Intelligent control method and system for highway inspection

Abstract

The application is suitable for the technical field of highway inspection, and provides an intelligent control method and system for highway inspection, which comprises the following steps: when it is detected that there is an obstacle in front, a dynamic perception blind area blocked by the obstacle is calculated; a vehicle bypassing track for safely bypassing the obstacle is planned, a virtual sweeping track starting from the beginning of the dynamic perception blind area and extending along the direction of the lane is derived on the road surface of the originally planned inspection lane, and a spatial mapping relationship between the vehicle bypassing track and the virtual sweeping track at each time point is established; according to the spatial mapping relationship, a control instruction is generated in real time to drive the bearing cloud platform of the main inspection sensor to move, so that the observation focus of the main inspection sensor continuously tracks the virtual sweeping track. By constructing a virtual sweeping track which is always anchored on the originally planned inspection lane, the application realizes continuous scanning of the target lane and fundamentally eliminates data discontinuity.

Description

Technical Field

[0001] This invention relates to the field of highway inspection technology, specifically to an intelligent control method and system for highway inspection. Background Technology

[0002] With the development of intelligent transportation and highway maintenance management technologies, the use of autonomous or assisted driving vehicles equipped with various sensors (such as high-definition cameras, LiDAR, and line array scanners) for automated highway inspection has become an important trend to improve inspection efficiency and ensure operational safety.

[0003] In real-world open highway environments, temporary and sudden static obstacles frequently appear along inspection routes, such as broken-down vehicles, road maintenance work areas, and scattered goods. When inspection vehicles encounter such obstacles, they must perform a detour to ensure driving safety. In existing solutions, the vehicle's path planning and control module replans a safe driving trajectory to avoid the obstacle based on obstacle information input from the perception system, and the vehicle follows this new trajectory. While this process solves the vehicle's own traffic safety problem, it introduces a critical inspection operational issue: during the detour, the observation field of the vehicle's main inspection sensor (usually fixed to the road surface directly in front of or below the vehicle) inevitably deviates from the originally planned inspection lane, causing an interruption in the scanning of a section of the road surface (especially areas obscured by obstacles and their adjacent areas), creating a perception gap. This severely affects the continuity and completeness of inspection data and the accuracy of subsequent defect analysis. Therefore, an intelligent control method and system for highway inspection is needed to solve the above problems. Summary of the Invention

[0004] In view of the shortcomings of the existing technology, the purpose of this invention is to provide an intelligent control method and system for highway inspection, so as to solve the problems existing in the above-mentioned background technology.

[0005] This invention is implemented as follows: an intelligent control method for highway inspection, the method comprising the following steps: Perform obstacle detection on the inspection path. When an obstacle is detected ahead, calculate the dynamic perception blind spot that is blocked by the obstacle. Plan a vehicle detour trajectory that allows the inspection vehicle to safely bypass the obstacle. On the road surface of the original planned inspection lane, derive a virtual sweep trajectory that starts from the starting point in front of the dynamic perception blind spot and extends along the lane direction. Establish the spatial mapping relationship between the vehicle detour trajectory and the virtual sweep trajectory at each time point. Based on the spatial mapping relationship, control commands are generated in real time to drive the gimbal of the main inspection sensor to move, so that the observation focus of the main inspection sensor continuously tracks the virtual sweep trajectory. Activate the side assist sensor of the inspection vehicle and set the observation task of the side assist sensor to focus on the originally planned inspection lane; During the vehicle's detour, data collected by the main inspection sensor and the lateral auxiliary sensor are combined to generate continuous road surface perception information.

[0006] As a further aspect of the present invention, the step of calculating the dynamic perception blind spot obstructed by obstacles specifically includes: Acquire the three-dimensional contour information of the obstacle and the real-time position of the obstacle, wherein the three-dimensional contour information includes at least length, width and height; Call the inherent parameter model of the main inspection sensor, including the sensor's installation height, default pitch angle, horizontal field of view, and vertical field of view; Based on the inherent parameter model, calculate the visual cone emanating from the optical center of the sensor that covers the originally planned inspection lane area; project the three-dimensional contour onto the visual cone, and calculate the invalid area where the visual cone is occluded by obstacles. This invalid area is defined as the dynamic perception blind spot. Output the geometric parameters of the dynamic perception blind spot, including the longitudinal distance from the starting point of the dynamic perception blind spot to the vehicle, the longitudinal length, and the lateral width.

[0007] As a further aspect of the present invention, the step of deriving the virtual sweep trajectory specifically includes: On the center line of the originally planned inspection lane, determine an initial point at a preset safe distance in front of the starting point of the dynamic perception blind spot, and set the expected movement speed of a virtual sweep point. Starting from the initial point, along the lane centerline, the movement path of the virtual sweep point over time is derived at the desired speed. This path extends to cover the entire dynamic perception blind spot and continues backward for a set distance. The motion path is discretized into a sequence of virtual sweep points arranged in a time series, forming a virtual sweep trajectory.

[0008] As a further aspect of the present invention, the step of establishing the spatial mapping relationship between the vehicle's detour trajectory and the virtual sweep trajectory at various time points specifically includes: Synchronize the vehicle detour trajectory and the virtual sweep trajectory in time so that for the same time index t, there exists a vehicle position point Pv(t) and a virtual sweep point Ps(t). Calculate the three-dimensional relative vector Δ(t) from the vehicle position point Pv(t) to the corresponding virtual sweep point Ps(t) at each time point; Record the set of relative vectors {Δ(t)} under all time indices. This set fully defines the spatial mapping relationship from vehicle trajectory points to virtual sweep points.

[0009] As a further aspect of the present invention, the step of generating control commands in real time based on the spatial mapping relationship specifically includes: Obtain the three-dimensional relative vector Δ(t') corresponding to the current time t'; based on the fixed installation matrix of the main inspection sensor on the vehicle, transform the relative vector Δ(t') to the sensor base coordinate system; Based on the transformed vector, the target yaw angle and target pitch angle required to drive the gimbal are solved to make the sensor optical axis point to the virtual sweep point; The target yaw angle and target pitch angle, along with the current actual angle feedback of the gimbal, are input into the gimbal servo controller to generate motor drive control commands.

[0010] As a further aspect of the present invention, the step of activating the lateral auxiliary sensor of the inspection vehicle specifically includes: Determine the detour direction: if the vehicle plans to detour from the left side of the obstacle, select the right side lateral assist sensor; if detour from the right side, select the left side sensor. Control the rotatable mechanism of the selected lateral auxiliary sensor to adjust its horizontal observation angle toward the originally planned inspection lane area.

[0011] Another object of the present invention is to provide an intelligent control system for highway inspection, the system comprising: The dynamic perception blind spot module is used to detect obstacles on the inspection path. When an obstacle is detected ahead, the dynamic perception blind spot occluded by the obstacle is calculated. The virtual sweep trajectory module is used to plan a vehicle detour trajectory that allows the inspection vehicle to safely bypass the obstacle. On the road surface of the originally planned inspection lane, a virtual sweep trajectory is derived, starting from the starting point in front of the dynamic perception blind spot and extending along the lane direction, and the spatial mapping relationship between the vehicle detour trajectory and the virtual sweep trajectory at each time point is established. The control command generation module is used to generate control commands in real time according to the spatial mapping relationship, drive the gimbal of the main inspection sensor to move, and make the observation focus of the main inspection sensor continuously track the virtual sweep trajectory. The auxiliary sensor control module is used to activate the side auxiliary sensors of the inspection vehicle and set the observation task of the side auxiliary sensors to focus on the originally planned inspection lane. The perception information generation module is used to integrate the data collected by the main inspection sensor and the lateral auxiliary sensor during the vehicle's detour to generate continuous road perception information.

[0012] Compared with the prior art, the beneficial effects of the present invention are: By constructing a virtual sweep trajectory independent of the vehicle's detour path but always anchored on the original planned inspection lane, and driving the main sensor gimbal to track this trajectory in real time, continuous scanning of the target lane, especially the area obscured by obstacles, is achieved, fundamentally eliminating data gaps and ensuring the integrity of inspection data.

[0013] By activating the lateral auxiliary sensor, it temporarily focuses on the originally planned inspection lane during the detour phase, forming an intersecting view with the main sensor, thus expanding the perception coverage and compensating for the potential degradation in edge data quality caused by the main sensor's viewing angle limitations. Attached Figure Description

[0014] Figure 1 A flowchart of an intelligent control method for highway inspection; Figure 2 A flowchart for determining dynamic perception blind spots in an intelligent control method for highway inspection; Figure 3 A flowchart for deriving a virtual sweep trajectory in an intelligent control method for highway inspection; Figure 4 A flowchart illustrating the establishment of spatial mapping relationships in an intelligent control method for highway inspection; Figure 5 A flowchart illustrating the generation of control commands in an intelligent control method for highway inspection; Figure 6 This is a flowchart illustrating the activation of a lateral auxiliary sensor in an intelligent control method for highway inspection. Figure 7 This is a schematic diagram of the structure of an intelligent control system for highway inspection. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0016] The specific implementation of the present invention will be described in detail below with reference to specific embodiments.

[0017] like Figure 1 As shown in the figure, this embodiment of the invention provides an intelligent control method for highway inspection, the method comprising the following steps: S100 performs obstacle detection on the inspection path. When an obstacle is detected ahead, it calculates the dynamic perception blind spot that is blocked by the obstacle. S200: Plan a vehicle detour trajectory that allows the inspection vehicle to safely bypass the obstacle. On the road surface of the original planned inspection lane, derive a virtual sweep trajectory and establish a spatial mapping relationship between the vehicle detour trajectory and the virtual sweep trajectory at each time point. S300, based on the spatial mapping relationship, generates control commands in real time to drive the movement of the gimbal supporting the main inspection sensor; S400, activate the side assist sensor of the inspection vehicle and set the observation task of the side assist sensor to focus on the originally planned inspection lane; S500, during the vehicle's detour, integrates the data collected by the main inspection sensor and the lateral auxiliary sensor to generate continuous road perception information.

[0018] In this embodiment of the invention, obstacles on the inspection path are first detected using LiDAR / camera. When an obstacle obstructing the vehicle's path is detected, a dynamic perception blind spot is calculated in real time based on the obstacle's three-dimensional dimensions and position. Parameters of the dynamic perception blind spot include the blind spot's starting point (distance from the vehicle's position), its length, and its lateral range within the originally planned inspection lane. Then, a detour trajectory is planned to allow the inspection vehicle to safely bypass the obstacle. Specifically, a detour strategy is determined based on the location of the dynamic perception blind spot and environmental perception, including changing lanes to the left or right of the obstacle. Next, a series of discrete path points are generated within adjacent lanes that comply with traffic rules. These path points are connected to form a detour reference path. Combining the inspection vehicle's dynamic constraints, including the minimum turning radius and maximum lateral acceleration, the detour reference path is smoothly optimized to generate a continuous and trackable vehicle detour trajectory. Finally, a timestamp or desired speed is assigned to each path point of the vehicle detour trajectory, forming a spatiotemporal trajectory.

[0019] In this embodiment of the invention, a virtual sweep trajectory is derived on the road surface of the originally planned inspection lane, starting from the starting point in front of the dynamic perception blind spot and extending along the lane direction. A spatial mapping relationship between the vehicle's detour trajectory and the virtual sweep trajectory is established at each time point. This embodiment calculates the dynamic perception blind spot caused by the relative motion between obstacles and vehicles in real time, using this as the starting point and key coverage area for generating the virtual sweep trajectory. This makes the allocation of sensing resources precise and adaptive, proactively responding to changes in occlusion rather than passively accepting data loss. Then, based on the spatial mapping relationship, control commands are generated in real time to drive the movement of the main inspection sensor's mounting platform, ensuring that the main inspection sensor's observation focus continuously tracks the virtual sweep trajectory. In this way, a virtual sweeping trajectory is constructed that is independent of the vehicle's detour trajectory but is always anchored on the original planned inspection lane. The main sensor gimbal is driven to track this trajectory in real time. This ensures that the main sensor's line of sight is locked on the target lane during the entire detour process when the vehicle deviates laterally. This achieves near-continuous scanning of the target lane, especially the area obscured by obstacles, fundamentally eliminating data gaps and ensuring the integrity of the inspection data.

[0020] In this embodiment of the invention, the lateral auxiliary sensors (lateral lidar or high-definition camera) of the inspection vehicle are also activated. These lateral auxiliary sensors typically remain dormant or perform low-priority tasks (such as observing road shoulders) during normal cruise. Here, the observation task of the lateral auxiliary sensors is set to focus on the originally planned inspection lane. By strategically activating and redirecting the lateral auxiliary sensors originally used for other tasks (such as observing road shoulders or adjacent lanes), they are made to temporarily focus on the originally planned inspection lane during the detour phase. Their scanning angle is adjusted so that their beam or line of sight can cross the lane lines, covering the edge area of ​​the originally planned inspection lane, forming an intersecting view with the main sensor. In this way, without increasing hardware costs, the sensing coverage is expanded, data redundancy and reliability are improved, and in particular, the degradation in edge data quality that may occur due to the viewing angle limitations of the main sensor is compensated for.

[0021] Finally, during the vehicle's detour, the data collected by the main inspection sensor and the lateral auxiliary sensor are fused to generate continuous road surface perception information. Specifically, image stitching or point cloud registration algorithms are used to spatially fuse and stitch the main sensor data and the lateral auxiliary sensor data to generate a seamless and continuous road surface condition information map covering the originally planned inspection lane. This embodiment of the invention enables inspection vehicles to continuously perform core inspection tasks while adhering to safety regulations and completing obstacle avoidance maneuvers normally. This significantly improves the practicality and operational value of the intelligent inspection system in complex road conditions, making full-section, no-missed inspection possible.

[0022] like Figure 2As shown, in a preferred embodiment of the present invention, the step of calculating the dynamic perception blind spot obstructed by obstacles specifically includes: S101, Obtain the three-dimensional contour information of the obstacle and the real-time position of the obstacle, wherein the three-dimensional contour information includes at least length, width and height; S102, call the inherent parameter model of the main inspection sensor, including the sensor's installation height, default pitch angle, horizontal field of view and vertical field of view; S103, based on the inherent parameter model, calculate the visual cone emanating from the optical center of the sensor that covers the originally planned inspection lane area; project the three-dimensional contour into the visual cone, and calculate the invalid area where the visual cone is occluded by obstacles, which is defined as the dynamic perception blind zone. S104, output the geometric parameters of the dynamic perception blind spot, including the longitudinal distance from the starting point of the dynamic perception blind spot to the vehicle, the longitudinal length, and the lateral width.

[0023] In this embodiment of the invention, the three-dimensional contour information and real-time position of the obstacle are determined by LiDAR / camera. Then, the inherent parameter model of the main inspection sensor is called, and based on the inherent parameter model, a viewing cone emanating from the optical center of the sensor and covering the originally planned inspection lane area is calculated. The three-dimensional contour is projected onto the viewing cone, and the invalid area where the viewing cone is occluded by the obstacle is calculated. This invalid area is the dynamic perception blind spot.

[0024] like Figure 3 As shown, in a preferred embodiment of the present invention, the step of deriving the virtual sweep trajectory specifically includes: S201, on the center line of the originally planned inspection lane, determine an initial point located at a preset safe distance in front of the starting point of the dynamic perception blind spot; S202, Starting from the initial point, along the lane centerline direction, derive the motion path of the virtual sweep point as time changes with the desired motion speed; S203, the motion path is discretized into a sequence of virtual sweep points arranged in a time series to form a virtual sweep trajectory.

[0025] In this embodiment of the invention, an initial point is determined on the centerline of the originally planned inspection lane, located at a preset safe distance ahead of the starting point of the dynamic perception blind spot, and a desired movement speed is set. Starting from this initial point, the virtual sweep point moves along the lane centerline at the desired speed, deriving a movement path over time that covers the entire dynamic perception blind spot and extends backward for the preset distance. Finally, the movement path is discretized into a sequence of virtual sweep points arranged in a time series, thus obtaining the virtual sweep trajectory.

[0026] like Figure 4 As shown, in a preferred embodiment of the present invention, the step of establishing the spatial mapping relationship between the vehicle detour trajectory and the virtual sweep trajectory at various time points specifically includes: S204, synchronize the vehicle detour trajectory and the virtual sweep trajectory in time, so that for the same time index t, there exists a vehicle position point Pv(t) and a virtual sweep point Ps(t). S205, calculate the three-dimensional relative vector Δ(t) from the vehicle position point Pv(t) to the corresponding virtual sweep point Ps(t) at each time point; S206 records the set of relative vectors {Δ(t)} under all time indices, which fully defines the spatial mapping relationship from vehicle trajectory points to virtual sweep points.

[0027] In this embodiment of the invention, to determine the spatial mapping relationship, it is necessary to synchronize the vehicle's detour trajectory and the virtual sweep trajectory in time, and to establish a one-to-one correspondence between the vehicle position point Pv(t) and the virtual sweep point Ps(t) under each time index t. Then, the three-dimensional relative vector Δ(t) from the vehicle position point Pv(t) to the corresponding virtual sweep point Ps(t) at each time point is calculated. Finally, the three-dimensional relative vectors under all time indices are summarized to obtain all the spatial mapping relationships from the vehicle trajectory points to the virtual sweep points.

[0028] like Figure 5 As shown, in a preferred embodiment of the present invention, the step of generating control commands in real time based on the spatial mapping relationship specifically includes: S301, obtain the three-dimensional relative vector Δ(t') corresponding to the current time t'; based on the fixed installation matrix of the main inspection sensor on the vehicle, transform the relative vector Δ(t') to the sensor base coordinate system; S302, based on the transformed vector, solves inversely the target yaw angle and target pitch angle required to drive the gimbal, so that the sensor optical axis points to the virtual sweep point; S303, the target yaw angle and target pitch angle, along with the current actual angle feedback of the gimbal, are input into the gimbal servo controller to generate motor drive control commands.

[0029] In this embodiment of the invention, control commands for the gimbal are generated in real time. Specifically, the three-dimensional relative vector Δ(t') corresponding to the current time t' is obtained. Then, the fixed installation matrix of the main inspection sensor on the vehicle is retrieved, and the relative vector Δ(t') is transformed into the sensor base coordinate system. Based on the transformed vector, the target yaw angle and target pitch angle required to drive the gimbal are solved to make the sensor optical axis point to the virtual sweep point. For example, the transformed vector is V=[Vx,Vy,Vz]. TIn this context, Vx typically points directly in front of the sensor base (the initial direction of the optical axis), Vy points to the left, and Vz points upwards. The target pitch angle θ is the angle between the sensor's optical axis in the vertical plane and the horizontal plane (XOY plane), calculated using the arctangent function: θ = arctan(Vz / √(Vx²+Vy²)). This result indicates the angle by which the sensor needs to rotate upwards or downwards to point at the target point. The target yaw angle ψ is the angle between the projection vector of the sensor's optical axis in the horizontal plane (XOY plane) and the directly in front of the base (X-axis), calculated using the arctangent function: ψ = arctan(Vy / Vx). This result indicates the angle by which the sensor needs to rotate left or right to point at the target point. Finally, the target yaw angle and target pitch angle, along with the actual angle feedback of the current gimbal, are input into the gimbal servo controller, which automatically generates motor drive control commands.

[0030] like Figure 6 As shown, in a preferred embodiment of the present invention, the step of activating the lateral auxiliary sensor of the inspection vehicle specifically includes: S401, determine the detour direction. If the vehicle plans to detour from the left side of the obstacle, select the right side lateral assist sensor; if detour from the right side, select the left side sensor. S402 controls the rotatable mechanism of the selected lateral auxiliary sensor to adjust its horizontal observation angle toward the originally planned inspection lane area.

[0031] In this embodiment of the invention, before activating the lateral assist sensor, the vehicle's detour direction is determined. If the vehicle plans to detour from the left side of the obstacle, the right-side lateral assist sensor is selected; if it detours from the right side, the left-side sensor is selected. After selecting the lateral assist sensor, its rotatable mechanism is controlled to adjust its horizontal observation angle toward the originally planned inspection lane area, so that it focuses on collecting road surface images or point cloud data to supplement the blind spot information that the main inspection sensor may miss due to changes in viewing angle.

[0032] like Figure 7 As shown in the figure, this embodiment of the invention also provides an intelligent control system for highway inspection, the system comprising: The dynamic perception blind spot module 100 is used to detect obstacles on the inspection path. When an obstacle is detected ahead, the dynamic perception blind spot occluded by the obstacle is calculated. The virtual sweep trajectory module 200 is used to plan a vehicle detour trajectory that enables the inspection vehicle to safely bypass the obstacle. On the road surface of the originally planned inspection lane, a virtual sweep trajectory is derived, starting from the starting point in front of the dynamic perception blind spot and extending along the lane direction, and the spatial mapping relationship between the vehicle detour trajectory and the virtual sweep trajectory at each time point is established. The control command generation module 300 is used to generate control commands in real time according to the spatial mapping relationship, drive the gimbal of the main inspection sensor to move, and make the observation focus of the main inspection sensor continuously track the virtual sweep trajectory. The auxiliary sensor control module 400 is used to activate the side auxiliary sensor of the inspection vehicle and set the observation task of the side auxiliary sensor to focus on the originally planned inspection lane. The perception information generation module 500 is used to generate continuous road perception information by fusing data collected by the main inspection sensor and the lateral auxiliary sensor during the vehicle's detour.

[0033] In a preferred embodiment of the present invention, the dynamic sensing blind spot module 100 includes: A three-dimensional contour information unit is used to acquire the three-dimensional contour information of an obstacle and the real-time position of the obstacle. The three-dimensional contour information includes at least length, width, and height. The parameter model calling unit is used to call the inherent parameter model of the main inspection sensor, including the sensor's installation height, default pitch angle, horizontal field of view, and vertical field of view. The blind spot determination unit is used to calculate, based on the inherent parameter model, a visual cone emanating from the optical center of the sensor that covers the originally planned inspection lane area; project the three-dimensional contour onto the visual cone; and calculate the invalid area where the visual cone is occluded by obstacles, which is defined as the dynamic perception blind spot. The geometric parameter output unit is used to output the geometric parameters of the dynamic perception blind spot, including the longitudinal distance from the starting point of the dynamic perception blind spot to the vehicle, the longitudinal length, and the lateral width.

[0034] In a preferred embodiment of the present invention, the virtual sweep trajectory module 200 includes: The initial point speed determination unit is used to determine an initial point located at a preset safe distance in front of the starting point of the dynamic perception blind zone on the center line of the original planned inspection lane, and to set the expected movement speed of a virtual sweep point. The motion path determination unit is used to derive the motion path of the virtual sweep point over time from the initial point along the lane centerline direction at the desired motion speed. The path extends to cover the entire dynamic perception blind zone and continues backward for a set distance. The virtual sweep trajectory unit is used to discretize the motion path into a sequence of virtual sweep points arranged in a time series, thus forming a virtual sweep trajectory.

[0035] In a preferred embodiment of the present invention, the virtual sweep trajectory module 200 further includes: The trajectory time synchronization unit is used to synchronize the vehicle detour trajectory and the virtual sweep trajectory in time, so that for the same time index t, there exists a vehicle position point Pv(t) and a virtual sweep point Ps(t). A three-dimensional relative vector unit is used to calculate the three-dimensional relative vector Δ(t) from the vehicle position point Pv(t) to the corresponding virtual sweep point Ps(t) at each time point; The vector set determination unit is used to record the relative vector set {Δ(t)} under all time indices. This set completely defines the spatial mapping relationship from vehicle trajectory points to virtual sweep points.

[0036] The above description only details the preferred embodiments of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0037] It should be understood that although the steps in the flowcharts of the various embodiments of the present invention are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the various embodiments may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least a portion of the sub-steps or stages of other steps.

[0038] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0039] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the disclosure in the specification and embodiments. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the claims.

Claims

1. An intelligent control method for highway inspection, characterized in that, The method includes the following steps: Perform obstacle detection on the inspection path. When an obstacle is detected ahead, calculate the dynamic perception blind spot that is blocked by the obstacle. Plan a vehicle detour trajectory that allows the inspection vehicle to safely bypass the obstacle. On the road surface of the original planned inspection lane, derive a virtual sweep trajectory that starts from the starting point in front of the dynamic perception blind spot and extends along the lane direction. Establish the spatial mapping relationship between the vehicle detour trajectory and the virtual sweep trajectory at each time point. Based on the spatial mapping relationship, control commands are generated in real time to drive the gimbal of the main inspection sensor to move, so that the observation focus of the main inspection sensor continuously tracks the virtual sweep trajectory. Activate the side assist sensor of the inspection vehicle and set the observation task of the side assist sensor to focus on the originally planned inspection lane; During the vehicle's detour, data collected by the main inspection sensor and the lateral auxiliary sensor are integrated to generate continuous road surface perception information; The steps for deriving the virtual sweep trajectory specifically include: determining an initial point at a preset safe distance in front of the starting point of the dynamic perception blind spot on the centerline of the originally planned inspection lane; setting a desired movement speed for the virtual sweep point; starting from the initial point, deriving the movement path of the virtual sweep point over time along the lane centerline at the desired movement speed, with the path extending to cover the entire dynamic perception blind spot and continuing backward for a preset distance; and discretizing the movement path into a sequence of virtual sweep points arranged in a time series to form the virtual sweep trajectory. The step of establishing the spatial mapping relationship between the vehicle detour trajectory and the virtual sweep trajectory at each time point specifically includes: synchronizing the vehicle detour trajectory and the virtual sweep trajectory in time, so that for the same time index t, there exists a vehicle position point Pv(t) and a virtual sweep point Ps(t); calculating the three-dimensional relative vector Δ(t) from the vehicle position point Pv(t) to the corresponding virtual sweep point Ps(t) at each time point; and recording the set of relative vectors {Δ(t)} under all time indices, which fully defines the spatial mapping relationship from the vehicle trajectory point to the virtual sweep point.

2. The intelligent control method for highway inspection according to claim 1, characterized in that, The step of calculating the dynamic perception blind spot obstructed by obstacles specifically includes: Acquire the three-dimensional contour information of the obstacle and the real-time position of the obstacle, wherein the three-dimensional contour information includes at least length, width and height; Call the inherent parameter model of the main inspection sensor, including the sensor's installation height, default pitch angle, horizontal field of view, and vertical field of view; Based on the inherent parameter model, calculate the visual cone emanating from the optical center of the sensor that covers the originally planned inspection lane area; project the three-dimensional contour onto the visual cone, and calculate the invalid area where the visual cone is occluded by obstacles. This invalid area is defined as the dynamic perception blind spot. Output the geometric parameters of the dynamic perception blind spot, including the longitudinal distance from the starting point of the dynamic perception blind spot to the vehicle, the longitudinal length, and the lateral width.

3. The intelligent control method for highway inspection according to claim 1, characterized in that, The step of generating control commands in real time based on the spatial mapping relationship specifically includes: Obtain the three-dimensional relative vector Δ(t') corresponding to the current time t'; based on the fixed installation matrix of the main inspection sensor on the vehicle, transform the relative vector Δ(t') to the sensor base coordinate system; Based on the transformed vector, the target yaw angle and target pitch angle required to drive the gimbal are solved to make the sensor optical axis point to the virtual sweep point; The target yaw angle and target pitch angle, along with the current actual angle feedback of the gimbal, are input into the gimbal servo controller to generate motor drive control commands.

4. The intelligent control method for highway inspection according to claim 1, characterized in that, The step of activating the lateral auxiliary sensors of the inspection vehicle specifically includes: Determine the detour direction: if the vehicle plans to detour from the left side of the obstacle, select the right side lateral assist sensor; if detour from the right side, select the left side sensor. Control the rotatable mechanism of the selected lateral auxiliary sensor to adjust its horizontal observation angle toward the originally planned inspection lane area.

5. An intelligent control system for highway inspection, characterized in that, The system includes: The dynamic perception blind spot module is used to detect obstacles on the inspection path. When an obstacle is detected ahead, the dynamic perception blind spot occluded by the obstacle is calculated. The virtual sweep trajectory module is used to plan a vehicle detour trajectory that allows the inspection vehicle to safely bypass the obstacle. On the road surface of the originally planned inspection lane, a virtual sweep trajectory is derived, starting from the starting point in front of the dynamic perception blind spot and extending along the lane direction, and the spatial mapping relationship between the vehicle detour trajectory and the virtual sweep trajectory at each time point is established. The control command generation module is used to generate control commands in real time according to the spatial mapping relationship, drive the gimbal of the main inspection sensor to move, and make the observation focus of the main inspection sensor continuously track the virtual sweep trajectory. The auxiliary sensor control module is used to activate the side auxiliary sensors of the inspection vehicle and set the observation task of the side auxiliary sensors to focus on the originally planned inspection lane. The perception information generation module is used to integrate the data collected by the main inspection sensor and the lateral auxiliary sensor during the vehicle's detour process to generate continuous road perception information. The virtual sweep trajectory module includes: an initial point speed determination unit, used to determine an initial point located at a preset safe distance in front of the starting point of the dynamic perception blind zone on the center line of the originally planned inspection lane, and set a desired movement speed for the virtual sweep point; a movement path determination unit, used to derive the movement path of the virtual sweep point over time along the lane center line direction from the initial point, with the desired movement speed, the path extending to cover the entire dynamic perception blind zone and continuing backward for a preset distance; and a virtual sweep trajectory unit, used to discretize the movement path into a sequence of virtual sweep points arranged in a time sequence, forming a virtual sweep trajectory. The virtual sweep trajectory module further includes: a trajectory time synchronization unit, used to synchronize the vehicle detour trajectory and the virtual sweep trajectory in time, so that for the same time index t, there exists a vehicle position point Pv(t) and a virtual sweep point Ps(t); a three-dimensional relative vector unit, used to calculate the three-dimensional relative vector Δ(t) from the vehicle position point Pv(t) to the corresponding virtual sweep point Ps(t) at each time point; and a vector set determination unit, used to record the relative vector set {Δ(t)} under all time indices, which fully defines the spatial mapping relationship from the vehicle trajectory point to the virtual sweep point.

6. The intelligent control system for highway inspection according to claim 5, characterized in that, The dynamic perception blind spot module includes: A three-dimensional contour information unit is used to acquire the three-dimensional contour information of an obstacle and the real-time position of the obstacle. The three-dimensional contour information includes at least length, width, and height. The parameter model calling unit is used to call the inherent parameter model of the main inspection sensor, including the sensor's installation height, default pitch angle, horizontal field of view, and vertical field of view. The blind spot determination unit is used to calculate, based on the inherent parameter model, a visual cone emanating from the optical center of the sensor that covers the originally planned inspection lane area; project the three-dimensional contour onto the visual cone; and calculate the invalid area where the visual cone is occluded by obstacles, which is defined as the dynamic perception blind spot. The geometric parameter output unit is used to output the geometric parameters of the dynamic perception blind spot, including the longitudinal distance from the starting point of the dynamic perception blind spot to the vehicle, the longitudinal length, and the lateral width.

Images