration of the present invention, this embodiment selects a typical working condition of nondestructive testing for corrosion of large low-carbon steel plates as the application scenario, and performs complete parameter derivation and trajectory example calculation. The specific working condition configuration and quantitative results are as follows:

[0098] The object being tested is a large low-carbon steel plate. The overall dimensions of the plate, equipment structural parameters, process parameters, coordinate parameters, etc., are configured as follows.

[0099] 1) Establishing the coordinate system:

[0100] Origin point O: The starting point of the left edge of the board.

[0101] X-axis: along the length of the sheet material

[0102] Y-axis: Along the width of the sheet material (perpendicular to the direction of movement).

[0103] 2) The coordinate positions and values ​​are shown in Table 1:

[0104] Table 1:

[0105] <![CDATA[L area ]]> 5000mm Detection area length <![CDATA[W area ]]> 2000mm Detection area width <![CDATA[W probe ]]> 120mm Effective detection width of the probe <![CDATA[P overlap ]]> 12.5% Probe detection overlap coverage <![CDATA[L car ]]> 500mm Car body length <![CDATA[W car ]]> 350mm Car body width <![CDATA[d x ]]> 430mm Distance in the longitudinal direction from the center of the probe to the center of the vehicle body <![CDATA[d y ]]> 0mm Distance in width direction from the center of the probe to the center of the vehicle body <![CDATA[W overlap ]]> <![CDATA[W probe ×P overlap =15mm]]> Probe detection overlap coverage width D <![CDATA[W probe -W overlap =105mm]]> Track spacing

[0106] Car body center C: (W car / 2,L car / 2,0)=(175,250,0)

[0107] Ultrasound module P: (W car / 2+d y ,L car / 2+d x,0)=(175,680,0)

[0108] X range: [115, 4855],

[0109] Y range: [680, 2000],

[0110] ,

[0111] ,

[0112] 3) 90° scan

[0113] X-coordinate of the parallel line segment of the trajectory:

[0114] X1=W car / 2=175mm,

[0115] Xi=X1+(i-1)×D,i=1,2,...,45,

[0116] X45 = 175 + 44 × 105 = 4795 mm

[0117] Scan Y range:

[0118] Y start =L car / 2+d x =680mm,

[0119] Y end =W area =2000mm.

[0120] The trolley is initially positioned at the lower left corner of the area to be inspected, with the center P of the ultrasound module located on the parallel line segment of the first trajectory. Let the X-coordinate of the parallel line segment of the first trajectory be X90. start The initial coordinates of the ultrasound module center P are: (x...) P_0 ,y P_0 ) = (W car / 2+d y ,L car / 2+d x ).

[0121] under normal assembly conditions d y =0, at this time the initial coordinates of the vehicle center C are (W car / 2,L car / 2,0); the initial coordinates of the ultrasound module center P are (W car / 2+d y ,L car / 2+d x Strictly match and calibrate the eccentric vector to ensure zero deviation of the initial working reference.

[0122] In the initial state of operation, the wall-climbing trolley is precisely positioned in the lower left corner of the detection area, completing the initial pose calibration: the initial coordinates of the trolley center C match the calibration parameters of the trolley structure dimensions, and the ultrasonic module center P is positioned according to the preset eccentric offset vector (d). x ,d y The system is positioned on the first parallel line segment of the preset trajectory to ensure that the starting reference for the full-area scanning is accurate and consistent, and to avoid the problem of overall trajectory deviation caused by initial positioning deviation.

[0123] like Figure 3 As shown, the full-area scanning process of this invention is strictly executed in a fixed sequence, including four orthogonal working conditions: 90° forward main scanning, 0° right-side supplementary scanning, 270° reverse scanning, and 180° left-side supplementary scanning. The four trajectories are spliced ​​together and coordinated to completely cover all blind spots in front, rear, left, and right of the vehicle body. Each zone scanning corresponds to a fixed trajectory boundary interval, accurately defining the detection coverage range in each direction.

[0124] (1) 90° forward main scan: the ultrasound module along the world coordinate system Y W Traveling along the positive axis, covering the core detection area in front of the vehicle, with the trajectory boundary constraint as follows: X W Directional detection starting point X90 start =W car / 2, X90 end =W car / 2+(N L -1)xD,Y W Directional detection starting point Y90 start =L car / 2+d x Y90 end =W area .

[0125] (2) 0° right-side supplementary scan: ultrasound module along X W Traveling along the positive axis, specifically covering the missed detection area on the right side of the vehicle body, with the trajectory boundary constraint as follows: X W Directional detection starting point X0 start =X90 end X0 end =L area Y W Detection of starting point Y0 in direction start =W car / 2, Y0 end =Y0 start +(N W -1)xD.

[0126] (3) 270° reverse scanning: ultrasound module along Y WThe trajectory moves in the opposite direction along the axis, covering the rear undetected area caused by the vehicle body and connecting parts. The trajectory boundary constraint is: X W Directional detection starting point X270 start =X90 end X270 end =W car / 2, Y W Directional detection starting point Y270 start =L car / 2+d x Y270 end =0.

[0127] (4) 180° left-side supplementary scan: ultrasound module along X W The axis moves in the opposite direction, covering the missed area on the left side of the vehicle body. The trajectory boundary constraint is: X W Directional detection starting point X180 start =W car / 2-W probe / 2, X180 end =0, Y W Directional detection starting point Y180 start =W car / 2, Y180 end =Y180 start +(N W -1)xD.

[0128] The 90° forward main scan is the core main scan procedure. The ultrasonic module moves forward along the YW axis of the world coordinate system, covering most of the core detection area in front of the vehicle. It adopts a refined six-segment complete cycle motion logic to complete a set of standardized actions in sequence, including straight-line scanning, backward fixed-point turning, attitude rotation displacement compensation, straight-line backward, forward turning, and secondary straight-line compensation.

[0129] like Figures 5a-5g As shown, the specific step-by-step motion logic is as follows:

[0130] ① Straight line scanning stage: Control the trolley to move at a constant speed along the preset trajectory, from the starting point of the current trajectory to the end of the trajectory, to complete the full coverage scanning of a single line segment.

[0131] ② Reverse turning phase: Let the current position of point P on the ultrasound module be (x P ,y P The starting point of the next line's trajectory is (x) next ,y next The current orientation angle of the vehicle body is θ, and the offset vector of point P relative to the center C of the vehicle body is (d). x ,d yThe control objective for this stage is to ensure that point P accurately reaches the starting point of the next row, and that the line connecting PC coincides with the parallel line segment of the next row's trajectory. The orientation angle of the next row's trajectory is defined as θ. next The value is usually 0 or π;

[0132] The vehicle's target orientation angle satisfies ;

[0133] The target coordinates of the vehicle center satisfy: , .

[0134] The corresponding vehicle body rotation angle Δθ and displacement ΔS are:

[0135] ,

[0136] The position offset is the Euclidean distance between the two points; the entire journey is based on the current position of point P (x P ,y P ( ) is the unique center of rotation, and the radius of rotation is: ,

[0137] ③ Linear compensation stage after rotation: After the vehicle body completes its fixed-point rotation, if point P does not accurately align with the starting point of the next row, the system automatically calculates the positional deviation: ,

[0138] Perform linear displacement compensation along the next trajectory direction to eliminate minor rotational errors.

[0139] ④ Straight backward movement phase: After point P completes the full scan of the current line and reaches the end of the trajectory, control the trolley to move backward in a straight line along the current trajectory and reset to the starting position of the current line to prepare for changing lanes and turning.

[0140] ⑤ Forward Turning Phase: After the car returns to the current starting point, it performs a forward turning action. Based on the real-time coordinates of point P and the target trajectory posture, the angle △θ and displacement △S are calculated to ensure that point P accurately lands on the next trajectory segment after the action is completed, and the line connecting P and C is parallel and coincident with the trajectory.

[0141] ⑥ Secondary straight-line compensation stage after rotation: After the forward turn is completed, a small straight-line backward compensation is performed to compensate for the residual position deviation, ensuring that point P is accurately aligned with the starting point of the trajectory.

[0142] Repeat the above six actions in a loop until all trajectory segments in the 90° positive direction have been scanned.

[0143] Throughout this process, the center P of the ultrasonic module serves as the sole center of rotation, ensuring that the detection benchmark remains constant during line changes. Only minor adjustments to the vehicle's attitude are made, eliminating trajectory offset and landing point misalignment issues introduced by the line change action. Simultaneously, based on the orientation angle and eccentricity vector of the next trajectory, the system accurately calculates the target attitude angle and target position coordinates of the vehicle, achieving seamless connection between attitude switching and trajectory docking. This ensures that the trajectory in the main area is smooth, continuous, and without offset or gaps.

[0144] After the forward main scan is completed, the equipment completes the transition of working conditions and attitude switching by rotating around the center of the ultrasonic module and axially translating, and enters the 0° right-side supplementary scan procedure (see...). Figure 6a , Figure 6b ).

[0145] Transition process from the 0° right-side supplementary scanning starting position: After the 90° scan is completed, the trolley stops in the upper right corner area of ​​the board, with the trolley body facing θ=90°, and point P located at (X90). end Y90 end The system completes the working condition switch in two steps: the first step is fixed-point rotation, with point P as the rotation center and d as the rotation point. x The first step involves rotating the vehicle by 90° to change its orientation from θ=90° to θ=0°, keeping point P fixed throughout the rotation. The second step involves axial translation, controlling the vehicle to move along the Y-axis. W Shift the axis in the negative direction, changing the ordinate of point P from Y90. end Adjust to Y0 start =W car / 2, complete the 0° scan starting attitude and position calibration.

[0146] After attitude calibration is completed, the standardized scanning process begins: ① In the straight-line scanning phase, all four wheels are controlled to move forward at a constant speed, so that point P moves along the X-axis. W The positive axis starts from the trajectory starting point (X0) start Y0 start Walk to the end of the track (X0) end Y0 end The linear travel distance for a single scan is |X0 end -X0 start |;

[0147] ② During the trajectory changing phase, after a single trajectory is scanned, the current position of point P is used as the rotation center, and the trajectory is shifted laterally by a standard step distance D using an arc-shaped backward movement to settle at the starting point of the next trajectory; the straight-line scanning and arc-shaped trajectory changing actions are repeated cyclically until point P reaches the end of the last 0° trajectory (X0). end Y0 end The 0° right-side supplementary scan was deemed complete. During the 0° right-side supplementary scan, the ultrasound module moved along the X-ray... WThe system moves along the axis in the forward direction, employing a strategy of linear constant-speed scanning combined with fixed-point arc reversal and line changing to accurately fill blind spots obstructed by the right side structure of the vehicle body and improve the detection coverage of the edge area in the width direction of the sheet metal.

[0148] After the right-side scan is completed, the system adjusts the vehicle's attitude and working position, and performs a 270° reverse scan (see...). Figure 7a , Figure 7b ).

[0149] 270° Reverse Scan Initial Adjustment Procedure: After the 0° scan is completed, control the vehicle body to rotate to (θ=3π / 2), so that the ultrasonic module orientation switches to Y. W In the negative axis direction, simultaneously move the trolley to 90° to scan the final trajectory segment, ensuring that point P accurately matches the starting coordinates (X270). start Y270 start ).

[0150] The 270° reverse scan employs the same six-segment zigzag path strategy as the 90° forward scan, only the direction of travel is reversed, along the Y... W The entire trajectory is traversed in the negative axis direction until point P reaches the end coordinate (X270). end Y270 end The scanning process is now complete. This procedure reuses the same set of trajectory segments as the 90° forward scan, but in the opposite direction. It can specifically cover the obstructed or missed areas on the front and rear sides of the vehicle body. At the same time, through bidirectional scanning verification, it eliminates the trajectory seam error of the unidirectional scan, greatly improving the trajectory fitting accuracy and detection completeness.

[0151] Finally, the equipment was reset to its initial posture and working position, and a 180° left-side supplementary scan was performed (see...). Figure 8 ).

[0152] 180° Left Side Re-scan Starting Requirements: After completing the 270° scan, control the trolley to return to the lower left corner of the detection area, and calibrate the coordinates of point P to the starting position (X180). start Y180 start The overall workflow of this procedure follows the 90° forward scan logic, employing a combination strategy of arc-shaped retraction and line-based constant-speed scanning. The ultrasonic module moves along the X-ray... W Moving in the opposite direction along the axis, it completes full coverage of parallel trajectory segments line by line, accurately filling the blind spot on the left side of the vehicle body.

[0153] Four sets of orthogonal processes work together in a closed-loop sequence, requiring no changes to the mechanical structure. The problem of blind spots caused by the eccentric structure is solved only through innovative operation strategies, achieving full coverage non-destructive testing.

[0154] Meanwhile, the entire scanning control system of this invention is not limited to conventional ultrasonic testing modules, but can also directly replace and adapt electromagnetic ultrasonic testing modules. Without modifying the core algorithm logic of trajectory planning, eccentricity correction, slippage compensation, and dual-loop control, it can still achieve coupling agent-free, high-efficiency automated electromagnetic ultrasonic non-destructive testing, effectively realizing the application of this invention in various industrial scenarios such as metal plate flaw detection, weld inspection, and equipment corrosion inspection.

[0155] Furthermore, due to limitations imposed by the vehicle's shape, mounting brackets, and the physical obstructions of the adsorption structure, there will be a small number of triangular and rectangular blind spots at the four corners of the sheet metal that cannot be covered by algorithmic strategies. These are inherent obstruction areas of the equipment's mechanical structure. In actual industrial testing conditions, blind spot compensation can be achieved through various adaptation methods: First, after the testing operation is completed, manual spot inspection can be performed on the blind spots at the four corners of the sheet metal; second, the ultrasonic module installation structure can be optimized by placing the ultrasonic module at the bottom or side of the vehicle body to eliminate obstruction at the structural level; third, a retractable ultrasonic probe module can be installed or a vehicle body tilt adjustment strategy can be matched to expand the testing coverage and achieve full-area, blind-spot-free testing of the sheet metal.

[0156] Furthermore, during operation on vertical and inclined walls, the wall-climbing vehicle is highly susceptible to wheel spin and lateral slippage due to factors such as wall roughness, gravitational forces, negative pressure adsorption fluctuations, and uneven load. This leads to a mismatch between the mileage data collected by the encoder and the actual vehicle posture, resulting in trajectory drift, trajectory errors, and data failure. Traditional fixed-threshold slippage detection methods cannot adapt to all speed conditions, exhibiting defects such as false positives at low speeds and false negatives at high speeds, resulting in extremely poor fault tolerance and accuracy. To address this, this invention establishes a multi-parameter adaptive slippage detection mechanism based on the discrete characteristics of four-wheel speeds, coupled with a multi-level compensation strategy of active anti-slip and posture reversal, to achieve accurate identification, rapid suppression, and precise trajectory repair of slippage faults.

[0157] The system collects independent speed data of the four wheels in real time throughout the entire scanning process, and calculates the standard deviation σ of the four wheel speeds in real time using a standardized formula. v This method precisely quantifies the consistency and dispersion of vehicle motion to characterize the occurrence and severity of slippage faults. The specific quantitative calculation logic for slippage determination is as follows:

[0158] (1) Calculation of the mean and standard deviation of the four wheel speeds

[0159] First, calculate the real-time average speed of the four wheels: ,

[0160] Next, calculate the standard deviation of the four wheel speeds to quantify the dispersion of each wheel speed:

[0161] ,

[0162] In the formula, v fl vfr v rl v rr These represent the real-time linear velocities of the left front, right front, left rear, and right rear wheels, respectively, v. i This refers to the real-time speed of a single wheel. When the vehicle is moving smoothly, the force on all four wheels is uniform, the rotation speed is consistent, and the speed dispersion remains at an extremely low level. When one or more wheels slip, spin freely, or sideslip, the rotation speed of the faulty wheel deviates abnormally from the normal range, the consistency of the four-wheel motion is disrupted, and the speed standard deviation increases significantly.

[0163] (2) Adaptive slip threshold calculation

[0164] To adapt to the operational requirements across the entire speed range, this invention employs an adaptive threshold that is positively correlated with the vehicle's travel speed, overcoming the technical shortcomings of traditional fixed thresholds. The threshold calculation formula is as follows: ,

[0165] In the formula, k is a preset proportional coefficient, and the value of k < 1 under normal operating conditions; at the same time, the system is configured with a minimum threshold σ. min This mechanism is used for low-speed limiting protection to prevent the adaptive threshold from approaching zero when the vehicle speed is too low, thus avoiding misjudging minute speed fluctuations as slippage faults. In low-speed conditions, the threshold is automatically lowered to improve recognition sensitivity and accurately capture minute slippage; in high-speed conditions, the threshold is automatically raised to filter out normal vehicle speed fluctuations and avoid invalid misjudgments.

[0166] (3) Slippage fault judgment conditions and fault location

[0167] The system employs a dual judgment mechanism of numerical over-limit and time-limit locking to effectively filter instantaneous disturbances: when the standard deviation of the four-wheel speed meets σ v >σ thresh Furthermore, the duration of the abnormal state exceeds the preset slippage duration T. slip (For example, under normal operating conditions, the value is within 1 second), which can determine that a wheel has a valid slippage fault. The fault location logic is as follows: compare the absolute value of the deviation between the real-time speed and the average speed of the four wheels, and the wheel with the largest speed deviation is determined to be the abnormal slippage wheel, so as to achieve precise location of the slippage position.

[0168] Based on accurate slippage identification, this invention configures a two-level linkage quantification compensation and repair mechanism. By combining quantifiable torque anti-slip suppression with historical pose retraction and reset, it achieves comprehensive and accurate repair of slippage faults. The specific quantification execution logic is as follows:

[0169] (1) First stage: Dynamic torque deslippage and active elimination of sideslip force

[0170] Upon detecting a slippage fault, the system immediately executes an anti-slip procedure: pausing the vehicle's continuous feed motion, uniformly zeroing the real-time speeds of all four wheels, and locking the vehicle's instantaneous motion state. For the identified abnormal slipping wheel, its drive torque and output speed are reduced to below the normal operating threshold; simultaneously, the drive torque output of the remaining normal wheels is moderately increased, using the controllable torque difference between wheels to counteract the lateral slippage force on the wall, thus suppressing slippage and sideslipping tendencies from the power source.

[0171] The duration of anti-slip regulation can be quantified and calculated: T comp =ν slip / a decel ,

[0172] In the formula, ν slip For the instantaneous speed of the vehicle body when it skids, a decel A preset deceleration is provided for the vehicle body. The system monitors the speed dispersion of the four wheels in real time, and when σ is satisfied... v <λ×σ thresh When the wheel sideslip force is completely eliminated and the four wheels return to uniform motion, the anti-slip control state is exited and the vehicle body resumes normal scanning motion; where the coefficient λ is an empirical threshold coefficient, with a benchmark value of 0.5, or it can be adaptively selected in the range of 0.4 to 0.6 according to the actual wall conditions and equipment load.

[0173] (2) Second level: Historical pose queue backtracking compensation

[0174] The system establishes a dynamic posture history queue throughout the process, and caches the posture information of the vehicle body and ultrasonic module for the most recent N control cycles in real time. The queue length N is configurable, for example, the normal value range is 3 to 100, and the steady-state operation posture data is completely retained.

[0175] When a slippage fault is triggered, the system automatically traverses the historical pose queue and selects the last steady-state normal pose before the slippage occurred. The steady-state determination criterion is that the standard deviation of the four wheel speeds in that cycle satisfies σ. v <λ×σ thresh The system forcibly resets the current vehicle center and ultrasonic detection module reference pose to this stable pose, while clearing all abnormal trajectory data stored after this normal pose, eliminating trajectory drift errors caused by slippage. After completing the pose reset, the system continues to execute the original preset scanning trajectory from the precise reset point, ensuring the continuity of the scanning operation and the integrity of the detection data, and preventing the continuous spread of accumulated slippage errors.

[0176] Furthermore, the basic four-wheel speed standard deviation slippage determination mechanism of this invention can be used independently or expanded and optimized through multi-source sensor data fusion. For example, it can be combined with attitude and angular velocity data from the IMU inertial measurement unit to assist in correcting the slippage determination results, improving the determination accuracy under attitude disturbance conditions; and it can superimpose wheel-ground contact pressure data from wheel pressure sensors to effectively distinguish between vehicle unloaded speed fluctuations and actual wheel slippage faults, reducing the probability of misjudgment. At the same time, this system can also introduce machine learning classification algorithms, train the model through working condition samples, and achieve high-precision slippage identification under complex wall surfaces and extreme load conditions, further improving the robustness and versatility of fault determination.

[0177] In addition, to address the problems of traditional single-loop PID control being unable to adapt to the characteristics of four-wheel independent drive, having lag in dynamic response, weak anti-disturbance capability, and low trajectory tracking accuracy, this invention adopts a dual-loop PID closed-loop control architecture with a layered nested position outer loop and a speed inner loop, which is decoupled and independent. This architecture achieves accurate macroscopic trajectory tracking and dynamic correction of microscopic single-wheel power, respectively, and is fully adapted to the dynamic operation characteristics of the wall-climbing robot on the vertical wall.

[0178] The real-time position (x) of the outer ring of the position relative to the center of the ultrasonic detection module. P ,y P ) and target location (x) P_target ,y P_target The deviation is the control input, and the target linear velocity v of the vehicle body is calculated through PID control. target With the target angular velocity ω target It focuses on controlling the accuracy of macro-trajectory alignment, effectively avoiding problems such as overall trajectory deviation, path deviation, and trajectory exceeding the boundary, and ensuring that the overall scanning path is smooth and regular and fits the preset trajectory.

[0179] The speed inner loop is equipped with an independent closed-loop control unit for each of the four drive wheels, based on the real-time linear velocity (v) of each wheel. fl_target ,v fr_target ,v rl_target ,v rr_target The deviation from the target linear velocity is used as the basis for adjustment. The real-time power output of each motor is independently corrected, and the speed disturbance caused by single-wheel resistance fluctuation, slight slippage, and uneven load is quickly suppressed. The consistency of power output of the four wheels is balanced, and the stability of vehicle movement is improved.

[0180] This invention employs a differentiated independent sampling period design for its dual-loop PID control. The outer position loop uses a low-frequency sampling mechanism (e.g., sampling period Δt = 10~50ms) to ensure smooth overall trajectory control without overshoot or jitter, adapting to long-distance trajectory tracking requirements. The inner velocity loop uses a high-frequency sampling mechanism (e.g., sampling period Δt_v = 1~5ms) to achieve millisecond-level instantaneous dynamic correction, quickly responding to dynamic conditions such as changes in wall resistance, attitude disturbances, and slight slippage. Simultaneously, the dual-loop PID parameters are systematically tuned using the critical proportional gain method or the Ziegler-Nichols method, fully adapting to the dynamic operating conditions of the climbing robot on a vertical wall.

[0181] The dual-loop hierarchical decoupling control mode separates macroscopic trajectory control from microscopic dynamic correction, while maintaining their coordination. This balances system stability, trajectory accuracy, and dynamic response capabilities, significantly improving trajectory tracking accuracy and vehicle motion stability under complex wall conditions. It provides reliable closed-loop control for high-precision, blind-spot-free, and continuous ultrasonic non-destructive testing operations across the entire domain.

[0182] The above-disclosed embodiments are merely one or more preferred embodiments of this application and should not be construed as limiting the scope of this application. Those skilled in the art will understand that all or part of the processes for implementing the above embodiments and equivalent variations made in accordance with the claims of this application are still within the scope of this application.

Claims

1. A method for full-coverage ultrasonic inspection of a wall-climbing vehicle based on a four-wheel encoder, wherein the four-wheel wall-climbing vehicle is fixedly equipped with an ultrasonic inspection module, each wheel is equipped with an independent encoder, and a main control unit is built into the vehicle body; characterized in that, The center of the ultrasonic testing module is offset from the center of the vehicle body. The scanning method specifically includes the following steps: S1. Multi-coordinate system establishment and parameter calibration: Construct a matching world coordinate system, vehicle coordinate system and ultrasonic module coordinate system, calibrate and store the two-dimensional center offset vector of the ultrasonic module center relative to the vehicle center, and calibrate the vehicle structure parameters, wheel running parameters, ultrasonic effective detection width and preset detection overlap rate to provide unified reference parameters for subsequent steps. S2. Real-time differential pose calculation and offset correction for four wheels: In each control cycle, the pulse increments of the four-wheel encoders are collected. Based on the pre-calibrated wheel running parameters and vehicle structure parameters, the real-time linear velocity of each wheel is calculated. Then, the linear velocity and angular velocity of the vehicle body are calculated, and the pose of the vehicle center is iteratively updated. Based on the rigid body transformation relationship of three sets of coordinate systems, combined with the vehicle body attitude angle and two-dimensional offset vector, the real-time detection pose of the ultrasonic module center in the world coordinate system is obtained, and the ultrasonic module center is used as the reference point for the whole trajectory tracking. S3, Four-way partition blind spot compensation scanning: Based on the calibrated ultrasonic parameters, a basic scanning trajectory is generated. For the blind spot caused by the eccentric installation of the detection module and the obstruction of the vehicle structure, a primary and secondary partition complementary scanning mechanism is adopted. The blind spot is closed by four sets of orthogonal angle partitions working independently and the trajectories are spliced ​​together. With the help of multi-posture collaborative motion, the eccentric detection dead zone is filled, and the wall surface is fully covered by scanning. S4. Multi-level slippage recognition and trajectory compensation: Real-time monitoring of the motion status of the four wheels, identification of wheel slippage faults, and correction of trajectory deviations through a multi-level method of wheel torque adjustment to eliminate slippage and historical normal posture retraction correction, and then continue to complete the scanning operation. S5. Dual-loop PID closed-loop precision control: Employing a dual-loop PID control architecture with an outer position loop and an inner speed loop, the outer position loop outputs the vehicle's target motion parameters based on the deviation between the actual position of the ultrasonic detection module's center and the target position. The inner speed loop adjusts the speed of each wheel drive motor based on the deviation between the actual linear speed of the four wheels and the target linear speed, thereby achieving closed-loop precise tracking of the scanning trajectory.

2. The ultrasonic inspection method for a wall-climbing vehicle based on a four-wheel encoder with full coverage scanning as described in claim 1, characterized in that, Step S2 employs a split-type velocity calculation and planar coordinate matching method to achieve pose correction, specifically including: The system collects independent pulse increments from all four wheels, combines these with vehicle structure parameters and wheel operating parameters to calculate the linear velocity of each wheel, and then uses the average velocity of the left and right wheels to calculate the vehicle's motion parameters. Based on the rigid body translation transformation relationship of three sets of coordinate systems, the center pose transformation of the ultrasonic module is completed by combining the two-dimensional plane offset and the vehicle body attitude angle.

3. The ultrasonic inspection method for a wall-climbing vehicle based on a four-wheel encoder with full coverage scanning as described in claim 1, characterized in that, Step S3 involves configuring a dedicated parameter arrangement logic for the trajectory generation task, specifically including: The pre-calibrated effective ultrasound detection width and preset detection overlap rate parameters are retrieved. Based on the ultrasound detection coverage parameters, multiple parallel scanning trajectories are planned, and adjacent scanning trajectories are arranged according to the preset overlap amount.

4. The ultrasonic inspection method for a wall-climbing vehicle based on a four-wheel encoder with full coverage scanning as described in claim 1, characterized in that, In step S3, the primary and secondary partition complementary scanning mechanism operates sequentially for four sets of orthogonal angle partitions. It performs multi-pose cooperative motion according to a fixed time sequence of 90° forward primary scanning, 0° right-side supplementary scanning, 270° reverse scanning, and 180° left-side supplementary scanning. The trajectories of each partition are generated independently and spliced ​​together. In the initial state of operation, the wall-climbing trolley is positioned in the lower left corner of the detection area. Based on the dimensions of the trolley body, a fixed offset vector between the center of the trolley body and the center of the ultrasonic module is calibrated, so that the center of the ultrasonic module is positioned on the first scanning trajectory line segment. The 90° forward main scan adopts a zigzag cyclic motion logic, which sequentially completes the straight-line scanning, backward turning with the center of the ultrasound module as the rotation reference, linear displacement compensation after rotation, straight backward, forward turning and secondary linear compensation actions, thus completing the main area trajectory traversal. After the 90° forward main scan is completed, the vehicle attitude and working position are switched by rotating around the center of the ultrasonic module and axially translating. The vehicle then enters the 0° right-side supplementary scan condition and completes the lateral blind zone trajectory coverage by using a straight constant speed scan and a fixed-point arc retraction and line change method. Then, the vehicle's attitude and position were adjusted, and a 270° reverse scan was performed using a zigzag trajectory. The forward and reverse scans were coordinated to cover the same set of trajectory segments. Finally, switch the initial posture and position of the operation, and perform a 180° left-side supplementary scan according to the 90° scan motion logic.

5. The ultrasonic inspection method for a wall-climbing vehicle based on a four-wheel encoder with full coverage scanning as described in claim 1, characterized in that, In step S4, the slippage fault identification adopts a multi-parameter joint judgment mechanism, which simultaneously collects wheel motion data and vehicle posture data, compares the real-time operating parameters with the preset normal operating parameter range, locates the slippage abnormal wheel based on the parameter deviation state, and outputs a slippage judgment signal.

6. The ultrasonic inspection method for a wall-climbing vehicle based on a four-wheel encoder with full coverage scanning as described in claim 5, characterized in that, Based on the multi-parameter joint judgment mechanism, the specific judgment logic for wheel slippage is as follows: calculate the standard deviation of the four wheel speeds in real time, combine the adaptive threshold that is positively correlated with the vehicle speed and the preset duration to complete the slippage judgment, and judge the wheel with the largest speed deviation as the abnormal slipping wheel.

7. The ultrasonic inspection method for a wall-climbing vehicle based on a four-wheel encoder with full coverage scanning as described in claim 5, characterized in that, In step S4, the wheel torque adjustment anti-slip method matches the vehicle's motion state and outputs the corresponding wheel set torque. The lateral slip generated during wheel operation is offset by dynamic wheel difference torque distribution, thereby achieving active suppression of lateral slip force.

8. The ultrasonic inspection method for a wall-climbing vehicle based on a four-wheel encoder with full coverage scanning as described in claim 5, characterized in that, In step S4, the specific compensation method for the historical pose rollback is as follows: pre-caching the historical pose data of the vehicle body and the ultrasonic detection module under normal operating conditions, and after determining the wheel slippage fault, resetting the current pose of the vehicle body and the ultrasonic detection module to the last normal pose before slippage, and continuing to execute the preset scanning trajectory based on the reset pose.

9. The ultrasonic inspection method for a wall-climbing vehicle based on a four-wheel encoder with full coverage scanning as described in claim 1, characterized in that, In step S5, the dual-loop PID control adopts an independent control architecture with inner and outer loops nested in a layered manner. The outer loop of position is used to achieve trajectory tracking control, and the inner loop of speed is used to achieve motor power control. The inner and outer loops work together to complete the closed-loop dynamic tracking of the scanning trajectory.

10. The ultrasonic inspection method for a wall-climbing vehicle based on a four-wheel encoder with full coverage scanning as described in claim 9, characterized in that, The outer position loop calculates and outputs the vehicle target motion parameters based on the position deviation of the ultrasonic module center, while the inner speed loop independently adjusts the corresponding drive motor speed based on the speed deviation of each wheel. The position loop and speed loop are configured with independent sampling periods and are adapted to the dynamic working conditions of wall scanning to complete the dual-loop PID parameter calibration.

Images