A method for offline calibration of autonomous driving vehicle sensors based on commercial vehicles

Abstract

The present invention belongs to the field of automotive electronics technology, and in particular relates to a method for offline calibration of autonomous driving vehicle sensors based on commercial vehicles, comprising the following steps: positioning and aligning different vehicles, the main manufacturer establishing a database of front wheelbases of different vehicles through the VMS system, the wheel parts are assembled together to form a block, and the front wheelbase mark is assigned to the front wheelbase; the target automatically moves according to different vehicle models; different sensors are calibrated with each other; and the position accuracy of the target is ensured. The present invention has the function of positioning and aligning vehicles with different wheelbases, front overhang lengths, vehicle lengths, vehicle widths, and vehicle heights, and has the function of moving the targets of the camera, front-view integrated machine, laser radar, forward millimeter-wave radar, and side millimeter-wave radar according to the different wheelbases, front overhang lengths, vehicle lengths, vehicle widths, and vehicle heights, and has the function of mutual calibration of the camera, front-view integrated machine, forward millimeter-wave radar, and side millimeter-wave radar.

Description

Technical Field

[0001] The present invention belongs to the field of automotive electronics technology, and in particular relates to an off-line calibration method for autonomous driving vehicle sensors based on commercial vehicles. Background Art

[0002] Traditional vehicles equipped with assisted driving systems such as AEBS, LDWS, and BSD only require calibration of two sensors: cameras and millimeter-wave radars. Autonomous vehicles are equipped with a front-facing integrated camera, surround-view cameras, forward-facing millimeter-wave radars, side millimeter-wave radars, lidars, and inertial navigation systems. These systems include assisted driving functions, as well as autonomous driving features such as automatic obstacle avoidance. These systems utilize a greater number of sensors and place higher demands on sensor calibration. Mass production of autonomous vehicles requires calibration of more than six sensor types, particularly lidars. High vehicle positioning accuracy and target repeatability are required. Otherwise, with so many sensors, data cannot be aligned during fusion.

[0003] Chinese invention patent CN 112834239 B discloses a calibration scheme for assisted driving systems (AEBS) and BSD. This scheme uses a vehicle alignment system to construct a calibration coordinate system to calibrate the front-view camera and millimeter-wave radar sensors.

[0004] Chinese invention patent CN 114488048 A discloses a calibration scheme for forward millimeter-wave radar, which uses a calibration plate, absorbing materials, a calibration computer, and a robotic arm to complete the calibration of the lidar.

[0005] Both of the above solutions have the following disadvantages:

[0006] 1) The solution involves the calibration of two sensors and cannot cover the calibration of multiple sensors;

[0007] 2) The calibration between sensors is isolated, and mutual calibration and verification between multiple sensors is not achieved;

[0008] 3) No solution is given on how to solve the target repeatability problem.

[0009] To this end, we propose an off-line calibration method for autonomous driving vehicle sensors based on commercial vehicles to solve the above problems. Summary of the Invention

[0010] The purpose of the present invention is to address the above-mentioned problems and provide a method for offline calibration of autonomous driving vehicle sensors based on commercial vehicles.

[0011] To achieve the above objectives, the present invention provides the following technical solution: a method for offline calibration of autonomous driving vehicle sensors based on commercial vehicles, comprising the following steps:

[0012] Step S1. Positioning and straightening the vehicle;

[0013] Step S2. Moving the sensor target to the target position based on the sensor installation parameters of the vehicle;

[0014] Step S3. Mutual calibration between different sensor components, including:

[0015] The ADCU (domain controller) obtains the lateral distance and angle information of the LiDAR target; the ADCU obtains the longitudinal distance of the LiDAR target and transmits the longitudinal distance to the LiDAR processor;

[0016] The LiDAR scans the LiDAR target and transmits the acquired point cloud information to the LiDAR processor to obtain the X, Y, and Z coordinate position information of the LiDAR target;

[0017] The LiDAR processor integrates the target position information of different sensor components and combines it with the actual position information of the vehicle on the straightener to complete the calibration of sensor components including the LiDAR.

[0018] Step S4. Obtain analog data of the motor winding current of the sensor target through an AD converter, convert it into corresponding decimal data and then enter the current loop PID algorithm of the sensor target to improve the position accuracy of the sensor target.

[0019] In the above-mentioned offline calibration method for autonomous driving vehicle sensors based on commercial vehicles, the steps of positioning and aligning the vehicle in step S1 include:

[0020] The OEM uses the VMS system (vehicle and parts data management system) to establish a database of front wheel track values ​​for different vehicles, organizes wheel components into wheel blocks, and attaches front wheel track markers assigned with front wheel track values ​​to the wheel blocks, so that the wheel blocks carry the front wheel track values.

[0021] The wheel assembly block is connected to the vehicle data so that the vehicle data includes the parameter value of the front wheel track. The host computer accesses the VMS system through the network to obtain the parameter value of the front wheel track, and inputs the parameter value of the front wheel track into the alignment control system.

[0022] The straightening control system calculates the movement stroke of the lead screw of the straightening device of the straightening control system according to the parameter value of the front wheel track, and straightens the vehicle based on the calculation result.

[0023] In the above-mentioned offline calibration method for autonomous driving vehicle sensors based on commercial vehicles, the step of moving the sensor target to the target position based on the vehicle's sensor installation parameters in step S2 includes:

[0024] The OEM uses the VMS system to establish an installation database for different vehicle cameras and sensor components, including millimeter-wave radars and lidars, in the vehicle coordinate system. The sensor components are grouped into sensor blocks, and sensor component tags assigned with installation parameters are attached to the sensor blocks, so that the sensor blocks carry the values ​​of the sensor component installation parameters.

[0025] The sensor assembly block is linked to the vehicle data so that the vehicle data has the value of the installation parameter. The host computer accesses the VMS system through the network to obtain the value of the installation parameter, and inputs the value of the installation parameter of the sensor part into the alignment control system.

[0026] The alignment control system calculates the movement stroke of the lead screw of the alignment device of the alignment control system according to the value of the installation parameter, and moves the sensor target to the target position based on the calculation result.

[0027] In the above-mentioned offline calibration method for autonomous driving vehicle sensors based on commercial vehicles, before obtaining analog data of the motor winding current of the sensor target through the AD converter in step S4, the method further includes:

[0028] Based on the sensor installation parameters, the controller calculates the distance the sensor target needs to move; combined with the transfer function formula of the servo control system, the controller converts the position and speed information of the sensor target into servo signals and transmits them to the servo control system, forming a position closed loop and a speed closed loop for the sensor target.

[0029] In the above-mentioned offline calibration method for autonomous driving vehicle sensors based on commercial vehicles, the calculation formula for the motion stroke of the lead screw of the straightener is as follows:

[0030] PB≥VmaxX103X60 / NR / i

[0031] In the above formula, PB is the screw lead, mm; Vmax is the maximum operating speed of the mechanism, m / s; NR is the rated speed of the motor, r / min; and i is the transmission ratio from the motor to the screw.

[0032] In the above-mentioned offline calibration method for autonomous driving vehicle sensors based on commercial vehicles, step S3 also includes obtaining the height information of the lidar target through the VMS and transmitting it to the ADCU.

[0033] In the above-mentioned offline calibration method for autonomous driving vehicle sensors based on commercial vehicles, in step S3, the step of the ADCU obtaining the lateral distance and angle information of the lidar target includes:

[0034] The camera is used to obtain image information of the LiDAR target at the target location, and the image information of the LiDAR target is preprocessed to extract image features, classify and match the image features to complete the recognition of the LiDAR target; the camera outputs the lateral distance and angle information of the LiDAR target to the ADCU.

[0035] In the above-mentioned offline calibration method for an autonomous driving vehicle sensor based on a commercial vehicle, in step S3, the step of the ADCU obtaining the longitudinal distance of the lidar target includes:

[0036] A linear frequency modulation pulse is generated by the synthesizer. Based on the linear frequency modulation pulse, the millimeter wave radar outputs the longitudinal distance of the lidar target to the ADCU.

[0037] In the above-mentioned offline calibration method for autonomous driving vehicle sensors based on commercial vehicles, in step S3, the target position information of the different sensor parts includes: the lidar target position information output by the camera, the lidar target position information output by the forward millimeter-wave radar, the lidar target position information output by the forward-looking integrated machine, and the lidar target information obtained by the lidar point cloud.

[0038] In the above-mentioned offline calibration method for autonomous driving vehicle sensors based on commercial vehicles, the point cloud information in step S3 includes the coordinate information of the laser radar target corner points and the RCS reflection intensity.

[0039] In the above-mentioned offline calibration method of autonomous driving vehicle sensors based on commercial vehicles,.

[0040] Compared with the existing technology, the present invention provides a method for offline calibration of autonomous driving vehicle sensors based on commercial vehicles, which has the following beneficial effects:

[0041] 1. This offline calibration method for autonomous driving vehicle sensors based on commercial vehicles achieves vehicle positioning and alignment by having the vehicle enter the calibration site and drive to the alignment station. It can position and align vehicles with different wheelbases, front overhang lengths, vehicle lengths, vehicle widths, and vehicle heights.

[0042] 2. The offline calibration method of the autonomous driving vehicle sensor based on commercial vehicles establishes an installation database of cameras, millimeter-wave radars, and lidars of different vehicles in the vehicle coordinate system through the VMS system. The sensor parts are gathered together to form a block, and the sensor flag is assigned the installation parameter. The sensor flag is hung into the previously formed block so that the block has the value of the installation parameter. The block is hung into the whole vehicle data, and the whole vehicle data has the value of the installation parameter. The upper computer at the calibration site accesses the VMS system through the network to obtain the sensor installation parameters, and inputs the sensor installation parameter values ​​into the target control system. The target calculates the movement stroke of the screw according to the wheelbase, and the target moves to the corresponding position. The camera, front-view integrated machine, lidar, forward millimeter-wave radar, and side millimeter-wave radar targets have the function of moving the target according to the different wheelbases, front overhang lengths, vehicle lengths, vehicle widths, and vehicle heights of the vehicle.

[0043] 3. This offline calibration method for autonomous driving vehicle sensors based on commercial vehicles uses a lidar processor to fuse the lidar target position information output by the camera, the lidar target position information output by the forward millimeter-wave radar, the lidar target position information output by the front-view integrated device, and the lidar target information obtained by the lidar point cloud. It combines the actual position information of the vehicle on the straightener to complete the lidar calibration, and has the function of mutual calibration of the camera, forward-view integrated device, forward millimeter-wave radar, and side millimeter-wave radar.

[0044] To sum up: the present invention has the function of positioning and straightening vehicles with different wheelbases, front overhang lengths, vehicle lengths, vehicle widths, and vehicle heights. It has the function of moving the camera, front-view integrated machine, laser radar, forward millimeter-wave radar, and side millimeter-wave radar targets according to the different wheelbases, front overhang lengths, vehicle lengths, vehicle widths, and vehicle heights of the vehicles. It has the function of calibrating and mutually verifying the camera, front-view integrated machine, forward millimeter-wave radar, and side millimeter-wave radar. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 This is a flow chart of a method for offline calibration of autonomous driving vehicle sensors based on commercial vehicles proposed in the present invention;

[0046] Figure 2 This is a flowchart of an embodiment of positioning and aligning different vehicles in an offline calibration method for autonomous driving vehicle sensors based on commercial vehicles proposed by the present invention;

[0047] Figure 3 This is a flowchart of an embodiment of the automatic movement of targets according to different vehicle models in an offline calibration method for autonomous driving vehicle sensors based on commercial vehicles proposed by the present invention;

[0048] Figure 4This is a schematic diagram of the mutual calibration and verification process between different sensors in the offline calibration method of autonomous driving vehicle sensors based on commercial vehicles proposed by the present invention;

[0049] Figure 5 This is a flowchart of the target repeatability accuracy in the offline calibration method for autonomous driving vehicle sensors based on commercial vehicles proposed by the present invention. DETAILED DESCRIPTION

[0050] The following examples are for illustrative purposes only and are not intended to limit the scope of the present invention.

[0051] See also Figure 1 , a commercial vehicle-based autonomous driving vehicle sensor offline calibration method, comprising:

[0052] Step S1. Positioning and aligning the vehicle.

[0053] The specific steps of step S1 include:

[0054] The OEM establishes a database of front wheel track values ​​for different vehicles through the VMS system, groups wheel parts into wheel blocks, and attaches a front wheel track marker assigned with the front wheel track value to the wheel blocks, so that the wheel blocks carry the front wheel track value.

[0055] The wheel assembly block is connected to the vehicle data so that the vehicle data includes the parameter value of the front wheel track. The host computer accesses the VMS system through the network to obtain the parameter value of the front wheel track, and inputs the parameter value of the front wheel track into the alignment control system.

[0056] The straightening control system calculates the movement stroke of the lead screw of the straightening device of the straightening control system according to the parameter value of the front wheel track, and straightens the vehicle based on the calculation result.

[0057] Step S2: Based on the sensor installation parameters of the vehicle, move the sensor target to the target position.

[0058] Step S2 specifically includes:

[0059] The OEM uses the VMS system to establish an installation database for different vehicle cameras and sensor components, including millimeter-wave radars and lidars, in the vehicle coordinate system. The sensor components are grouped into sensor blocks, and sensor component tags assigned with installation parameters are attached to the sensor blocks, so that the sensor blocks carry the values ​​of the sensor component installation parameters.

[0060] The sensor assembly block is linked to the vehicle data so that the vehicle data has the value of the installation parameter. The host computer accesses the VMS system through the network to obtain the value of the installation parameter, and inputs the value of the installation parameter of the sensor part into the alignment control system.

[0061] The alignment control system calculates the movement stroke of the lead screw of the alignment device of the alignment control system according to the value of the installation parameter, and moves the sensor target to the target position based on the calculation result.

[0062] Step S3. Mutual calibration between different sensor components, including:

[0063] ADCU obtains the lateral distance and angle information of the LiDAR target; ADCU obtains the longitudinal distance of the LiDAR target and transmits the longitudinal distance to the LiDAR processor;

[0064] The LiDAR scans the LiDAR target and transmits the acquired point cloud information to the LiDAR processor to obtain the X, Y, and Z coordinate position information of the LiDAR target;

[0065] The lidar processor fuses the target position information of different sensor parts and combines it with the actual position information of the vehicle on the straightener to complete the calibration of sensor parts including the lidar.

[0066] The step S3 also includes obtaining the height information of the laser radar target through the VMS and transmitting it to the ADCU.

[0067] In step S3, the step of the ADCU obtaining the lateral distance and angle information of the laser radar target includes:

[0068] The camera is used to obtain image information of the LiDAR target at the target location, and the image information of the LiDAR target is preprocessed to extract image features, classify and match the image features to complete the recognition of the LiDAR target; the camera outputs the lateral distance and angle information of the LiDAR target to the ADCU.

[0069] In step S3, the step of the ADCU obtaining the longitudinal distance of the laser radar target includes:

[0070] A linear frequency modulation pulse is generated by the synthesizer. Based on the linear frequency modulation pulse, the millimeter wave radar outputs the longitudinal distance of the lidar target to the ADCU.

[0071] In step S3, the target position information of the different sensor parts includes: the laser radar target position information output by the camera, the laser radar target position information output by the forward millimeter-wave radar, the laser radar target position information output by the forward-looking integrated machine, and the laser radar target information obtained by the laser radar point cloud.

[0072] The point cloud information in step S3 includes the laser radar target corner coordinate information and RCS reflection intensity.

[0073] Step S4. Obtain analog data of the motor winding current of the sensor target through an AD converter, convert it into corresponding decimal data and then enter the current loop PID algorithm of the sensor target to improve the position accuracy of the sensor target.

[0074] Before the step S4 of acquiring analog data of the motor winding current of the sensor target through the AD converter, the method further includes:

[0075] Based on the sensor installation parameters, the controller calculates the distance the sensor target needs to move; combined with the transfer function formula of the servo control system, the controller converts the position and speed information of the sensor target into servo signals and transmits them to the servo control system, forming a position closed loop and a speed closed loop for the sensor target.

[0076] The calculation formula for the motion stroke of the screw of the straightener is as follows:

[0077] PB≥VmaxX103X60 / NR / i

[0078] In the above formula, PB is the screw lead, mm; Vmax is the maximum operating speed of the mechanism, m / s; NR is the rated speed of the motor, r / min; and i is the transmission ratio from the motor to the screw.

[0079] Specific combination of Figure 2-5 As shown in the figure, the offline calibration method of autonomous driving vehicle sensors based on commercial vehicles includes the following steps:

[0080] Step S1. Positioning and alignment of different vehicles

[0081] VMS (vehicle and parts data management system) establishes the vehicle front axle wheelbase data of 2000mm, VMS establishes the vehicle front axle wheelbase data of 3000mm, VMS establishes the vehicle second and third axle distance data of 1000mm, and assigns the data to the front wheel track mark PD001, the first and second axle distance mark PD002, and the second and third axle distance mark PD003 respectively, and hangs the assigned marks into the 10-digit wheel code block 4600A-L0000;

[0082] The wheel assembly 4600A-L0000 consists of the 12-digit tire code 4600000-L000, the 12-digit rim code 4600000-L001, and the 12-digit spoke code 4600000-L0002. This ensures the integrity of all parts in the vehicle system and forms a minimum set of parts. This ensures that the mark value can be transferred to complete the calibration, the vehicle can be selected from different configurations, and the organization of the vehicle parts can be guaranteed.

[0083] The component data (including component drawings, component quantity, and standard component assembly) of the vehicle module, as well as the module's logo and logo value information, are obtained. The vehicle's front axle wheelbase data (2000mm), vehicle front axle wheelbase data (3000mm), and vehicle second and third axle distance data (1000mm) are transmitted to the calibration controller via the factory's scanner. The calibration controller assigns the pulse equivalent variable a value of δ = 10. At this time, the number of movement turns of the screw is n = 20, and the stepper motor step angle α = 360.

[0084] Use Formula 1 and Formula 2 to calculate the lead of the screw drive. The worktable moving with the target moves a distance D = 100. The horizontal extension distance of the straightening device is 2000 mm. One axis of the straightening device is stationary, and the V-shaped block of the second axis moves to the position x = -3000. One axis of the straightening device is stationary, and the V-shaped block of the third axis moves to the position x = -4000. At this time, the straightening device is in place.

[0085] Step S2. The target automatically moves according to different vehicle models

[0086] VMS establishes the laser radar installation height data of 3000mm, the laser radar installation offset data of 0mm, and the laser radar installation depth data of 1200mm. The data are assigned to the laser radar installation height mark AT001, the laser radar installation offset mark AT002, and the laser radar installation depth mark AT003 respectively, and the three marks are hung in the laser radar block 10-bit code 5600A-L0000;

[0087] The laser radar assembly consists of the laser radar parts with 12-digit code 5600000-L0002, the laser radar bracket parts with 12-digit code 5600000-L0001, and the laser radar cleaning device parts with 12-digit code 5600000-L0000;

[0088] This ensures the integrity of all vehicle system components, forming a minimum set of components. This ensures that the flag values ​​can be transferred to complete calibration, allowing the vehicle to select different configurations, while also ensuring the organization of the vehicle components. Since the perception system of autonomous vehicles is basically installed in the cab, in order to facilitate unified management and calibration of sensors, the LiDAR module 5600A-L0000 is hung in the cab with a flag, and is managed on a unified platform inside the cab.

[0089] The vehicle acquires the parts data (including part drawings, part quantities, and standard parts), assembly markers, and marker value information from the cab. The factory's scanner transmits the LiDAR height data (3000mm), LiDAR installation offset (0mm), and LiDAR installation depth (1200mm) to the calibration controller. The calibration controller assigns the pulse equivalent variable a value of δ = 5. Based on Formulas 1 and 2, the number of turns of the lead screw is calculated to be n = 10, the stepper motor step angle α = 180, the translation distance D = 50 for the workbench carrying the target, LiDAR target 1 moves 1000mm, and LiDAR target 2 moves 1000mm. At this point, the LiDAR targets are in place.

[0090] Step S3. Mutual calibration between different sensors

[0091] The camera acquires the image information of the LiDAR target, and then pre-processes the image information of four LiDAR targets with a size of 1000mm x 800mm, extracts image features, and classifies the image features. The camera then matches the LiDAR target images and finally completes the LiDAR target recognition. Finally, the camera outputs the lateral distance and angle information of the LiDAR target to the ADCU (domain controller);

[0092] The ADCU transparently transmits the above information to the LiDAR processor. The on-site calibration equipment obtains the LiDAR target height information of 2000mm through the VMS. Because the LiDAR target height information does not change with the movement of the target, it can be obtained through the system.

[0093] The on-site calibration device transmits the LiDAR target height information (2000mm) to the ADCU through the OBD port, and the ADCU transparently transmits this information to the LiDAR processor.

[0094] The millimeter-wave radar synthesizer generates a linear frequency modulation pulse, which is transmitted by the transmitting antenna (TX antenna). The reflection of the linear frequency modulation pulse by the lidar target generates a reflected linear frequency modulation pulse captured by the receiving antenna (RX antenna). The millimeter-wave radar outputs the longitudinal distance of the lidar target (9000mm) to the ADCU through the transceiver signal. The ADCU transmits the longitudinal distance of the lidar target to the lidar processor.

[0095] The LiDAR scans the LiDAR target and transmits the point cloud information to the LiDAR processor to obtain the X, Y, and Z coordinate information of the LiDAR target;

[0096] The LiDAR processor integrates the LiDAR target position information output by the camera, the forward millimeter-wave radar, the forward-looking integrated device, and the LiDAR target information obtained from the LiDAR point cloud, and completes the LiDAR calibration by combining it with the actual position information of the vehicle on the straightener.

[0097] Step S4. Ensure the position accuracy of the target

[0098] The controller obtains the sensor installation location information through VMS, calculates the coordinates of the sensor in the current calibration coordinate system, and calculates the distance the sensor needs to move, 1000mm, based on the sensor's initial coordinates and the coordinates to which it needs to move.

[0099] The sensor controller inputs the position information of the target to be moved to the target controller. According to the transfer function formula of the servo control system, the target controller converts the position information and speed information into servo signals and transmits them to the servo system, forming the position closed loop and speed closed loop of the sensor target.

[0100] The analog data of the sensor target motor winding current is obtained by the AD converter, converted into corresponding decimal data and then enters the current loop PID algorithm of the sensor target. The current loop mainly controls the rotational torque of the servo motor of the sensor target and is the most basic guarantee for the servo motor's response speed to the command pulse. The addition of the current loop can improve the overall closed-loop rigidity of the sensor target servo system and improve the position accuracy of the target.

[0101] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for offline calibration of autonomous driving vehicle sensors based on commercial vehicles, characterized by: The steps include: Step S1. Positioning and straightening the vehicle; Step S2. Moving the sensor target to the target position based on the sensor installation parameters of the vehicle; Step S3. Mutual calibration between different sensor components, including: The ADCU obtains the lateral distance and angle information of the LiDAR target through the camera; the ADCU obtains the longitudinal distance of the LiDAR target through the millimeter-wave radar and transmits the longitudinal distance to the LiDAR processor; The LiDAR scans the LiDAR target and transmits the acquired point cloud information to the LiDAR processor to obtain the X, Y, and Z coordinate position information of the LiDAR target; The LiDAR processor integrates the target position information of different sensor components and combines it with the actual position information of the vehicle on the straightener to complete the calibration of sensor components including the LiDAR. Step S4. Obtain analog data of the motor winding current of the sensor target through an AD converter, convert it into corresponding decimal data, and then enter the current loop PID algorithm of the sensor target to improve the position accuracy of the sensor target; The steps of positioning and straightening the vehicle in step S1 include: The OEM establishes a database of front wheel track values ​​for different vehicles through the VMS system, groups wheel parts into wheel blocks, and attaches a front wheel track marker assigned with the front wheel track value to the wheel blocks, so that the wheel blocks carry the front wheel track value. The wheel assembly block is connected to the vehicle data so that the vehicle data includes the parameter value of the front wheel track. The host computer accesses the VMS system through the network to obtain the parameter value of the front wheel track, and inputs the parameter value of the front wheel track into the alignment control system. The straightening control system calculates the movement stroke of the lead screw of the straightening device of the straightening control system according to the parameter value of the front wheel track, and straightens the vehicle based on the calculation result.

2. The offline calibration method for autonomous driving vehicle sensors based on commercial vehicles according to claim 1, characterized in that: The step S2 of moving the sensor target to the target position based on the sensor installation parameters of the vehicle includes: The OEM uses the VMS system to establish an installation database for different vehicle cameras and sensor components, including millimeter-wave radars and lidars, in the vehicle coordinate system. The sensor components are grouped into sensor blocks, and sensor component tags assigned with installation parameters are attached to the sensor blocks, so that the sensor blocks carry the values ​​of the sensor component installation parameters. The sensor assembly block is linked to the vehicle data so that the vehicle data has the value of the installation parameter. The host computer accesses the VMS system through the network to obtain the value of the installation parameter, and inputs the value of the installation parameter of the sensor part into the alignment control system. The alignment control system calculates the movement stroke of the lead screw of the alignment device of the alignment control system according to the value of the installation parameter, and moves the sensor target to the target position based on the calculation result.

3. The offline calibration method for autonomous driving vehicle sensors based on commercial vehicles according to claim 1, characterized in that: Before obtaining the analog data of the motor winding current of the sensor target through the AD converter in step S4, the method further includes: Based on the sensor installation parameters, the controller calculates the distance the sensor target needs to move; combined with the transfer function formula of the servo control system, the controller converts the position and speed information of the sensor target into servo signals and transmits them to the servo control system, forming a position closed loop and a speed closed loop for the sensor target.

4. The method for offline calibration of sensors for autonomous driving vehicles based on commercial vehicles according to claim 1 or 2, characterized in that: The calculation formula for the motion stroke of the screw of the straightener is as follows: PB≥VmaxX103X60 / NR / i In the above formula, PB is the screw lead, mm; Vmax is the maximum operating speed of the mechanism, m / s; NR is the rated speed of the motor, r / min; and i is the transmission ratio from the motor to the screw.

5. The offline calibration method for autonomous driving vehicle sensors based on commercial vehicles according to claim 1, characterized in that: The step S3 also includes obtaining the height information of the laser radar target through the VMS and transmitting it to the ADCU.

6. The offline calibration method for autonomous driving vehicle sensors based on commercial vehicles according to claim 1, characterized in that: In step S3, the step of the ADCU obtaining the lateral distance and angle information of the laser radar target includes: The camera is used to obtain image information of the LiDAR target at the target location, and the image information of the LiDAR target is preprocessed to extract image features, classify and match the image features to complete the recognition of the LiDAR target; the camera outputs the lateral distance and angle information of the LiDAR target to the ADCU.

7. The offline calibration method for autonomous driving vehicle sensors based on commercial vehicles according to claim 1, characterized in that: In step S3, the step of the ADCU obtaining the longitudinal distance of the laser radar target includes: A linear frequency modulation pulse is generated by the synthesizer. Based on the linear frequency modulation pulse, the millimeter wave radar outputs the longitudinal distance of the lidar target to the ADCU.

8. The offline calibration method for autonomous driving vehicle sensors based on commercial vehicles according to claim 1, characterized in that: In step S3, the target position information of the different sensor parts includes: the laser radar target position information output by the camera, the laser radar target position information output by the forward millimeter-wave radar, the laser radar target position information output by the forward-looking integrated machine, and the laser radar target information obtained by the laser radar point cloud.

9. The offline calibration method for autonomous driving vehicle sensors based on commercial vehicles according to claim 1, characterized in that: The point cloud information in step S3 includes the laser radar target corner coordinate information and RCS reflection intensity.

Images