Nondestructively realized anthropomorphic unmanned robot

Abstract

The invention relates to the technical field of unmanned driving, and provides a lossless-realized anthropomorphic unmanned robot. The lossless-realized anthropomorphic unmanned robot comprises a mounting seat assembly, a control cabinet, and a steering wheel control assembly, a camera assembly, a brake control assembly, an accelerator control assembly and a gear engaging control assembly which are connected with the control cabinet; the steering wheel control assembly comprises a steering wheel left clamping arm and a steering wheel right clamping arm which are installed on the installation base assembly, a steering wheel left support is arranged on the steering wheel left clamping arm, a steering wheel right support is arranged on the steering wheel right clamping arm, and the steering wheel left support and the steering wheel right support are each provided with a steering wheel control motor, a steering wheel steering claw and a camera assembly. The steering wheel control motor drives the steering wheel to rotate by controlling the steering wheel steering claw to rotate; the steering wheel steering claw is used for clamping a steering wheel. The camera assembly is used for simulating a driver to obtain road condition data. The brake control assembly is used for controlling braking; the accelerator control assembly is used for controlling an accelerator, and the gear engaging control assembly is used for controlling gear engaging.

Description

Technical Field

[0001] The present invention relates to the technical field of driverless technology, and in particular, to an anthropomorphic driverless robot implemented without damage. Background Art

[0002] The statements in this part only provide background technical information related to the present invention and do not necessarily constitute prior art.

[0003] With the rapid development of artificial intelligence and autonomous driving technology, driverless cars are gradually moving from the laboratory to reality and becoming an important trend in future transportation. For the existing large number of manned vehicles, how to achieve their conversion to the driverless mode has become an urgent problem to be solved. However, the existing driverless technologies mostly target newly designed driverless vehicles. Upgrading traditional manned vehicles to driverless vehicles usually requires large-scale modification of the vehicle, such as installing a drive-by-wire system, sensors, or even replacing core components, etc., which has problems such as high cost, poor adaptability, and damage to the original vehicle structure, with high cost and potential safety hazards.

[0004] Therefore, the existing upgrading of traditional manned vehicles to driverless vehicles has the following problems: 1. Traditional autonomous driving modification: It is necessary to deeply invade the vehicle's electronic control system (such as cracking the CAN protocol), which has legal risks and cannot be compatible with multiple vehicle models.

[0005] 2. The existing autonomous driving systems are difficult to adapt to diverse vehicle control interfaces (such as different forms of shift mechanisms and pedal types). Summary of the Invention

[0006] To solve at least one of the above technical problems in the background art, the present invention provides an anthropomorphic driverless robot implemented without damage. The anthropomorphic driverless robot designed by the present invention is a driver-like robot that does not require vehicle modification. By simulating the operation behaviors of human drivers, it realizes the lossless unmanned upgrade of manned vehicles.

[0007] To achieve the above object, the present invention adopts the following technical solutions: The first aspect of the present invention provides an anthropomorphic driverless robot implemented without damage, including: A mounting seat assembly, a control cabinet, and a steering wheel control assembly, a camera assembly, a brake control assembly, an accelerator control assembly, and a gear shifting control assembly all connected to the control cabinet; The steering wheel control assembly includes a left steering wheel clamp arm and a right steering wheel clamp arm mounted on the mounting seat assembly. A left steering wheel bracket is provided on the left steering wheel clamp arm, and a right steering wheel bracket is provided on the right steering wheel clamp arm. A steering wheel control motor, a steering wheel steering claw, and a camera assembly are provided on both the left steering wheel bracket and the right steering wheel bracket. The steering wheel control motor drives the steering wheel to rotate by controlling the rotation of the steering wheel steering claw; the steering wheel steering claw is used to clamp the steering wheel, and the camera assembly is used to simulate a driver to obtain road condition data. A brake control assembly is used to control the brakes; an accelerator control assembly is used to control the accelerator, and a gear shifting control assembly is used to control gear shifting.

[0008] Further, a plurality of threaded holes are provided on both the left steering wheel clamp arm and the right steering wheel clamp arm. A screw rod is used to cooperate with the threaded holes to connect the left steering wheel clamp arm and the right steering wheel clamp arm, and they are fastened by a steering wheel clamp arm fastening nut; a steering wheel bracket screw is used to cooperate with the threaded holes to connect the left steering wheel clamp arm and the left steering wheel bracket, and a steering wheel bracket screw is used to cooperate with the threaded holes to connect the right steering wheel clamp arm and the right steering wheel bracket.

[0009] Further, a camera bracket is provided on both the left steering wheel bracket and the right steering wheel bracket, and a camera system is provided on the camera bracket.

[0010] Further, a groove is provided on the left steering wheel bracket, and a left steering wheel slider that slides up and down is provided on the groove; a groove is provided on the right steering wheel bracket, and a right steering wheel slider that slides up and down is provided on the groove; the left steering wheel bracket is connected to the steering wheel control motor base through the left steering wheel slider, the right steering wheel bracket is connected to the steering wheel control motor base through the right steering wheel slider, the steering wheel control motor is installed on the steering wheel control motor base, and the output shaft of the steering wheel control motor is coaxial with the steering wheel during application.

[0011] Further, the mounting seat assembly includes a left base plate and a right base plate. A brake piece base, a steering wheel clamp arm base, and a control piece bottom plate are installed on both the left base plate and the right base plate. Base screw holes are provided on both the left base plate and the right base plate, threaded holes are provided on the brake piece base, the steering wheel clamp arm base, and the control piece bottom plate. The brake piece base is installed on the left base plate and the right base plate through brake piece base fastening screws, the steering wheel clamp arm base is installed on the left base plate and the right base plate through steering wheel clamp arm base fastening screws, and the control piece bottom plate is installed on the left base plate and the right base plate through control piece bottom plate fastening screws.

[0012] Furthermore, the brake control assembly includes a brake bracket and a brake linear motor. The brake bracket is installed on the brake piece base, and the brake linear motor is installed on the brake bracket. The central axis of the brake linear motor abuts against the brake pedal plate.

[0013] Further, the throttle control component includes a throttle bracket and a throttle linear motor. The throttle bracket is installed on the brake part base, and the throttle linear motor is installed on the throttle bracket. The central axis of the throttle linear motor abuts against the throttle pedal plate.

[0014] Further, both the left steering wheel clamping arm and the right steering wheel clamping arm are installed on the steering wheel clamping arm base through the steering wheel clamping arm fastening block.

[0015] Further, the control cabinet is fixed on the control part bottom plate through the cabinet fastening block.

[0016] Further, the gear shifting control component includes a three-axis robotic arm base, a three-axis robotic arm, and a gear shifting socket head. The three-axis robotic arm base is installed on the control part bottom plate. The three-axis robotic arm is installed on the three-axis robotic arm base, and the tail end of the three-axis robotic arm is connected to the gear shifting socket head for gear shifting.

[0017] Further, one steering wheel steering claw includes two steering wheel fastening claws, and the two steering wheel fastening claws are fixed on the steering wheel through fastening nuts.

[0018] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention provides a humanoid driverless robot that can be realized without damage, which can realize the non-destructive transformation of a manned vehicle, ensure that the original structure and functions of the vehicle are not damaged at all, and can flexibly and conveniently switch between the manned driving and driverless modes to meet different scenarios and user needs.

[0019] The present invention provides a humanoid driverless robot that can be realized without damage, which has the ability to accurately receive and deeply process multi-source information from the vehicle itself, road infrastructure, and the cloud, and then make scientific and reasonable driving decisions, comprehensively improving the safety and efficiency of driving.

[0020] The advantages of the additional aspects of the present invention will be partially given in the following description, partially become obvious from the following description, or be understood through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The specification drawings constituting a part of the present invention are used to provide a further understanding of the present invention. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation to the present invention.

[0022] Figure 1 It is a structural diagram of the humanoid driverless robot that can be realized without damage provided by the embodiment of the present invention; Figure 2 It is a structural diagram of the steering wheel steering claw provided by the embodiment of the present invention; Among them, 1. Base screw hole, 2. Right base plate, 3. Left base plate, 4. Three-axis robotic arm base, 5. Control cabinet, 6. Cabinet fastening block, 7. Control component base plate, 8. Steering wheel clamping arm fastening block, 9. Steering wheel clamping arm base, 10. Braking component base, 11. Braking component base fastening screw, 12. Brake pedal plate, 13. Brake linear motor, 14. Brake bracket, 15. Throttle linear motor, 16. Throttle bracket, 17. Left steering wheel clamping arm, 18. Steering wheel bracket screw, 19. Steering wheel clamping arm fastening nut, 20. Right steering wheel clamping arm, 21. Left steering wheel bracket, 22. Steering wheel steering claw, 23. Steering wheel control motor, 24. Left steering wheel slider, 25. Right steering wheel slider, 26. Steering wheel control motor base, 27. Camera bracket, 28. Camera system, 29. Triple-axis robotic arm, 30. Gear shift sleeve head, 31. Steering wheel control motor output shaft, 32. Fastening nut, 33. Steering wheel, 34. Steering wheel fastening claw. Detailed implementation manners

[0023] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0024] It should be noted that the following detailed description is illustrative and is intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs.

[0025] It should be noted that the terms used herein are only for describing specific implementation manners and are not intended to limit the exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular forms are also intended to include the plural forms. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0026] Embodiment 1 As Figure 1 、 Figure 2 shown, this embodiment provides a humanoid driverless robot that realizes lossless implementation, including: A mounting seat assembly, a control cabinet 5, and a steering wheel control assembly, a camera assembly, a brake control assembly, a throttle control assembly, and a gear shift control assembly all connected to the control cabinet 5; The steering wheel control assembly comprises a left steering wheel clamp arm 17 and a right steering wheel clamp arm 20 mounted on the mounting seat assembly, a left steering wheel bracket 21 is provided on the left steering wheel clamp arm 17, a right steering wheel bracket is provided on the right steering wheel clamp arm 20, a steering wheel control motor 23, a steering wheel steering claw 22 and a camera assembly are provided on both the left steering wheel bracket 21 and the right steering wheel bracket, the steering wheel control motor 23 controls the steering wheel steering claw 22 to rotate, so as to drive the steering wheel 33 to rotate; the steering wheel steering claw 22 is used to clamp the steering wheel 33, and the camera assembly is used to simulate the driver to obtain road condition data; The brake control component is used to control the brake; the throttle control component is used to control the throttle, and the gear control component is used to control the gear.

[0027] The steering wheel control component is closely connected to the main body, and through the synergy of its own high-precision locking device and the mechanical arm, it ensures that it always maintains a stable position and state during the working process. The servo motor can achieve precise steering control of the steering wheel, whether in a straight line, a curve, or parking in various complex driving scenarios, it can ensure that the vehicle turns accurately according to the predetermined driving trajectory, effectively improving the vehicle's driving controllability and safety.

[0028] In some embodiments, the left steering wheel clamp arm 17 and the right steering wheel clamp arm 20 are both provided with a plurality of threaded holes, and the left steering wheel clamp arm 17 and the right steering wheel clamp arm 20 are connected by screws in combination with the threaded holes, and are fastened by the steering wheel clamp arm fastening nuts 19; the left steering wheel clamp arm 17 and the left steering wheel bracket 21 are connected by steering wheel bracket screws 18 in combination with the threaded holes, and the right steering wheel clamp arm 20 and the right steering wheel bracket are connected by steering wheel bracket screws 18 in combination with the threaded holes.

[0029] In some embodiments, a camera bracket 27 is provided on the left steering wheel bracket 21 and the right steering wheel bracket, and a camera system is provided on the camera bracket 27 .

[0030] In some embodiments, the left steering wheel bracket 21 is provided with a groove, and a left steering wheel slider 24 that slides up and down is provided on the groove; the right steering wheel bracket is provided with a groove, and a right steering wheel slider 25 that slides up and down is provided on the groove; the left steering wheel bracket 21 is connected to the steering wheel control motor base 26 through the left steering wheel slider 24, and the right steering wheel bracket is connected to the steering wheel control motor base 26 through the right steering wheel slider 25. The steering wheel control motor 23 is installed on the steering wheel control motor base 26, and when in use, the steering wheel control motor output shaft 31 is coaxial with the steering wheel 33.

[0031] In some embodiments, the mounting seat assembly includes a left base plate 3 and a right base plate 2. Brake part bases 10, steering wheel clamping arm bases 9, and control part base plates 7 are all mounted on the left base plate 2 and the right base plate 2. Base screw holes 1 are provided on both the left base plate 3 and the right base plate 2. Threaded holes are provided on the brake part bases 10, the steering wheel clamping arm bases 9, and the control part base plates 7. The brake part bases 10 are mounted on the left base plate 3 and the right base plate 2 through brake part base fastening screws 11. The steering wheel clamping arm bases 9 are mounted on the left base plate 3 and the right base plate 2 through steering wheel clamping arm base fastening screws. The control part base plates 7 are mounted on the left base plate 3 and the right base plate 2 through control part base plate fastening screws.

[0032] In some embodiments, the brake control assembly includes a brake bracket 14 and a brake linear motor 13. The brake bracket 14 is mounted on the brake part base 10. A brake linear motor 13 is mounted on the brake bracket 14. The central axis of the brake linear motor 13 abuts against the brake pedal plate 12.

[0033] In some embodiments, the throttle control assembly includes a throttle bracket 16 and a throttle linear motor 15. The throttle bracket 16 is mounted on the brake part base 10. A throttle linear motor 15 is mounted on the throttle bracket 16. The central axis of the throttle linear motor 15 abuts against the throttle pedal plate.

[0034] The brake control assembly and the throttle control assembly are firmly connected to the main body. With the synergistic effect of the main body structure and the high-performance servo motor, it is ensured that a stable position and posture are always maintained during operation. Through precise control of the servo motor, fine operation of the brake and throttle is achieved, ensuring that the vehicle can accelerate and brake smoothly according to the preset controllable acceleration, greatly improving the comfort and safety of driving.

[0035] In some embodiments, the left steering wheel clamping arm 17 and the right steering wheel clamping arm 20 are both mounted on the steering wheel clamping arm base 9 through steering wheel clamping arm fastening blocks 8.

[0036] In some embodiments, the control cabinet 5 is fixed on the control part base plate 7 through a cabinet fastening block 6.

[0037] In some embodiments, the gear shifting control assembly includes a three-axis robotic arm base 4, a three-axis robotic arm 29, and a gear shifting sleeve head 30. The three-axis robotic arm base 4 is mounted on the control part base plate 7. A three-axis robotic arm 29 is mounted on the three-axis robotic arm base 4. The tail end of the three-axis robotic arm 29 is connected to the gear shifting sleeve head 30 for gear shifting.

[0038] The gear control assembly is equipped with an optional gear control mechanism that can accurately control various types of gear shifting devices, such as touch-type, button-type, hand-held gear shift lever type and traditional type. The gear shift control mechanism closely cooperates with the servo motor of the mechanical arm through its own high-precision locking device to always maintain a stable position and working state, thereby achieving reliable and accurate gear shifting operations to meet the gear switching needs of the vehicle under different driving conditions.

[0039] In some embodiments, one steering wheel turning claw 22 includes two steering wheel fastening claws 34 , and the two steering wheel fastening claws 34 are fixed on the steering wheel 33 by fastening nuts 32 .

[0040] The main frame can be detachably fixed to the vehicle driving seat, and the base includes a left and right double base, which is fixed by fastening screws connecting the driver's seat slide rail and the vehicle.

[0041] In order to ensure the stability and reliability of the overall base, the steering wheel support is clamped by the left and right double clamp arms, and the fastening method of the base and the vehicle is used to provide strong support for the steering wheel steering control.

[0042] In order to ensure that the left and right double clamp arms can clamp normally, a sliding groove is designed at the base of the steering wheel clamp arm to facilitate the left and right double clamp arms to adapt to steering wheel shaft jackets of different sizes.

[0043] The left and right steering wheel brackets connected to the left and right double clamp arms ensure that the base of the steering wheel control motor is parallel to the steering wheel surface through the steering wheel bracket screws that can adjust the angle.

[0044] Under the condition of ensuring parallelism, the left steering wheel slider 24 and the right steering wheel slider 25 are used to facilitate the adjustment of the axis center of the steering wheel control motor 23 and the axis center of the steering wheel.

[0045] Under the condition of ensuring parallelism and axis alignment, the steering wheel is controlled by the steering wheel control motor 23 to directly drive the steering wheel.

[0046] The steering wheel steering claw 22 directly controls the steering of the steering wheel, and adopts four steering wheel fastening claws, with two of each clamping one steering wheel radius rib, and tightening two steering wheel radius ribs in total. It is equivalent to a person's hands grasping the steering wheel, ensuring that the control motor and the steering wheel are precisely synchronized.

[0047] Based on the left and right brackets of the steering wheel, a camera system 28 is installed, including 4 sets of in-vehicle cameras, which are responsible for visual tracking of the left rearview mirror, the right rearview mirror, the front, and the rear mirror. The camera simulates the driver observing the road conditions in the above four directions.

[0048] Based on the above four cameras, we specially train and learn common road condition videos to perceive the surrounding conditions of the vehicle.

[0049] For the brake and throttle control, a linear servo motor is directly used for control respectively, and the vehicle control center is used for control and coordination respectively.

[0050] The gear control robotic arm (301) connected to the main body frame is equipped with a sleeve type at its end, and can support the operations of a touch type, a button type, a stalk shift lever type and a traditional shift device at the same time; To adapt to the shift tables of different heights, bases of different heights are equipped for the triple-axis robotic arm 29, so that the triple-axis robotic arm has better adaptability.

[0051] The steering robotic arm connected to the main body frame integrates a torque-angle double closed-loop control algorithm to compensate the steering wheel return torque in real time.

[0052] The lossless integration and switching mechanism enables the anthropomorphic driverless robot to be losslessly integrated into the existing manned vehicles and achieve smooth switching between the manned driving mode and the driverless driving mode.

[0053] The robotic arm anthropomorphic operating system includes a robotic arm, a sensor, a controller and an actuator.

[0054] The installation process of the anthropomorphic driverless robot described in the present invention includes: (1) Fix the robot base in the cockpit through the bottom of the seat and the vehicle locking device, install the brake piece base 10, adjust the position of the brake piece base 10 according to the size, and ensure that there is a gap of about 4-6 cm between the brake pedal plate of the brake linear motor 13 and the accelerator pedal plate of the accelerator linear motor 15 and the brake pedal and the accelerator pedal of the vehicle when in the shortest stroke.

[0055] (2) Install the steering control robotic arm of the steering wheel: Clamp the steering wheel support shaft by tightening the left clamping arm 17 and the right clamping arm 20 of the steering wheel, and the shaft provides a stable and reliable supporting force for the steering control of the steering wheel. Adjust the left support 21 and the right support of the steering wheel to make the steering control motor base parallel to the steering wheel surface, and then tighten the steering wheel support screw 18. Adjust the left slider 24 and the right slider 25 of the steering wheel to align the axis of the steering control motor 23 with the axis of the steering wheel, and then tighten the left slider 24 and the right slider 25 of the steering wheel. Adjust the steering claw 22 of the steering wheel so that the four fastening claws of the steering claw 22 respectively clamp two radius ribs in the steering wheel. Ensure that the steering angle is 0 degrees when the steering wheel is at the 12 o'clock position. Calibrate the steering control motor 23 to ensure that the initial position of the steering control motor 23 is at the 0-degree angle.

[0056] (3) For gear levers of different heights, triaxial robotic arm bases 4 of different heights have been pre-designed and the selection has been completed before installation. At this time, only the triaxial robotic arm 29 needs to be installed on the control part base plate 7, and the height and front-back position are checked.

[0057] (4) After the left and right steering wheel brackets are determined, the in-vehicle vision camera bracket 27 is fastened to the steering wheel bracket. The two left and right brackets of this bracket can be adjusted according to the distance between the left and right steering wheel brackets. The camera base can be adjusted to a certain height, and the camera itself can be adjusted in angle to ensure that the camera has the best vision.

[0058] (5) The control cabinet is directly installed on the control part base plate 7.

[0059] (6) Install radar and vision acquisition devices inside the cockpit and on the side of the vehicle, fix them through magnetic attraction or snap devices, and connect them to the main body through flexible wires or wirelessly.

[0060] (7) The power supply and control circuits of the gear robotic arm, in-vehicle camera, control cabinet, steering wheel control motor, brake and accelerator control linear motors are bundled and straightened, and the power lines are uniformly connected to the vehicle power input port.

[0061] The anthropomorphic driverless robot designed by the present invention can be fixed in the driver's cab, can be adapted to the shape and size characteristics of seats of various different vehicle models through simple assembly and adjustment, and the installation and disassembly processes are simple and fast, which can ensure that the main body always maintains a stable position and posture under various complex working conditions during vehicle driving, providing a solid and reliable foundation for the precise operation of various subsequent robotic arms.

[0062] The following is an introduction to the application of the anthropomorphic driverless robot described in this embodiment: Design for common modes of automatic vehicles: The vehicle has a brake gear, an accelerator gear, manual R gear, N gear, P gear, L gear. There are two types of manual handbrakes: button type and gear lever type. This design is based on the vehicle handbrake being button type or manual type, and the foot-operated type is not considered for the time being. In addition, the situation where pressing force needs to be applied to the manual gear lever is not considered for the time being.

[0063] The anthropomorphic driverless robot described in the present invention further includes a sensing module, which achieves seamless docking through the OBD interface of the vehicle, directly reads 28 parameters such as the torque output, wheel speed, and ESP status of the vehicle ECU through the OBD-III interface, and cleverly arranges sensors at key positions inside the vehicle to obtain key operation information such as the speed, acceleration, steering angle, and braking state of the vehicle in real time and accurately. At the same time, multiple advanced radar and visual acquisition devices are arranged inside the cockpit and on the side of the vehicle using convenient fixing devices such as magnetic attraction or buckles; a differential GPS module (centimeter-level positioning) is installed on the roof to form a global motion sensing network for making decisions on vehicle movement. These devices can be directly and tightly connected to the main body, or can maintain data interaction with the main body through soft wire connection or wireless communication technology, so as to achieve a full-range and non-blind-spot acquisition of the road conditions around the vehicle. The driverless robot also has a powerful autonomous learning ability, performs in-depth learning and analysis on the massive data collected through built-in algorithms, accurately calculates the optimal installation positions of each radar and visual acquisition device, and deeply perceives the road conditions and the complex relationship between the vehicle and the surrounding environment, providing rich and accurate data support for subsequent decision-making.

[0064] The control cabinet described in the present invention includes a decision-making module: using deep learning technology, based on a large amount of real driving data for in-depth training, a highly intelligent deep neural network model is constructed. This model can quickly and accurately make various driving decisions such as accelerating, decelerating, steering, and changing lanes according to the multiple information transmitted by the sensing module. At the same time, through data fusion algorithms, the vehicle's own state information, environmental perception information, and real-time information from the roadside and the cloud are organically fused and deeply analyzed, so as to comprehensively judge the optimal driving strategy to ensure that the vehicle can drive safely and efficiently under various complex road conditions.

[0065] In a closed or low-risk environment, the anthropomorphic driverless robot designed in the present invention repeatedly performs operations such as starting, accelerating, decelerating, steering, moving forward, reversing, and parking while the vehicle is moving slowly; vehicle state data and environmental data are collected from the OBD, in-vehicle video cameras, and cameras around the vehicle. Based on the collected data, the control parameters are adjusted through machine learning algorithms (such as reinforcement learning or PID parameter optimization) until reliable and accurate operations are achieved. During the slow operation of the vehicle, dynamic radar and visual data are collected; the optimal installation positions of the sensors are calculated through autonomous learning algorithms, and the device layout is adjusted to maximize the sensing range and accuracy.

[0066] The anthropomorphic driverless robot described in the present invention will automatically enter the autonomous learning mode after each vehicle startup. By driving slowly and repeatedly performing a series of typical driving operation processes such as starting, accelerating, decelerating, steering, moving forward, reversing, and parking, it autonomously and comprehensively collects the state information of the vehicle under various working conditions. At the same time, it monitors and deeply analyzes the rotational speed and rotation angle of the servo motors in each robotic arm control mechanism, and uses advanced data analysis algorithms to deeply mine and process the massive amounts of collected data, continuously optimizing and adjusting relevant parameters until a set of stable, reliable, and accurate operation parameters is obtained. These optimized parameters can ensure that the robot can always accurately, stably, and reliably control the throttle, brakes, gears, and steering device during subsequent driving, significantly improving the safety and stability of driverless driving.

[0067] The driverless robot designed in the present invention has the following advantages: autonomous learning time < 2 hours (covering 90% of mass-produced vehicle models), shift operation success rate > 98% (ASIL-C level verification), number of parameter optimization iterations < 50 times (compared with more than 300 times of the traditional trial-and-error method), supports cross-vehicle model knowledge transfer, and the adaptation time is shortened by more than 60%.

[0068] The present invention adopts incremental parameter identification. During the cold start phase, a progressive excitation signal (chirp sweep + pseudo-random sequence) is executed, and the throttle-acceleration transfer function is identified online through the least squares recursive algorithm.

[0069] The present invention adopts reinforcement learning optimization. An ε-greedy strategy exploration is superimposed on the benchmark parameters, and the ride comfort (jerk < 3 m / s³) and tracking accuracy (speed error RMS < 0.5 kph) are used as the reward function, and the deep deterministic policy gradient (DDPG) is used to continuously optimize the operation strategy. An operation mode library for typical scenarios (ramp / low adhesion road surface) is established to achieve cross-condition transfer learning.

[0070] Knowledge distillation mechanism: The cloud pre-trained model transfers knowledge to the vehicle-end model through feature distillation: L_KD = ∑KL(p_teacher(s)||p_student(s)) + λ||θ_t - θ_s||² Where: p_teacher(s) and p_student(s) respectively represent the output probability distributions of the teacher model (cloud large model) and the student model (vehicle-end small model) for the input s; θ_t and θ_s respectively represent the parameter vectors of the teacher model and the student model; λ is the regularization coefficient, which controls the weight balance between the KL divergence term and the L2 regularization term; KL(p_teacher(s)||p_student(s)) = ∑p_teacher(s) × ln(p_student(s) / p_teacher(s)), which means that minimizing the KL divergence is equivalent to making the prediction distribution of the student model approximate the "soft labels" of the teacher model, transmitting implicit knowledge similarity.

[0071] Spatio-temporal attention mechanism: Introduce a temporal convolution + spatial attention module in the robotic arm trajectory planning.

[0072] Dynamic reward engineering reward function design: R_t = ω1 × S_success + ω2 × exp(-J_erk) - ω3 × P_energy + ω4 × tanh(1 - D_safety) The jerk index (Jerk) adopts a fourth-order derivative constraint: J_erk = ‖d³x / dt³‖² Safety distance model: D_safety = min(θ_max - θ_current, θ_current - θ_min) Energy consumption model: P_energy = ∑(I_motor² × R_winding) Where: S_success is the task success rate index, which measures whether the agent completes the core task and is a continuous value in the interval [0, 1]; ω1 is the success reward coefficient, which is used to amplify the reward for task success and drive the agent to prioritize achieving the final goal; Jerk is the jerk of the action, which is used to describe the change rate of acceleration and characterize the smoothness of the action. The function form exp(−Jerk) means mapping the jerk to an exponentially decaying reward value. The higher the jerk, the lower the reward; ω2 is the jerk suppression coefficient, which is used to adjust the emphasis on action smoothness, suppress unnecessary jitters or mutations, and improve control stability; Penergy is the energy consumption, which is used to describe the energy consumption of the agent when performing actions; ω3 is the energy balance coefficient, which is used to balance task performance and energy efficiency and prevent the agent from sacrificing short-term benefits for high energy consumption; Dsafety is the safety distance, whose domain is [0, 1], which defines the minimum distance between the agent and the dangerous area (such as obstacles, other vehicles); the function form tanh(1 - D_safety) maps the safety distance to a non-linear reward. The closer the distance (the smaller D_safety), the faster the reward decays; ω4 is the safety sensitivity coefficient, which is used to enhance the sensitivity to safety and ensure that the agent significantly reduces the risk tendency when approaching danger.

[0073] The control method of the brake control component includes: applying a linearly adjustable pressure of 0 - 500N at the end; calculating the target deceleration a based on the vehicle mass and road surface adhesion coefficient, and generating a pressure control curve according to the formula F = ma + kΔv, where k is the hysteresis compensation coefficient of the hydraulic system; when it is detected that ESP intervenes, automatically switching to a pulsed braking strategy, and the single - pulse duration t satisfies: 50ms ≤ t ≤ 200ms.

[0074] The gear control component includes: a self - aligning snap - lock device that adjusts the clamping force in real time through a servo motor and a force - feedback sensor; adopting a control algorithm to configure independent operation logics for different types of shift devices; including steps of low - speed self - calibration, dynamic perception optimization, incremental learning decision - making, and vehicle - road - cloud collaborative control. The low - speed self - calibration stage realizes the compensation of sensor installation position error through SLAM technology.

[0075] This invention supports the following functions: multi - robotic - arm collaborative control, such as automatically adjusting the throttle opening during gear shifting and synchronously activating the brake robotic arm during reverse; environment perception and adaptability, detecting the surrounding environment through on - vehicle cameras and ultrasonic sensors; data recording and analysis, generating a shift performance evaluation report for optimizing the calibration of the vehicle transmission system.

[0076] The control cabinet adopts a heterogeneous architecture of FPGA and MCU. FPGA is responsible for generating real - time control instructions, and MCU executes safety monitoring; there is a hard - wire emergency stop circuit independent of the main control. When it is detected that the displacement of the robotic arm exceeds the safety threshold, the power supply of the actuator is cut off within 15ms.

[0077] This invention supports the following extended functions: receiving ADAS test scenario instructions through the V2X communication module and achieving a speed tracking accuracy of ±0.1kph in the ACC mode; being equipped with a hydrogen fuel cell power supply unit to provide continuous working power of ≥72 hours within the environmental temperature range of - 40°C to 85°C; generating a test report that complies with the ISO8888 standard, including the cloud map of pedal stroke distribution, equivalent fuel consumption calculation, and component life prediction data.

[0078] The present invention fixes the robot body to the driver's seat through a screw locking device; the collaborative control process: the cloud sends the global path to the vehicle end; the roadside device real-time pushes the local environment information (such as the signal light status); the robot generates operation instructions by integrating multi-source data and drives the robotic arm to perform steering, acceleration, or braking. It is fixed to the original vehicle through a snap locking device without damaging the vehicle structure; multi-robotic arm collaboration: modular robotic arms can be extended and adapted to the operation interfaces of different vehicle models with minor modifications; it perceives the external environment through vision, simulating the driver's perception of vehicle speed and driving state, without the need to precisely know information such as vehicle speed, acceleration, and steering angle like traditional driverless vehicles; the visual perception of the environment is trained and learned through a large amount of vehicle motion video data, combined with lidar technology and vehicle OBD, achieving the purpose of simulating the human body's perception of the vehicle's motion state through both eyes and somatosensation. Precise force control execution: The braking, throttle, and shifting robotic arms adopt servo motors and closed-loop feedback control to ensure operation accuracy and safety.

[0079] Modular mechanical execution system: detachable main frame structure (including multi-directional adjustment snap devices), multi-degree-of-freedom robotic arm group (including a three-axis linkage mechanism for shifting, steering, and braking), and a control panel recognition algorithm based on machine vision.

[0080] Technical effects: It can achieve the unmanned transformation of the vehicle within 1 hour (compared with more than 72 hours in the traditional scheme), support the physical control interface adaptation of more than 95% of mass-produced vehicle models, and the system response delay is <50 ms (reaching the ASIL-D functional safety level).

[0081] In some embodiments, a vehicle testing method uses the above-mentioned anthropomorphic driverless robot that can be realized without damage, and includes the steps: In the initial calibration stage, the amplitude-frequency response characteristics of the vehicle power system are excited by a chirp sweep signal; In the dynamic testing stage, based on the model predictive control algorithm, the pedal operation sequence within the next 3 seconds is optimized by rolling; When a low-adhesion road surface is detected, it automatically switches to the anti-lock mode and controls the slip ratio within the range of 10% - 30%; After the test, the joint current characteristic quantities of the robotic arm are extracted, and the remaining life of the transmission system is predicted through a pre-trained LSTM neural network.

[0082] The system executes the following autonomous learning process after the vehicle starts: Drive slowly (vehicle speed ≤ 10 km / h), and calibrate the throttle and brake operation parameters by repeatedly accelerating and decelerating (acceleration range ±0.5 m / s²); Execute repeated start, stop, and reverse operations, and record the corresponding relationship between the shifting operation force, displacement, and vehicle response; When the throttle / brake operation error ≤ 5%, the shift success rate ≥ 99%, and the vehicle acceleration fluctuation range ≤ ±0.1 m / s², it is determined that the learning is completed and the vehicle enters the normal driving stage.

[0083] In some embodiments, a vehicle testing method uses the above-mentioned anthropomorphic driverless robot with lossless implementation and includes the steps: In the initial learning stage, calibrate the vehicle power system parameters through repeated acceleration / deceleration and shifting operations; In the normal driving stage, generate optimal operation instructions based on the digital twin model; When the type change of the shifting device is detected, automatically switch to the corresponding control algorithm; After the test, generate a shift performance evaluation report including the force-displacement curve.

[0084] The autonomous learning algorithm based on hierarchical reinforcement learning and Bayesian optimization is used to achieve the adaptive learning and precise execution of vehicle throttle control and brake control. Aiming to realize the smooth transition of a manned vehicle to an unmanned driving mode through innovative robot design technology, filling the gaps in the prior art in terms of lossless transformation of vehicles and efficient cooperative driving while maintaining the original driving functions and safety of the vehicle.

[0085] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A lossless anthropomorphic unmanned robot, characterized in that: include: A mounting base assembly, a control cabinet, and a steering wheel control assembly, a camera assembly, a brake control assembly, a throttle control assembly, and a gear control assembly, all of which are connected to the control cabinet; The steering wheel control assembly comprises a left steering wheel clamp arm and a right steering wheel clamp arm mounted on the mounting seat assembly, a left steering wheel bracket is arranged on the left steering wheel clamp arm, and a right steering wheel bracket is arranged on the right steering wheel clamp arm, and a steering wheel control motor, a steering wheel steering claw and a camera assembly are arranged on the left steering wheel bracket and the right steering wheel bracket, and the steering wheel control motor drives the steering wheel to rotate by controlling the steering wheel steering claw to rotate; the steering wheel steering claw is used to clamp the steering wheel, and the camera assembly is used to simulate the driver to obtain road condition data; The brake control component is used to control the brake; the throttle control component is used to control the throttle, and the gear control component is used to control the gear.

2. The lossless anthropomorphic unmanned robot according to claim 1, characterized in that: The left steering wheel clamp arm and the right steering wheel clamp arm are both provided with multiple threaded holes, and the left steering wheel clamp arm and the right steering wheel clamp arm are connected by screws and threaded holes, and are fastened by steering wheel clamp arm fastening nuts; the left steering wheel clamp arm and the left steering wheel bracket are connected by steering wheel bracket screws and threaded holes, and the right steering wheel clamp arm and the right steering wheel bracket are connected by steering wheel bracket screws and threaded holes.

3. The lossless anthropomorphic unmanned robot according to claim 1, characterized in that: The left steering wheel bracket and the right steering wheel bracket are provided with camera brackets, and the camera brackets are provided with a camera system.

4. The lossless anthropomorphic unmanned robot according to claim 1, characterized in that: The left steering wheel bracket is provided with a groove, and the groove is provided with a left steering wheel slider that slides up and down; the right steering wheel bracket is provided with a groove, and the groove is provided with a right steering wheel slider that slides up and down; the left steering wheel bracket is connected to the steering wheel control motor base through the left steering wheel slider, and the right steering wheel bracket is connected to the steering wheel control motor base through the right steering wheel slider. The steering wheel control motor is installed on the steering wheel control motor base, and when in use, the steering wheel control motor output shaft is coaxial with the steering wheel.

5. The lossless anthropomorphic unmanned robot according to claim 1, characterized in that: The mounting base assembly includes a left base plate and a right base plate, on which a brake base, a steering wheel clamp arm base and a control base plate are installed, the left base plate and the right base plate are provided with base screw holes, the brake base, the steering wheel clamp arm base and the control base plate are provided with threaded holes, the brake base is installed on the left base plate and the right base plate by means of brake base fastening screws, the steering wheel clamp arm base is installed on the left base plate and the right base plate by means of steering wheel clamp arm base fastening screws, and the control base plate is installed on the left base plate and the right base plate by means of control base plate fastening screws.

6. The lossless anthropomorphic unmanned robot according to claim 5, characterized in that: The brake control assembly comprises a brake bracket and a brake linear motor. The brake bracket is mounted on a brake member base. The brake linear motor is mounted on the brake bracket. The central axis of the brake linear motor abuts against the brake pedal block.

7. The lossless anthropomorphic unmanned robot according to claim 5, characterized in that: The throttle control component comprises a throttle bracket and a throttle linear motor. The throttle bracket is mounted on the brake component base. The throttle linear motor is mounted on the throttle bracket. The central axis of the throttle linear motor abuts against the throttle pedal block.

8. The lossless anthropomorphic unmanned robot according to claim 5, characterized in that: The left steering wheel clamp arm and the right steering wheel clamp arm are both mounted on the steering wheel clamp arm base via a steering wheel clamp arm fastening block; or, The control cabinet is fixed on the control component bottom plate through a cabinet fastening block.

9. The lossless anthropomorphic unmanned robot according to claim 1, characterized in that: The gear shift control assembly includes a three-axis robotic arm base, a three-axis robotic arm and a gear shift sleeve. The three-axis robotic arm base is installed on the control component bottom plate. The three-axis robotic arm is installed on the three-axis robotic arm base. The tail end of the three-axis robotic arm is connected to the gear shift sleeve for shifting gears.

10. The lossless anthropomorphic unmanned robot according to claim 1, characterized in that: A steering wheel steering claw comprises two steering wheel fastening claws, and the two steering wheel fastening claws are fixed on the steering wheel through fastening nuts.

Images