A cruise control method, device and medium for an inspection robot

Abstract

This invention discloses a cruise control method, device, and medium for an inspection robot, relating to the field of robot automatic control technology. The method includes: collecting inspection environment information and equipment operating status data, performing unified coordinate mapping and spatial modeling to generate an inspection environment spatial map; constructing an inspection environment risk field model based on the inspection environment spatial map and equipment operating status data, employing an equipment operating risk layer, a spatial obstacle constraint layer, and an inspection target guidance layer; extracting the robot's current position from the inspection environment spatial map, performing risk gradient direction decomposition in conjunction with the inspection environment risk field model, and reconstructing the cruise direction to generate a cruise guidance direction. This invention improves cruise stability and task completion consistency in complex industrial environments by extracting local risk perception data and decomposing risk change trends into multiple candidate gradient directions.

Description

Technical Field

[0001] This invention relates to the field of robot automatic control technology, and in particular to a cruise control method, device and medium for an inspection robot. Background Technology

[0002] With the continuous development of industrial internet, intelligent inspection, and autonomous mobile robot technologies, inspection robots have evolved from early remote control replacing manual inspection to possessing autonomous localization, path planning, and multi-source perception capabilities. Currently, related technological systems typically focus on environmental perception and mapping, localization and navigation, local obstacle avoidance, and path tracking control, combined with task point queues to achieve route-based inspection. The operational status data of numerous devices in industrial settings can be accessed via fieldbus or industrial networks, providing inspection tasks with a two-dimensional information foundation. Existing methods achieve stable, continuous, and safe cruise control in environments with complex obstacles, narrow passages, and densely packed equipment. Related research is also gradually shifting from the accessibility of single geometric paths to comprehensive control strategies oriented towards safety, robustness, and task priority.

[0003] However, existing methods still have several shortcomings in cruise control for industrial scenarios, directly impacting inspection efficiency and operational safety. Firstly, many solutions still rely on static maps, fixed routes, and local obstacle avoidance. Path or speed adjustments are primarily based on geometric obstacles, making it difficult to spatially translate the risk factors represented by the equipment's operating status into calculable guidance constraints. This results in a lack of consistent quantitative data on cruise control responses to risk areas. Secondly, the guidance direction or reference trajectory output by trajectory planning lacks a comprehensive evaluation of the coupling relationship between risk level and motion stability when encountering ground adhesion differences, abrupt curvature changes, and disturbances in narrow passages. This leads to unpredictable stability of control commands in different risk sections. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a cruise control method for inspection robots to solve the problems of unstable cruise guidance direction and lack of reliable basis for motion control compensation.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides a cruise control method for an inspection robot, comprising: collecting inspection environment information and equipment operating status data and performing unified coordinate mapping and spatial modeling processing to generate an inspection environment spatial map; based on the inspection environment spatial map and equipment operating status data, constructing an inspection environment risk field model by employing an equipment operating risk layer, a spatial obstacle constraint layer, and an inspection target guidance layer; extracting the robot's current position in the inspection environment spatial map, performing risk gradient direction decomposition in conjunction with the inspection environment risk field model, and performing cruise direction reconstruction to generate a cruise guidance direction; determining the inspection robot's cruise trajectory based on the cruise guidance direction, and performing a comprehensive evaluation of motion stability using a risk coupling stability assessment method to obtain a cruise stability index; and performing motion control compensation on the cruise guidance direction based on the cruise stability index to generate cruise control commands.

[0008] In a preferred embodiment of the cruise control method for the inspection robot described in this invention, the inspection environment information includes equipment location data and obstacle location data.

[0009] As a preferred embodiment of the cruise control method for the inspection robot described in this invention, the inspection environment spatial map includes the location of equipment nodes, spatial distribution information of obstacles, location of inspection target points, and the current location of the robot.

[0010] The specific steps for collecting inspection environment information and equipment operating status data, performing unified coordinate mapping and spatial modeling processing, and generating an inspection environment spatial map are as follows:

[0011] Perform unified coordinate mapping on equipment location data and obstacle location data to generate unified coordinate environment data;

[0012] Based on unified coordinate environment data, spatial structure modeling is performed using a spatial raster modeling algorithm to generate a spatial map of the inspection environment.

[0013] As a preferred embodiment of the cruise control method for the inspection robot described in this invention, the step of constructing an inspection environment risk field model based on the inspection environment spatial map and equipment operating status data, employing an equipment operating risk layer, a spatial obstacle constraint layer, and an inspection target guidance layer, is as follows:

[0014] The equipment operation risk layer is based on equipment operation status data and equipment node location. It uses an equipment status risk mapping algorithm to spatially label the equipment operation status data and generate an equipment risk distribution matrix.

[0015] The spatial obstacle constraint layer is based on the spatial distribution information of obstacles. It uses a spatial safety distance expansion algorithm to perform geometric expansion of the area around the obstacle to generate a spatial passability constraint matrix.

[0016] The inspection target guidance layer calculates the target guidance vector for the robot to point towards the inspection target based on the location of the inspection target point and the robot's current position;

[0017] The risk field model of the inspection environment is constructed by superimposing the equipment operation risk layer, the spatial obstacle constraint layer, and the inspection target guidance layer.

[0018] As a preferred embodiment of the cruise control method for the inspection robot described in this invention, the specific steps of extracting the robot's current position from the inspection environment spatial map and performing risk gradient direction decomposition in conjunction with the inspection environment risk field model are as follows:

[0019] Extract the robot's current position from the inspection environment spatial map, input the inspection environment risk field model, and determine the corresponding local risk perception data;

[0020] Based on local risk perception data, the risk change trend at the robot's current position is calculated and decomposed into multiple candidate risk gradient directions to generate a set of risk gradient directions.

[0021] As a preferred embodiment of the cruise control method for the inspection robot described in this invention, the specific steps for generating the cruise guidance direction are as follows:

[0022] Based on the risk gradient direction set, a risk score is calculated using a risk direction screening algorithm, and a set of safe directions is selected.

[0023] Based on the location of the inspection target point, the set of safe directions is combined to generate the cruise guidance direction.

[0024] As a preferred embodiment of the cruise control method for the inspection robot described in this invention, the steps of determining the cruise trajectory of the inspection robot based on the cruise guidance direction and comprehensively evaluating the motion stability using the risk coupling stability assessment method to obtain the cruise stability index are as follows:

[0025] Based on the cruise guidance direction, point-by-point extrapolation is performed on the inspection environment spatial map, and navigability is verified by combining the spatial navigability constraint matrix to form a cruise trajectory reference line;

[0026] The target heading is obtained by performing trajectory tangent analysis on the cruise trajectory reference line, and then smoothed by curvature constraint to generate trajectory control reference quantity;

[0027] The cruise trajectory reference line is spatiotemporally aligned with the trajectory control reference quantity, and a risk coupling stability assessment is performed to obtain the cruise stability index.

[0028] As a preferred embodiment of the cruise control method for the inspection robot described in this invention, the specific steps for generating cruise control commands by performing motion control compensation on the cruise guidance direction based on the cruise stability index are as follows:

[0029] Cruise stability indicators are classified into cruise states, and hysteresis switching is determined to obtain compensation control gating parameters.

[0030] Based on the compensation control gating parameters, the cruise guidance direction is subjected to amplitude limiting, gradual change and continuity constraints to generate a compensated cruise guidance direction.

[0031] The curvature-constrained velocity mapping method is used to convert the compensated cruise guidance direction into linear velocity and angular velocity commands, thereby generating cruise control commands.

[0032] In a second aspect, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, it implements any step of the cruise control method for the inspection robot as described in the first aspect of the present invention.

[0033] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the cruise control method for the inspection robot as described in the first aspect of the present invention.

[0034] The beneficial effects of this invention are as follows: by extracting local risk perception data, the risk change trend is decomposed into multiple candidate gradient directions, and a set of safe directions is selected through directional risk scoring; by combining the inspection target points to perform directional combination and path continuity constraints, the risk information is directly transformed into executable and continuous cruise guidance input; it effectively avoids the local lock-up and directional jitter that are prone to occur in traditional potential field and single-path guidance, and improves the cruise stability and mission completion consistency in complex industrial environments. Attached Figure Description

[0035] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0036] Figure 1 This is a flowchart of the cruise control method for the inspection robot.

[0037] Figure 2 A flowchart for generating an inspection environment spatial map.

[0038] Figure 3A flowchart for constructing an inspection environment risk field model.

[0039] Figure 4 A flowchart for generating cruise guidance direction.

[0040] Figure 5 This is a schematic diagram illustrating enhanced cruise guidance continuity.

[0041] Figure 6 This is a verification diagram of cruise stability distribution. Detailed Implementation

[0042] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0043] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0044] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0045] Reference Figures 1-6 As one embodiment of the present invention, this embodiment provides a cruise control method for an inspection robot, comprising the following steps:

[0046] S1: Collect inspection environment information and equipment operation status data, perform unified coordinate mapping and spatial modeling processing, and generate an inspection environment spatial map.

[0047] S1.1: Inspection environment information includes equipment location data and obstacle location data.

[0048] Specifically, the equipment location data is collected by environmental sensing devices (such as lidar, depth cameras, and ultra-wideband positioning tag readers) measuring the spatial position within the inspection area. After the environmental sensing devices identify the outer contour of the equipment and the positioning mark set by the equipment, they output the position record of the equipment in the coordinate system of the inspection area and write the position record and the corresponding equipment identification information to generate equipment location data.

[0049] The environmental sensing device scans physical obstructions (such as pipe supports, valve bases, cable tray columns, and temporary stacks) in the passage area within the inspection area. After performing obstacle clustering and segmentation on the scanned point cloud or depth information, the environmental sensing device outputs the location records corresponding to the boundary range of the obstacles and writes them together to generate obstacle location data.

[0050] S1.2: Perform unified coordinate mapping on the equipment location data and obstacle location data to generate unified coordinate environment data.

[0051] Specifically, the system reads the device identification information, collection time, and coordinate values ​​in the inspection area coordinate system for each location record in the device location data; and reads the collection time and coordinate values ​​corresponding to the obstacle boundary range for each location record in the obstacle location data.

[0052] Unify the coordinate values ​​of equipment location data and obstacle location data to the same coordinate field format under the inspection area coordinate system, and unify the coordinate unit expression (e.g., unify to meters and retain a fixed number of decimal places). Sort the equipment location data and obstacle location data in ascending order of acquisition time.

[0053] The device location data and obstacle location data are timestamped using a fixed time window. When there are multiple device location data records in each time window, they are grouped according to the device identification information. Within each group, the coordinate values ​​are averaged and merged to generate a merged device location record within the time window. When there are multiple obstacle location data records in each time window, the set of grids covered by the obstacle boundary within the same time window is joined to generate a merged obstacle location record within the time window.

[0054] The merged device location records and merged obstacle location records within the same time window are written into a unified data table (including the acquisition time, merged device location records, and merged obstacle location records) according to the acquisition time, generating unified coordinate environment data.

[0055] S1.3: Based on unified coordinate environment data, spatial structure modeling is performed using a spatial raster modeling algorithm to generate an inspection environment spatial map.

[0056] Specifically, the algorithm reads the coordinate values ​​of the merged device position records and the coordinate values ​​corresponding to the obstacle boundary ranges of the merged obstacle position records from the unified coordinate environment data, and reads the grid size of the inspection area coordinate system (determined based on the robot's outline dimensions). The spatial grid modeling algorithm divides the inspection area into row and column grids based on the grid size, and rounds down the ratios of the horizontal and vertical components of each coordinate value to the grid size to obtain the grid column number and grid row number, thus determining the corresponding grid index.

[0057] The grid index corresponding to the merged device location record is written to the device node location, and the grid index corresponding to the obstacle boundary range is marked as obstacle spatial distribution information. The inspection target point location and the robot's current location are matched using the same grid positioning to obtain the inspection target point location and the robot's current location; the device node location, obstacle spatial distribution information, inspection target point location and robot's current location are summarized to generate an inspection environment spatial map.

[0058] The inspection environment spatial map includes the location of equipment nodes, spatial distribution information of obstacles, location of inspection target points, and the current position of the robot.

[0059] It should be noted that the spatial grid modeling algorithm divides the inspection area into row and column grids according to the grid size, and positions the coordinate values ​​in the unified coordinate environment data to the corresponding grid index by rounding the ratio. In this way, the device node position and obstacle spatial distribution information are marked on the grid and written into the inspection target point position and the robot's current position, generating an inspection environment spatial map.

[0060] S2: Based on the inspection environment spatial map and equipment operation status data, an inspection environment risk field model is constructed by adopting the equipment operation risk layer, spatial obstacle constraint layer and inspection target guidance layer.

[0061] S2.1: The equipment operation risk layer uses equipment operation status data and equipment node locations to perform spatial risk labeling on the equipment operation status data through an equipment status risk mapping algorithm, generating an equipment risk distribution matrix.

[0062] Specifically, the system reads the equipment identification information, collection time, and equipment operating status information contained in the equipment operating status data; it also reads the equipment node locations in the inspection environment spatial map and obtains the equipment identification information corresponding to the equipment node locations; and matches the equipment operating status data with the equipment node locations one by one according to the equipment identification information to obtain the equipment correspondence table.

[0063] The equipment status risk mapping algorithm summarizes all values ​​of the same parameter name for the equipment operating status information in the equipment corresponding table according to the collection time window, calculates the minimum and maximum values ​​within the group, and obtains the difference between the minimum and maximum values ​​within the group as the range item within the group; the difference between the value of each equipment operating status information and the minimum value is used as the offset item within the group; the ratio of the range item within the group to the offset item within the group is mapped to the relative position within the group and arranged in ascending order to form a risk ranking sequence.

[0064] The equipment status risk mapping algorithm divides the risk sorting sequence into three equal parts, which are low risk level, medium risk level and high risk level in ascending order, and maps them to risk level values ​​(the level value of low risk level is 1, the level value of medium risk level is 2, and the level value of high risk level is 3), and writes them into the equipment risk record table along with the equipment node position.

[0065] It should be noted that the state risk mapping algorithm summarizes the values ​​in the equipment operating status information according to the collection time window, and calculates the minimum value and the maximum value within the group to obtain the relative position within the group; sorts them according to the relative position within the group and performs trisection quantile division to output the risk level value; finally, the risk level value and the corresponding equipment node position are written into the equipment risk distribution matrix.

[0066] Read the coordinate values ​​corresponding to the device node locations and the grid size used in the inspection environment spatial map. Round the ratio of the horizontal component of the coordinate values ​​to the grid size to obtain the grid column number. Round the ratio of the vertical component of the coordinate values ​​to the grid size to obtain the grid row number. Combine the grid row number and grid column number to form the grid index corresponding to the device node location. Write the risk level from the device risk record table to the corresponding grid index to obtain the grid matrix. Write the grid indexes that are not written to the device node locations to the default risk level. Summarize the grid matrices to generate the device risk distribution matrix.

[0067] S2.2: The spatial obstacle constraint layer uses the spatial distribution information of obstacles to perform geometric expansion of the area around the obstacles through the spatial safety distance expansion algorithm, generating a spatial passability constraint matrix.

[0068] Specifically, the system reads the spatial distribution information of obstacles in the inspection environment spatial map and obtains the set of grid indices corresponding to the spatial distribution information of obstacles; it also reads the grid size used in the inspection environment spatial map and uses the grid size as the basis for the expansion step size of the spatial safety distance expansion algorithm.

[0069] The spatial safety distance expansion algorithm writes the set of grid indices corresponding to the spatial distribution information of obstacles into the queue to be expanded and into the expansion set; it retrieves the grid indices from the queue to be expanded and enumerates the adjacent grid indices of the grid indices. The adjacent grid indices are determined by the difference of one grid row number or one grid column number; the adjacent grid indices are added to the queue to be expanded, and the adjacent grid indices that have already been included are no longer written into the queue to be expanded.

[0070] Read the spatial safety distance (robot outline dimensions) and round down the ratio of the spatial safety distance to the grid size to obtain the number of expansion layers; record the number of grid index layers in the queue to be expanded, and stop adding adjacent grid indices after the number of grid index layers reaches the number of expansion layers.

[0071] Mark the raster indices corresponding to the extended set as impassable and mark the raster indices not marked as impassable as passable; write the impassable and passable indices into the raster matrix according to the raster index to form a spatial passability constraint matrix.

[0072] It should be noted that the spatial safety distance expansion algorithm uses the set of grid indices corresponding to the spatial distribution information of obstacles as the starting set, expands layer by layer according to adjacent grid indices to the expansion layer number corresponding to the spatial safety distance, and marks the expansion set as impassable and the grid indices as passable to generate a spatial passability constraint matrix.

[0073] The spatial navigability constraint matrix transforms the spatial distribution information of obstacles into grid constraint results that can be directly used in cruise decisions after being processed by the spatial safety distance extension algorithm. This enables the robot to not only avoid the obstacle itself, but also to simultaneously avoid the safety buffer zone around the obstacle. It provides a unified and definite spatial boundary basis for subsequent cruise trajectory reference line generation, navigability verification, and safety direction selection, thereby improving the safety, stability, and trajectory executability of the cruise process.

[0074] S2.3: The inspection target guidance layer calculates the target guidance vector for the robot to point towards the inspection target based on the location of the inspection target point and the robot's current position.

[0075] Specifically, the system reads the location of the inspection target point and the robot's current position from the inspection environment spatial map, and obtains the coordinate values ​​corresponding to the inspection target point and the robot's current position. In the inspection target guidance layer, the difference between the coordinate values ​​of the inspection target point and the robot's current position is used as the direction component, and the target guidance vector is calculated. The expression is as follows:

[0076]

[0077] Where G is the target guiding vector. Let P be the lateral coordinate of the inspection target point. Let P be the longitudinal coordinate of the inspection target point. Let q be the horizontal coordinate of the robot's current position. Let q be the vertical coordinate of the robot's current position.

[0078] It should be noted that the coordinate components of the position vector in the target guiding vector formula have the same dimension, which is the unit of length. The position difference vector has the same dimension as the length vector. Therefore, the target guiding vector is a dimensionless unit vector with consistent dimensions.

[0079] S2.4: Overlay and model the risk field of the equipment operation risk layer, the spatial obstacle constraint layer, and the inspection target guidance layer to construct an inspection environment risk field model.

[0080] Specifically, the system reads the equipment risk distribution matrix output by the equipment operation risk layer and extracts the grid row and column range of the equipment risk distribution matrix; it also reads the spatial accessibility constraint matrix output by the spatial obstacle constraint layer and extracts the grid row and column range of the spatial accessibility constraint matrix; and it extracts the direction component of the target guidance vector output by the inspection target guidance layer.

[0081] Using each grid row and column number combination within the grid range as the traversal grid, the passage marker of the spatial traversability constraint matrix is ​​read first at each grid row and column number combination position. When the passage marker is an impassable marker, the impassable risk marker is written into the corresponding position of the unified grid table and the target guiding vector direction component is written. When the passage marker is a traversable marker, the risk level of the equipment risk distribution matrix at the corresponding position of the unified grid table is read and written into the corresponding position of the unified grid table and the target guiding vector direction component is written. The unified grid table is summarized to form the inspection environment risk field model.

[0082] S3: Extract the robot's current position from the inspection environment spatial map, combine it with the inspection environment risk field model to perform risk gradient direction decomposition, and perform cruise direction reconstruction to generate cruise guidance direction.

[0083] S3.1: Extract the robot's current position from the inspection environment spatial map, input the inspection environment risk field model, and determine the corresponding local risk perception data.

[0084] Specifically, the robot's current position in the inspection environment spatial map is read and the corresponding coordinate value is obtained. The grid size used in the inspection environment spatial map is read and the coordinate value corresponding to the robot's current position is rasterized. The rasterization process rounds the ratio of the horizontal component of the coordinate value to the grid size to obtain the grid column number and the ratio of the vertical component of the coordinate value to the grid size to obtain the grid row number. The grid row number and the grid column number are combined to obtain the grid index corresponding to the robot's current position.

[0085] The system reads the inspection environment risk field model and locates the risk level, impassable risk marker, and target guidance vector direction component corresponding to the grid index of the robot's current position. Based on the grid index corresponding to the robot's current position, the system reads the risk level and impassable risk marker corresponding to the inspection environment risk field model one by one and writes them into the local grid table. It also writes the target guidance vector direction component corresponding to each grid index to obtain local risk perception data.

[0086] S3.2: Based on local risk perception data, calculate the risk change trend of the robot's current position and decompose it into multiple candidate risk gradient directions to generate a set of risk gradient directions.

[0087] Specifically, the system reads local risk perception data and locates the grid index corresponding to the robot's current position. It then reads the risk level of the grid index corresponding to the robot's current position and the set of adjacent grid indices (grid indices whose row numbers or column numbers differ by one grid).

[0088] For each adjacent set of grid indices, read the risk level of the corresponding grid index of the local risk perception data, and take the difference between the risk level value of the adjacent grid index and the risk level value of the grid index corresponding to the robot's current position as the risk difference value. Write the directional relationship between the risk difference value and the adjacent grid index relative to the grid index corresponding to the robot's current position into the directional difference table and sort it from largest to smallest according to the absolute value of the risk difference value to obtain the risk gradient direction set.

[0089] S3.3: Based on the risk gradient direction set, calculate the directional risk score through the risk direction screening algorithm, and screen the safe direction set.

[0090] Specifically, a risk direction screening algorithm is used to generate a forward grid index sequence for each candidate risk gradient direction in the risk gradient direction set, starting from the grid index corresponding to the robot's current position and proceeding along the candidate risk gradient direction. For each forward grid index sequence, the inaccessible risk marker and risk level of the corresponding grid index in the local risk perception data are read. If an inaccessible risk marker exists, the candidate risk gradient direction is added to the unavailable direction set; if no inaccessible risk marker exists, a direction risk score is calculated, expressed as:

[0091]

[0092] Where S is the directional risk score, and K is the number of forward raster index sequence steps. This represents the risk level value corresponding to the Kth raster index, where K is the raster index. It is the absolute value of the difference between the risk level values ​​of adjacent steps.

[0093] It should be noted that the risk level value and the inaccessible risk marker in the directional risk scoring formula are both dimensionless quantities. Risk ranking, summation and absolute value operations do not introduce new physical dimensions. Therefore, the directional risk score is dimensionless and the dimensions remain consistent.

[0094] The directional risk score and candidate risk gradient directions are written into the directional score table and sorted in the directional score table from smallest to largest. Candidate risk gradient directions that are not in the unavailable directional set are aggregated to form a safe directional set.

[0095] It should be noted that the risk direction screening algorithm takes the risk gradient direction set as input, generates a forward grid index sequence from the grid index corresponding to the robot's current position along each candidate direction, determines that the candidate direction is unusable when an impassable risk marker appears, and sums the risk level values ​​of the sequence to obtain a direction risk score when no impassable risk marker appears, and then screens the set of safe directions according to the direction risk score.

[0096] S3.4: Based on the location of the inspection target point, combine the safe direction set to generate the cruise guidance direction.

[0097] Specifically, the target point location and the robot's current location are read from the inspection environment space map. The difference between the horizontal component of the target point location coordinates and the horizontal component of the robot's current location coordinates is taken as the horizontal direction component. The difference between the vertical component of the target point location coordinates and the vertical component of the robot's current location coordinates is taken as the vertical direction component. The target direction vector (including the horizontal and vertical direction components) is obtained.

[0098] For each safe direction vector in the safe direction set, the direction difference between the safe direction vector and the target direction vector is calculated, and the safe direction vector with the smallest direction difference is selected as the cruise guidance direction.

[0099] S4: Determine the patrol trajectory of the inspection robot based on the cruise guidance direction, and use the risk coupling stability assessment method to comprehensively assess the motion stability and obtain the cruise stability index.

[0100] S4.1: Based on the cruise guidance direction, perform point-by-point extrapolation in the inspection environment spatial map, and combine it with the spatial navigability constraint matrix to perform navigability verification, forming a cruise trajectory reference line.

[0101] Specifically, the grid index corresponding to the robot's current position is determined in the inspection environment space map. The offset direction of the adjacent grid index is determined by the cruise guidance direction. Starting from the grid index corresponding to the robot's current position, the grid index corresponding to the next position is generated sequentially according to the offset direction of the adjacent grid index. The grid index corresponding to the next position that is retained each time is written into the trajectory point sequence in order.

[0102] For each grid index corresponding to the next position generated, the corresponding access marker is queried in the spatial accessibility constraint matrix. If the access marker is accessible, the next grid index corresponding to the next position is generated. If the access marker is inaccessible, generation stops and the current grid index corresponding to the next position is not written. The trajectory point sequence is connected in the order of generation to form the cruise trajectory reference line.

[0103] S4.2: Perform trajectory tangent analysis on the cruise trajectory reference line to obtain the target heading, and generate trajectory control reference quantities through curvature constraint smoothing.

[0104] Specifically, adjacent points are taken sequentially along the cruise trajectory reference line to form adjacent trajectory point pairs. The difference between the coordinate values ​​of the next trajectory point and the coordinate values ​​of the previous trajectory point in each adjacent trajectory point pair is taken as the tangent direction component. The tangent direction component is taken as the target heading and written into the heading sequence in the order of the trajectory point sequence.

[0105] The heading sequence is checked one by one according to the order of adjacent items to examine the change range of two adjacent headings, and the average of the previous heading, the current heading, and the next heading is calculated and written into the position of the current heading to obtain a smooth heading sequence. The smooth heading sequence is matched one-to-one with the cruise trajectory reference line according to the trajectory point sequence and summarized to generate trajectory control reference quantity.

[0106] S4.3: Spatiotemporally align the cruise trajectory reference line with the trajectory control reference quantity, perform a risk coupling stability assessment, and obtain cruise stability indicators.

[0107] Specifically, a one-to-one correspondence is established between the trajectory point sequence of the cruise trajectory reference line and the smooth heading sequence of the trajectory control reference quantity according to the sequence of trajectory points, and written into an alignment table (each row contains the coordinate values ​​of the trajectory points with the same sequence number and the smooth heading).

[0108] Based on the alignment table, locate the raster index corresponding to each trajectory point and query the corresponding risk level value and impassable risk marker in the inspection environment risk field model. If the impassable risk marker exists, replace the risk level value with the risk level value corresponding to the impassable risk marker and write it into the risk sequence; if the impassable risk marker does not exist, write the risk level value into the risk sequence.

[0109] Record the smoothed heading difference between adjacent rows in the alignment table according to row number order and obtain the heading change range. Then calculate the cruise stability index, expressed as:

[0110]

[0111] Where D is the cruise stability index, and N is the number of rows in the alignment table. Indicates the alignment table number The corresponding smooth heading, Indicates the alignment table number The risk level value corresponding to the smoothed course. To align the row number index of the table, It is the absolute value of the smooth heading difference between two adjacent rows.

[0112] It should be noted that the difference between adjacent smooth headings in the cruise stability index formula has an angular dimension, while the risk level value is dimensionless. The coupled accumulation maintains a unified angular dimension and does not introduce other physical dimensions. Therefore, the dimensions of the cruise stability index are consistent.

[0113] It should be noted that the risk coupling stability assessment method establishes a one-to-one correspondence between the cruise trajectory reference line and the trajectory control reference quantity according to the trajectory point sequence. It queries the risk level value corresponding to each trajectory point in the inspection environment risk field model to form a risk sequence, and obtains the course change amplitude by recording the difference between adjacent terms in the smoothed course sequence and outputs the cruise stability index.

[0114] S5: Based on the cruise stability index, perform motion control compensation on the cruise guidance direction to generate cruise control commands.

[0115] S5.1: Divide the cruise stability index into cruise states, perform hysteresis switching determination, and obtain compensation control gating parameters.

[0116] Specifically, the cruise stability index is written into the stability sequence at the end of each control cycle, and the most recently written cruise stability index is taken from the stability sequence as the comparison benchmark. The cruise stability index written in the current control cycle is compared with the comparison benchmark and the cruise state division result is output. When the cruise stability index of the current control cycle is greater than the comparison benchmark, the compensation state is output. When the cruise stability index of the current control cycle is less than the comparison benchmark, the stable state is output.

[0117] The system calculates the cruise state classification results corresponding to the most recent consecutive control cycles (e.g., the most recent 2 to 3 consecutive control cycles) in the statistical stability sequence, and allows the cruise state classification results to switch when the corresponding cruise state classification results are consistent; when the cruise state classification results corresponding to the most recent consecutive control cycles are inconsistent, the system outputs the cruise state classification results that maintain a stable state; and the system uses the cruise state classification results and hysteresis switching determination results together as compensation control gating parameters and outputs them.

[0118] S5.2: Based on the compensation control gating parameters, the cruise guidance direction is subjected to amplitude limiting, gradual change and continuity constraints to generate a compensated cruise guidance direction.

[0119] Specifically, the cruise guidance direction processing path is selected based on the cruise state division result output by the compensation control gating parameters. When the cruise state division result is a stable state, the cruise guidance direction is directly written into the compensation cruise guidance direction and then into the cruise guidance direction storage area. When the cruise state division result is a compensated state, the most recently written cruise guidance direction is taken from the cruise guidance direction storage area as the comparison direction, and the difference between the comparison direction and the cruise guidance direction is taken as the direction change.

[0120] When the change in direction exceeds the cruise stability index, the cruise guidance direction is corrected to the cruise guidance direction corresponding to the cruise stability index to obtain the limited cruise guidance direction; the limited cruise guidance direction and the comparison direction are gradually approximated and updated, and the gradually changing cruise guidance direction is output; the consistency check between the gradually changing cruise guidance direction and the direction of the initial segment of the cruise trajectory reference line is performed, and if the consistency check fails, the gradually changing cruise guidance direction is replaced with the direction corresponding to the smooth heading in the trajectory control reference quantity to generate the compensated cruise guidance direction.

[0121] S5.3: The curvature constraint velocity mapping method is used to convert the compensated cruise guidance direction into linear velocity and angular velocity commands, thereby generating cruise control commands.

[0122] Specifically, the smoothed heading sequence in the trajectory control reference quantity is read, and the grid size used in the inspection environment spatial map is read. The heading difference between adjacent items in the smoothed heading sequence is calculated to obtain the heading change sequence. The ratio of each item in the heading change sequence to the grid size is calculated to obtain the curvature sequence, which is used as the basis for curvature constraints.

[0123] The difference between the compensated cruise guidance direction and the first term of the smooth heading sequence is used as the initial heading difference. The differences between the initial heading difference and the maximum, median, and minimum values ​​of the curvature sequence are sorted from smallest to largest. When the first value in the sorted sequence corresponds to the minimum value, it is determined as the first curvature constraint level; when the first value corresponds to the median value, it is determined as the second curvature constraint level; and when the first value corresponds to the maximum value, it is determined as the third curvature constraint level. The curvature constraint levels are then determined.

[0124] The linear velocity command corresponding to the curvature constraint setting takes the linear velocity value corresponding to the minimum value of the curvature sequence, the linear velocity value corresponding to the median value of the curvature sequence, or the linear velocity value corresponding to the maximum value of the curvature sequence. The angular velocity command corresponding to the curvature constraint setting takes the angular velocity value corresponding to the minimum value of the curvature sequence, the angular velocity value corresponding to the median value of the curvature sequence, or the angular velocity value corresponding to the maximum value of the curvature sequence. The linear velocity command and angular velocity command are written into the command table in the order of the control cycle and summarized to generate the cruise control command.

[0125] It should be noted that the curvature constraint velocity mapping method uses the smoothed heading sequence of the trajectory control reference quantity to calculate the heading change sequence and combines it with the grid size used in the inspection environment spatial map to obtain the curvature sequence. Then, based on the curvature sequence, the curvature constraint gear is determined and the corresponding linear velocity command and angular velocity command are output to form the cruise control command.

[0126] like Figure 5The comparison results of the change in cruise guidance direction angle over time under narrow passage conditions are shown. The curves correspond to comparison group A - single-path guidance, comparison group B - traditional potential field, and experimental group C - the method of this invention, respectively. In comparison group A - single-path guidance, the target guidance direction is calculated based solely on the inspection target point and the robot's current position in each control cycle, and the cruise guidance direction is directly output. This is prone to direction jumps when encountering passage boundaries or risk disturbances. In comparison group B - traditional potential field, the cruise guidance direction is synthesized from the target attraction direction and the obstacle / risk repulsion direction in each cycle. This is prone to high-frequency jitter when the potential field gradient changes rapidly in narrow passages. In experimental group C - the method of this invention, local risk perception data is extracted first in each cycle, the risk change trend is decomposed into multiple candidate risk gradient directions to generate a risk gradient direction set, and then a safe direction set is selected by the direction risk score. Combined with the inspection target point execution direction combination and path continuity constraints, a continuous cruise guidance input is generated, thereby suppressing jitter. The red dashed rectangle in the figure marks the magnified window, and the red double-headed arrow indicates the magnified range. The magnified view below provides a detailed comparison of the differences within the magnified window: the local fluctuation peaks corresponding to characteristic peak 1 and characteristic peak 2 are marked to compare the peak amplitudes of each method. The point of greatest difference is marked with a red dashed line and arrow, indicating that there are differences in the positions of different methods at the same time. This can intuitively show that the peak value of the method of this invention in experimental group C is lower and there are fewer spikes in the key disturbance section, proving that the directional jitter is effectively suppressed and the cruise stability is improved.

[0127] like Figure 6 The results show a comparison of the statistical distribution of radians under narrow channel conditions, used to quantify the overall level of directional jitter and high-jitter tail events. The box plot represents the main distribution range of radians, with the upper and lower bands indicating the fluctuation range. The mean radian value in the figure is used to characterize the overall jitter level, and the 95th percentile of radians is used to characterize the intensity of high-jitter tail events (a higher value indicates more frequent large directional jumps). In comparison group A—single-path guidance lacks a candidate risk gradient direction and safety direction set screening mechanism, easily leading to repeated directional corrections when encountering narrow channel boundary constraints, resulting in an expanded radian distribution range. In comparison group B—traditional potential fields are easily affected by the left-right oscillation of the repulsive force direction within the channel, producing high-radian events, manifested as a longer upper band and a higher 95th percentile. The experimental group C - The method of this invention transforms risk information into executable and continuous cruise guidance input through the link of "risk gradient direction set → direction risk score → safe direction set → direction combination and path continuity constraint". This makes the arc distribution more concentrated, the box lower, and the 95th percentile lower, indicating that large-scale direction jump events are reduced. From a statistical perspective, this directly proves the beneficial effects of avoiding direction jitter, improving cruise stability and mission completion consistency.

[0128] This embodiment also provides a computer device applicable to the cruise control method of an inspection robot, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the cruise control method of the inspection robot as proposed in the above embodiment.

[0129] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0130] This embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements the cruise control method for the inspection robot as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0131] In summary, this invention extracts local risk perception data, decomposes risk change trends into multiple candidate gradient directions, and filters a set of safe directions through directional risk scoring; it combines inspection target points to perform directional combination and path continuity constraints, thereby directly transforming risk information into executable and continuous cruise guidance input; it effectively avoids the local lock-up and directional jitter that are prone to occur in traditional potential field and single-path guidance, and improves cruise stability and mission completion consistency in complex industrial environments.

[0132] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A cruise control method for an inspection robot, characterized in that, include: Collect inspection environment information and equipment operation status data, perform unified coordinate mapping and spatial modeling processing, and generate an inspection environment spatial map; Based on the spatial map of the inspection environment and equipment operation status data, an inspection environment risk field model is constructed by adopting the equipment operation risk layer, the spatial obstacle constraint layer and the inspection target guidance layer. Extract the robot's current position from the inspection environment spatial map, combine it with the inspection environment risk field model to perform risk gradient direction decomposition, and perform cruise direction reconstruction to generate cruise guidance direction; The patrol trajectory of the inspection robot is determined based on the patrol guidance direction, and the motion stability is comprehensively evaluated using the risk coupling stability assessment method to obtain the patrol stability index. Based on the cruise stability index, motion control compensation is performed on the cruise guidance direction to generate cruise control commands.

2. The cruise control method for the inspection robot as described in claim 1, characterized in that, The inspection environment information includes equipment location data and obstacle location data.

3. The cruise control method for the inspection robot as described in claim 2, characterized in that, The inspection environment spatial map includes the location of equipment nodes, spatial distribution information of obstacles, location of inspection target points, and the current location of the robot. The process of collecting inspection environment information and equipment operating status data, performing unified coordinate mapping and spatial modeling, and generating an inspection environment spatial map involves the following steps: Perform unified coordinate mapping on equipment location data and obstacle location data to generate unified coordinate environment data; Based on unified coordinate environment data, spatial structure modeling is performed using a spatial raster modeling algorithm to generate a spatial map of the inspection environment.

4. The cruise control method for the inspection robot as described in claim 1, characterized in that, Based on the inspection environment spatial map and equipment operation status data, an inspection environment risk field model is constructed using an equipment operation risk layer, a spatial obstacle constraint layer, and an inspection target guidance layer. The specific steps are as follows: The equipment operation risk layer is based on equipment operation status data and equipment node location. It uses an equipment status risk mapping algorithm to spatially label the equipment operation status data and generate an equipment risk distribution matrix. The spatial obstacle constraint layer is based on the spatial distribution information of obstacles. It uses a spatial safety distance expansion algorithm to perform geometric expansion of the area around the obstacles and generates a spatial passability constraint matrix. The inspection target guidance layer calculates the target guidance vector for the robot to point towards the inspection target based on the location of the inspection target point and the robot's current position; The risk field model of the inspection environment is constructed by superimposing the equipment operation risk layer, the spatial obstacle constraint layer, and the inspection target guidance layer.

5. The cruise control method for the inspection robot as described in claim 4, characterized in that, The steps for extracting the robot's current position from the inspection environment spatial map and performing risk gradient direction decomposition based on the inspection environment risk field model are as follows: Extract the robot's current position from the inspection environment spatial map, input the inspection environment risk field model, and determine the corresponding local risk perception data; Based on local risk perception data, the risk change trend of the robot's current position is calculated, and the risk change trend is decomposed into multiple candidate risk gradient directions to generate a set of risk gradient directions.

6. The cruise control method for the inspection robot as described in claim 1, characterized in that, The specific steps for generating the cruise guidance direction are as follows: Based on the risk gradient direction set, a risk risk score is calculated using a risk direction screening algorithm, and a set of safe directions is selected. Based on the location of the inspection target point, the set of safe directions is combined to generate the cruise guidance direction.

7. The cruise control method for the inspection robot as described in claim 6, characterized in that, The process involves determining the inspection robot's cruise trajectory based on the cruise guidance direction and using a risk-coupled stability assessment method to comprehensively evaluate the degree of motion stability and obtain cruise stability indicators. The specific steps are as follows: Based on the cruise guidance direction, point-by-point extrapolation is performed on the inspection environment spatial map, and navigability is verified by combining the spatial navigability constraint matrix to form a cruise trajectory reference line; The target heading is obtained by performing trajectory tangent analysis on the cruise trajectory reference line, and then smoothed by curvature constraint to generate trajectory control reference quantity; The cruise trajectory reference line is spatiotemporally aligned with the trajectory control reference quantity, and a risk coupling stability assessment is performed to obtain the cruise stability index.

8. The cruise control method for the inspection robot as described in claim 1, characterized in that, The specific steps for performing motion control compensation on the cruise guidance direction based on the cruise stability index and generating cruise control commands are as follows: Cruise stability indicators are classified into cruise states, and hysteresis switching is determined to obtain compensation control gating parameters. Based on the compensation control gating parameters, the cruise guidance direction is subjected to amplitude limiting, gradual change and continuity constraints to generate a compensated cruise guidance direction. The curvature-constrained velocity mapping method is used to convert the compensated cruise guidance direction into linear velocity and angular velocity commands, thereby generating cruise control commands.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the cruise control method for the inspection robot according to any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the cruise control method for the inspection robot according to any one of claims 1 to 8.

Images