A vehicle-mounted radar terrain panoramic recognition system

Abstract

The application provides a vehicle-mounted radar terrain panoramic recognition system, which comprises a plurality of sensor arrays arranged in a gooseneck area, a bridge area and a center area of a vehicle bottom, wherein the plurality of sensor arrays comprise a plurality of laser radars and a plurality of imaging radars; a terrain modeling module is used for fusing depth data of the laser radars with texture features of the imaging radars, generating a terrain data matrix containing elevation gradients and ground hardness, and constructing a 3D terrain model according to the terrain data matrix; and a dynamic leveling controller is used for generating a set of leg coordination action instructions according to the terrain data matrix generated by the terrain modeling module. The application solves the problem that the vehicle-mounted radar cannot recognize a roadbed when the vehicle-mounted radar is urgently erected in a complex temporary position, thereby causing a control lag, and also solves the problem that an operator cannot panoramically understand the erection roadbed state.

Description

Technical Field

[0001] This invention relates to the field of vehicle-mounted radar automatic leveling technology, specifically to a vehicle-mounted radar terrain panoramic recognition system. Background Technology

[0002] Vehicle-mounted radar automatic leveling systems are widely used in the leveling control of radar equipment's outriggers. They not only provide stable engineering support and a horizontal reference, but are also the prerequisite and foundation for completing various radar functions. Leveling outriggers are generally installed on the vehicle, and depending on the vehicle model, they are typically arranged with four main leveling outriggers; four auxiliary leveling outriggers (optional), the selection of which depends on the vehicle's span; and two anti-overturning outriggers (optional), the selection of which depends on wind load, array size, and the equipment's center of gravity. Currently, vehicle-mounted radar automatic leveling systems have evolved to the fourth generation.

[0003] The first-generation radar's leveling system was a single-point hand-cranked leveling system. It mainly relied on the readings of a bubble level to manually crank each outrigger to complete the leveling of each point on the radar. This type of outrigger was not electrically driven. Setup typically required 8 people for 2 hours, achieving an accuracy of 1 degree, with a single leg travel of 500mm.

[0004] The second-generation radar's leveling system is a single-point electric leveling system. It uses a small electric device mounted on a hand-cranked shaft to perform electric leveling. Because this leveling system uses the values ​​of a dual-axis digital display level as the accuracy standard, and there is no software interconnection between the outriggers, each outrigger can only be automatically leveled independently. Setup typically requires 5 people for 1 hour, achieving an accuracy of 20 minutes, with a single outrigger travel of 500mm.

[0005] The third-generation radar's leveling system is a multi-point automatic leveling system. It relies on a tilt sensor to determine the highest point using positional errors, and then uses an automatic leveling controller to process multi-point data to achieve height-tracking leveling. However, this leveling method is prone to failure to level within the effective travel range when the radar installation site has a large slope or is twisted. Installation typically requires two people and 20 minutes, achieving an accuracy of 6 minutes, with a single-leg travel of 500mm.

[0006] The fourth-generation radar's leveling system is a multi-point intelligent leveling system. This method requires two high-precision dual-axis tilt sensors to provide real-time feedback on the platform's tilt angles. Then, an intelligent algorithm adjusts the extension and retraction of the outriggers to control the platform's tilt angle until both the X and Y tilt angles are 0 or within the leveling accuracy range. At this point, the platform is considered leveled. Due to limitations in the placement of the tilt sensors, this leveling method can only reflect the specific angles at the intersection points of the outriggers, and the leveling display screen can only provide accuracy information in the X and Y dimensions. Setup typically takes one person 3 minutes, achieving an accuracy within 3 minutes, with a single outrigger travel of 500mm.

[0007] In summary, the existing fourth-generation vehicle-mounted radar automatic leveling technology has shortcomings such as poor terrain adaptability, low leveling efficiency, and lack of panoramic perception. Summary of the Invention

[0008] To address the technical problems existing in the background technology, this invention proposes a vehicle-mounted radar terrain panoramic recognition system.

[0009] It should be noted that the vehicle mentioned in this application is a vehicle equipped with radar, including semi-trailers.

[0010] This invention proposes a vehicle-mounted radar terrain panoramic recognition system, comprising:

[0011] A variety of sensor arrays are arranged in the gooseneck area, axle area and center area under the vehicle, the variety of sensor arrays including multiple lidar and multiple imaging radar;

[0012] Terrain Modeling Module: This module fuses the depth data from LiDAR with the texture features from imaging radar to generate a terrain data matrix that includes elevation gradient and surface hardness, and then constructs a 3D terrain model based on the terrain data matrix.

[0013] Dynamic leveling controller: Generates a set of outrigger coordinated action instructions based on the terrain data matrix generated by the terrain modeling module.

[0014] Preferably, it also includes a three-dimensional panoramic display module: used to integrate the three-dimensional model of the radar vehicle platform with the 3D terrain model generated by the terrain modeling module in the same coordinate system to output a three-dimensional panoramic view including the projection area of ​​the outriggers.

[0015] Preferably, the 3D panoramic display module includes a control panel, a display screen, and control buttons on the control panel.

[0016] Preferably, the dynamic leveling controller calculates the optimal movement trajectory of each leg in real time based on the terrain matrix. When it detects that the suspension distance on one side is greater than the preset value, it triggers the priority leg extension mode and reduces the speed of the opposite leg.

[0017] Preferably, the field of view of each lidar has an overlapping area.

[0018] Preferably, the field of view of each imaging radar has an overlapping area.

[0019] Preferably, there are 12 lidar units, of which: 2 are installed on both sides of the front end of the gooseneck area at the bottom of the vehicle, 2 are installed on both sides of the rear end of the gooseneck area at the bottom of the vehicle, 2 are installed on both sides of the front end of the axle area at the bottom of the vehicle, and the remaining 4 are installed around the center of the bottom of the vehicle; and 6 imaging radar units are provided, of which 1 is centrally located in the center of the gooseneck area at the bottom of the vehicle, 1 is installed in the center of the axle area, and the remaining 4 are installed around the area between the rear end of the gooseneck area and the front end of the axle area at the bottom of the vehicle.

[0020] This invention utilizes multiple sets of visual sensors (imaging radar) and distance sensors (LiDAR) deployed at the bottom of the vehicle to form a terrain data matrix below the vehicle's boundaries. A terrain modeling module then constructs a 3D terrain model from this data matrix, providing a comprehensive view of the terrain beneath the vehicle. Finally, a dynamic leveling controller generates a set of coordinated outrigger movement commands based on the terrain data matrix generated by the modeling module. This not only reduces leveling time and errors but also allows users to perceive the deployment status in real time, achieving high-quality vision and ease of operation. Deployment can be completed in one minute by one person, with an accuracy within 30 arcseconds and a single-leg travel of 500mm. This solves the problem of vehicle-mounted radar failing to identify the roadbed during emergency deployment in complex temporary terrain, causing control lag, and also addresses the difficulty for operators to have a panoramic view of the roadbed deployment status. Attached Figure Description

[0021] Fig. 1 This is a schematic diagram showing the location distribution of various areas of the vehicle in a vehicle-mounted radar terrain panoramic recognition system proposed in this invention;

[0022] Fig. 2 This is a map showing the positional distribution of 12 lidars and 6 imaging radars at the bottom of a vehicle in a vehicle-mounted radar terrain panoramic recognition system proposed in this invention.

[0023] Fig. 3 This is a schematic diagram of the structure of the three-dimensional panoramic display module in a vehicle-mounted radar terrain panoramic recognition system proposed in this invention. Detailed Implementation

[0024] Reference Figs. 1-3 The present invention proposes a vehicle-mounted radar terrain panoramic recognition system, comprising: 12 lidar units and 6 imaging radar units, a terrain modeling module, a dynamic leveling controller, a 3D panoramic display module, and a dynamic leveling controller, wherein:

[0025] Twelve lidar units (numbered J1-J12) and six imaging radars (numbered C1-C6) are deployed in the gooseneck area Q1, axle area Q2, and center area Q3 under the vehicle. Specifically: lidars J1 and J12 are installed on both sides of the front end of the gooseneck under the vehicle; lidars J2 and J11 are installed on both sides of the rear end of the gooseneck under the vehicle; lidars J5 and J8 are installed on both sides of the front end of the axle under the vehicle; and lidars J6 and J7 are installed on both sides of the rear end of the vehicle. LiDAR J3, LiDAR J4, LiDAR J9, and LiDAR J10 are installed around the center of the vehicle's underside. Imaging radar C6 is centrally located in the center of the gooseneck area Q1 on the vehicle's underside. Imaging radar C3 is installed in the center of the axle area Q2. Imaging radars C1, C2, C4, and C5 are installed around the area between the rear end of the gooseneck area Q1 and the front end of the axle area Q2 on the vehicle's underside, and are located around the periphery of LiDARs J3, J4, J9, and J10. All-terrain coverage is achieved through the coordinated data acquisition of 12 LiDARs and 6 imaging radars.

[0026] The terrain modeling module is used to fuse the depth data of LiDAR with the texture features of imaging radar to generate a terrain data matrix that includes elevation gradient and surface hardness, and to build a 3D terrain model based on the terrain data matrix.

[0027] The 3D panoramic display module is used to integrate the 3D model of the radar vehicle platform with the 3D terrain model generated by the terrain modeling module in the same coordinate system to output a 3D panoramic view including the projection area of ​​the outriggers. Specifically, the 3D panoramic display module includes a control panel 1, a display screen 2, and control buttons 3 set on the control panel 1.

[0028] The dynamic leveling controller generates a set of outrigger coordinated action instructions based on the terrain data matrix generated by the terrain modeling module. Specifically, the dynamic leveling controller calculates the optimal motion trajectory of each outrigger in real time based on the terrain matrix. When it detects that the suspension distance on one side is greater than the preset value, it triggers the priority leg extension mode and reduces the speed of the opposite outrigger so that both outriggers can contact the ground synchronously.

[0029] In a further embodiment, the dynamic leveling controller executes the following decision logic:

[0030] When a Class III soft foundation is identified (echo intensity < 50dB), the outriggers are controlled to extend in a three-segment motion curve with a stepped speed increase.

[0031] When the elevation difference between adjacent outriggers is detected to be greater than the preset value, the anti-overturning compensation algorithm is automatically activated.

[0032] The specific working steps of this system are as follows:

[0033] Step 1: System Startup and Initialization

[0034] Once the vehicle enters the position and the support blocks are in place, the lidars J1 to J12 and the imaging radars C1 to C6 are synchronously powered on and initialized.

[0035] Step 2: Topographic Data Acquisition and Processing

[0036] LiDARs J1 to J12 calculate the depth values ​​of each point in the terrain; imaging radars C1 to C6 respectively acquire high-resolution scene images of the gooseneck area Q1 (imaging radar C6), the bridge area Q2 (imaging radar C3), and the center area Q3 (imaging radars C1 / C2 / C4 / C5).

[0037] Step 3: 3D Terrain Model Construction

[0038] The terrain modeling module uses the built-in ICP algorithm to fuse the depth data of the lidar with the texture features of the imaging radar, generating a terrain data matrix that includes elevation gradient and surface hardness. Based on the terrain data matrix, a 3D terrain subsidence feature model is generated, and then a 3D terrain model is established.

[0039] Step 4: 3D panoramic display:

[0040] The 3D panoramic display module integrates the 3D model of the radar vehicle platform with the 3D terrain model generated by the terrain modeling module in the same coordinate system to output a 3D panoramic view including the projection area of ​​the outriggers.

[0041] Step 5: Dynamic Leveling

[0042] The dynamic leveling controller calculates the optimal movement trajectory of each outrigger in real time based on the terrain matrix. When it detects that the suspension distance on one side is greater than the preset value, it triggers the priority leg extension mode and reduces the speed of the opposite outrigger so that both outriggers can contact the ground synchronously.

[0043] In a further embodiment, the mounting axis of the lidar J1 / J12 in the gooseneck area Q1 forms an angle of 32°-38° with the longitudinal axis of the vehicle, the overlap area of ​​the field of view of each lidar is ≥15%, and the overlap area of ​​the field of view of each imaging radar is ≥15%.

[0044] As can be seen from the above, this invention forms a terrain data matrix under the vehicle boundary by arranging multiple sets of visual sensors (imaging radar C1-C6) and distance sensors (lidar J1-J12) at the bottom of the vehicle body. A 3D terrain model is then constructed using a terrain modeling module to provide feedback on the overall terrain below the vehicle body. Finally, the dynamic leveling controller generates a set of coordinated outrigger action commands based on the terrain data matrix generated by the terrain modeling module. This not only reduces leveling time and minimizes leveling errors but also allows users to perceive the erection status in real time, achieving high-quality vision and ease of operation. Erecting can be completed in one minute by one person, with an accuracy within 30 arcseconds and a single-leg travel of 500mm. It solves the problem of vehicle-mounted radar being unable to identify the roadbed during emergency erection in complex temporary terrain conditions, causing control lag. It also addresses the difficulty for operators to have a panoramic view of the roadbed status during erection. Compared to existing technologies, the outstanding advantages of this invention are:

[0045] 1. Enhanced the adaptability of vehicle-mounted radar in mountainous, plain, and Class III hard ground conditions;

[0046] 2. The roadbed images generated by the identification system can help operators grasp the terrain status of the vehicle in real time. When the radar accuracy requirement is relatively high, it can be fine-tuned in real time according to environmental changes.

[0047] 3. Feedback on the terrain condition of the roadbed to the identification system can improve the leveling efficiency during erection. When some leveling legs are far from the ground, some legs can extend faster to ensure that multiple leveling legs contact the ground at the same time, shorten the erection cycle, and improve the mobility of the vehicle.

[0048] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. An on-board radar terrain panoramic recognition system, characterized in that, include: A variety of sensor arrays are arranged in the gooseneck area, axle area and center area under the vehicle, including multiple lidar and multiple imaging radars; Specifically: 12 lidar units are installed, including 2 installed on both sides of the front end of the gooseneck area at the bottom of the vehicle, 2 installed on both sides of the rear end of the gooseneck area at the bottom of the vehicle, 2 installed on both sides of the front end of the axle at the bottom of the vehicle, 2 installed on both sides of the rear end of the bottom of the vehicle, and the remaining 4 installed around the center of the bottom of the vehicle; 6 imaging radar units are installed, including 1 centrally located in the center of the gooseneck area at the bottom of the vehicle, 1 installed in the center of the axle area, and the remaining 4 installed around the area between the rear end of the gooseneck area and the front end of the axle area at the bottom of the vehicle. Terrain Modeling Module: This module fuses the depth data from LiDAR with the texture features from imaging radar to generate a terrain data matrix that includes elevation gradient and surface hardness, and then constructs a 3D terrain model based on the terrain data matrix. Dynamic leveling controller: Generates a set of outrigger coordinated action instructions based on the terrain data matrix generated by the terrain modeling module.

2. The vehicle-mounted radar terrain panoramic recognition system according to claim 1, characterized in that, It also includes a 3D panoramic display module: used to integrate the 3D model of the radar vehicle platform with the 3D terrain model generated by the terrain modeling module in the same coordinate system to output a 3D panoramic view including the projection area of ​​the outriggers.

3. The vehicle-mounted radar terrain panoramic recognition system of claim 2, wherein, The 3D panoramic display module includes a control panel, a display screen, and control buttons on the control panel.

4. The vehicle-mounted radar terrain panoramic recognition system of claim 1, wherein, The dynamic leveling controller calculates the optimal movement trajectory of each outrigger in real time based on the terrain matrix. When it detects that the suspension distance on one side is greater than the preset value, it triggers the priority leg extension mode and reduces the speed of the opposite outrigger.

5. The vehicle-mounted radar terrain panoramic recognition system of claim 1, wherein, The fields of view of each lidar have overlapping areas.

6. The vehicular radar terrain panoramic recognition system of claim 1, wherein, The fields of view of each imaging radar have overlapping areas.

Images