A four-wheel four-rotation unmanned vehicle linear control chassis control method

Abstract

This invention discloses a four-wheel, four-rotation driverless vehicle (HV-WHM) chassis control method, relating to the field of HV-WHM chassis control technology for autonomous vehicles. By treating the steering phase of the four wheels as a four-node coupled oscillator system, a dynamic equation for the phase evolution of each wheel is established. A phase offset matrix is ​​determined based on the desired motion mode, and the mode is encoded as an external excitation term. The steering angle and vehicle speed of each wheel are collected in real time, and the natural frequency and current phase difference are calculated. The steering phase of each wheel is updated using a numerical integration method. The updated phase is converted into a target steering angle command and output to the HV-WHM actuator. During mode switching, a smooth transition is achieved by adjusting the phase offset matrix through linear interpolation. In the event of a single wheel failure, the system automatically downgrades to a three-node oscillator system for fault-tolerant resynchronization. Compared to existing technologies, this invention has advantages such as shock-free mode switching and low computational burden, making it suitable for scenarios requiring frequent switching of motion modes, such as smart logistics and agricultural harvesting.

Description

Technical Field

[0001] This invention relates to the field of drive-by-wire chassis control technology for unmanned vehicles, and particularly to a drive-by-wire chassis control method for a four-wheel, four-rotation unmanned vehicle. Background Technology

[0002] Four-wheel, four-turn unmanned vehicles (i.e., driverless vehicles with all four wheels capable of independent steering) have shown broad application prospects in fields such as smart logistics, agricultural harvesting, military reconnaissance, and last-mile delivery due to their extremely high maneuverability. Unlike traditional front-wheel steering vehicles, four-wheel, four-turn unmanned vehicles can achieve special movement modes such as rotating on the spot, lateral crabbing, and small-radius turning, effectively solving practical problems such as difficulty in turning around in narrow alleys, damaging seedlings when turning in fields, and difficulty in parallel parking.

[0003] Currently, the drive-by-wire chassis control of four-wheel, four-rotor unmanned vehicles mainly adopts a hierarchical control architecture (path tracking → stability control → torque distribution), including three levels: top-level trajectory planning, middle-level stability control, and bottom-level torque distribution. Among them, the middle-level control widely uses algorithms such as model predictive control, sliding mode control, and linear quadratic regulators to generate the desired steering angle and driving torque of each wheel by solving optimization problems or designing sliding surfaces.

[0004] However, existing methods such as model predictive control require solving constrained optimization problems in each control cycle, resulting in high computational demands and high computing power requirements for the onboard controller, making it difficult to run in real time on low-cost embedded platforms. Furthermore, when switching between different motion modes such as straight driving, crab driving, and stationary rotation, existing methods typically employ discrete state machines combined with instruction sequences, leading to abrupt changes in the steering angle commands of each wheel, causing significant shocks and vibrations, affecting ride comfort and actuator lifespan. Therefore, there is an urgent need for a four-wheel, four-turn autonomous vehicle drive-by-wire chassis control scheme that achieves computationally lightweight operation and smooth mode switching. To this end, a four-wheel, four-turn autonomous vehicle drive-by-wire chassis control method is proposed. Summary of the Invention

[0005] The main objective of this invention is to provide a four-wheel, four-rotation driverless vehicle drive-by-wire chassis control method, which can effectively solve the problems in the background art.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] This invention provides a method for controlling the drive-by-wire chassis of a four-wheel, four-rotation unmanned vehicle, comprising the following steps: Step S1: Establish the phase dynamic model of the coupled oscillator Steering phase of the four wheels , , , Treating it as a four-node coupled oscillator system, the dynamic equations for the phase evolution of each wheel are established, as follows: ,in, Let i be the steering phase of the i-th wheel, in radians; Let be the natural frequency of the i-th wheel, which is proportional to the current vehicle speed, and its unit is rad / s; The coupling strength coefficient between wheels i and j reflects the strength of the cooperative interaction between the wheels, in rad / s; The desired phase offset, determined by the desired motion mode, is expressed in rad. This is the external instruction stimulus term, used to introduce the desired instruction from the upper-level planner, in rad / s; i,j=1,2,3,4 correspond to the left front wheel, right front wheel, left rear wheel, and right rear wheel, respectively.

[0008] Step S2: Pattern Encoding and Stimulus Generation Based on the desired motion pattern output by the upper-level planner, the corresponding desired phase offset matrix is ​​determined by looking up a table. The desired motion modes include at least: straight-line mode, crab-like mode, stationary rotation mode, and small-radius turning mode; in: In straight-ahead mode, all ; Crab mode, ,the remaining ; In stationary rotation mode, , ; In small radius turning mode ,the remaining ,in It is determined by the turning radius and wheelbase.

[0009] Simultaneously, the desired motion pattern is encoded as an external stimulus term. The calculation method is as follows: ,in The excitation gain is expressed in rad / s (preferably 0.1-0.5 rad / s). Sensitivity coefficient (preferably 2); The desired phase of the i-th wheel under the current desired motion mode; This represents the actual phase of the current i-th wheel.

[0010] Step S3: Real-time status acquisition and parameter calculation Real-time acquisition of actual wheel steering angle signals, vehicle speed signals, and vehicle body yaw rate signals; calculation of the natural frequency of each wheel based on the current vehicle speed. ; Calculate the current phase difference using the actual rotation angle. .

[0011] in: When the desired motion mode is straight-line mode, crab mode, or small-radius turning mode, the natural frequency The calculation method is as follows: ,in The longitudinal velocity of the vehicle's center of gravity, in m / s; The effective rolling radius of the wheel, in meters (m). When the desired motion pattern is a stationary rotation pattern, the natural frequency The calculation method is as follows: ,in The linear velocity of rotation is expressed in m / s. , The radius of rotation, in meters, is the distance from the center of the wheel to the center of rotation of the vehicle. The desired rotational angular velocity, in rad / s, is the angular velocity command given by the upper-level planner or driver for the vehicle body to rotate around the vertical axis. Positive values ​​usually correspond to clockwise rotation, and negative values ​​correspond to counterclockwise rotation.

[0012] Step S4: Numerical integration of the dynamic equations Substitute the current state parameters obtained in step S3 into the dynamic equations of step S1, and update the steering phase of each wheel using a numerical integration method. Phase normalization is performed after each integration step: The phase value is mapped to the interval (-π, π].

[0013] The numerical integration method is either the fourth-order Runge-Kutta method or the Euler forward method.

[0014] Step S5: Control command generation and output Updated steering phase Convert to target steering angle command for each wheel And output to the steer-by-wire actuator.

[0015] Target corner command The calculation method is as follows: ,in Let be the steering ratio of the i-th wheel; Simultaneously apply safety limits: ,in , These are the minimum and maximum allowable turning angles, respectively.

[0016] Mode switching method When the desired motion mode changes, the phase of each wheel is not directly modified. Instead, the desired phase bias matrix is ​​used. From the current mode Linear interpolation transition to the target mode transition time No more than (Preferred time: 0.5 to 1.0 seconds); During this process, the phases of the four wheels are smoothly switched by automatic evolution of coupled oscillator dynamics, without the need for discrete logic judgment.

[0017] Fault-tolerant control methods When a steering actuator malfunction is detected in a single wheel, the phase of that wheel is... The current value is fixed, and the row and column corresponding to the wheel are removed from the coupling matrix K. The remaining three healthy wheels form a three-node coupled oscillator system and are automatically resynchronized to the new phase-locked state.

[0018] Furthermore, the coupling strength coefficient For the vehicle speed adaptive function, satisfying ,in It is a monotonically decreasing function. , To preset the critical speed, This is the transition width.

[0019] And coupling strength coefficient Satisfying symmetry constraints Furthermore, they are categorized into three types based on the physical position of the wheels: coaxial coupling , The value range is 0.5~2.0 rad / s; ipsilateral coupling , The value range is 1.0~3.0 rad / s; diagonal coupling , The value range is 0.8~2.5 rad / s.

[0020] Beneficial effects Compared with the prior art, the present invention has the following beneficial effects: This invention achieves this by continuously adjusting the desired phase offset. Instead of a direct command jump, the mode switching is achieved by utilizing the phase slip characteristics of the coupled oscillator dynamics, eliminating the shock and jitter caused by step commands.

[0021] The core calculation of this invention is only the numerical integration of four first-order differential equations. Each control cycle requires only a few dozen floating-point operations, which can be run in real time on an STM32F4-level MCU. Compared with model predictive control, the amount of computation is reduced by about 80%.

[0022] This invention utilizes the decentralized nature of the oscillator network. After an actuator failure, the coupled oscillators of the remaining wheels will automatically resynchronize, eliminating the need for explicit fault diagnosis and control redistribution, resulting in a rapid response.

[0023] The invention has low dependence on vehicle physical parameters; the main adjustable parameter is the coupling strength coefficient. With the excitation gain γ, the workload for actual vehicle calibration is small.

[0024] This invention has significant advantages in applications such as low-speed park logistics and agricultural harvesting, where frequent switching between straight-line movement, crab-like movement, and stationary rotation is required. Attached Figure Description

[0025] Figure 1 This is a schematic diagram of the overall structure of a four-wheel, four-rotor unmanned vehicle drive-by-wire chassis control method according to the present invention; Figure 2 In this embodiment of the invention, during the process of accelerating the vehicle speed from 0 to 10 m / s, K 12 A schematic diagram of the trajectory of (v) change; Figure 3 This is a schematic diagram of the wheel numbers and coordinate system of the four-wheeled, four-rotor unmanned vehicle in an embodiment of the present invention; Figure 4 This is a schematic diagram of the wheel phase vector diagram for four motion modes in an embodiment of the present invention; Figure 5 This is a schematic diagram of the vehicle speed adaptive coupling strength adjustment function σ(v) in an embodiment of the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0027] Example 1: System Hardware Configuration and Parameter Initialization This embodiment was implemented on a four-wheel, four-rotor unmanned vehicle platform. See [link to wheel numbering and coordinate system]. Figure 3 .

[0028] 1.1 Vehicle physical parameters

[0029] 1.2 Controller Hardware Main control chip: STM32F407 (ARM Cortex-M4, 168MHz) Control period: T s = 0.005 s (200 Hz) Communication interface: CAN bus (500 kbps) 1.3 Initialization Parameters Nominal coupling strength matrix K nom (Unit: rad / s): ; Control parameters:

[0030] Initial phase state: All wheels are initially set to 0 phase: θ1=θ2=θ3=θ4= 0, corresponding to straight driving mode.

[0031] Example 2: Conventional Driving Control Procedure This embodiment demonstrates the standard control process of a vehicle in straight-ahead mode.

[0032] 2.1 Expected Instruction Reception The upper-level planner issues the following instructions: Expected pattern: Straight ahead; Expected speed: v des = 3.0 m / s; 2.2 Status Acquisition (t=0 time) Sensor measured data:

[0033] Current vehicle speed: v = 2.8 m / s Vehicle body yaw rate: ω z = 0.02 rad / s 2.3 Natural Frequency Calculation (Straight-line Mode) In straight-line mode, all wheels naturally have the same frequency: .

[0034] 2.4 Calculation of External Excitation Desired phase in straight-ahead mode = 0.

[0035] Taking the left front wheel (i=1) as an example:

[0036] Similarly, the excitation terms for the remaining wheels are all close to 0, indicating that the system is close to the desired state.

[0037] 2.5 Calculation of Coupling Terms Taking the left front wheel (i=1) as an example, calculate the coupling terms:

[0038] 2.6 Dynamic Integrals (using Euler's forward method as an example) Left front wheel phase change rate:

[0039] Phase update:

[0040] 2.7 Phase Normalization It lies within the interval (-π, π] and does not require normalization.

[0041] 2.8 Generation of Cornering Commands Target corner:

[0042] The current vehicle speed is approximately 2.8 m / s. According to the limit table, the maximum permissible turning angle is approximately 60°. Since 48.6° < 60°, the command is valid.

[0043] 2.9 Repeated execution The above steps are executed cyclically every 5ms to form a continuous closed-loop control.

[0044] Example 3: Mode switching process (straight movement → crab movement) This example demonstrates the complete process of switching from straight-line mode to crab-like (lateral movement to the right) mode.

[0045] 3.1 Switching Trigger At t = 2.0 s, the upper-level planner issues a mode switching instruction: switch from straight line to crab line (right).

[0046] 3.2 Phase Offset Transition Do not directly modify θ i Instead, α ij Linear interpolation transition. Transition time T trans = 0.6 s.

[0047] Target α ij Value (crab mode):

[0048] Linear interpolation formula (with α) 13 (For example) .

[0049] 3.3 Natural Evolution of the System In α ij During the slow change, the phases of the four wheels are automatically adjusted through the dynamics of the coupled oscillator.

[0050] Simulation results (key time points):

[0051] Note: θ3-θ1 and θ4-θ1 stabilize at π / 2 (1.57 rad) at t=2.60s, and θ2-θ1 stabilizes at 0, achieving phase locking in the crab mode.

[0052] 3.4 Confirmation of Transition Completion When all phase differences meet the conditions, the transition is considered complete. Once the conditions are met, freeze α. ij The target value has been reached; mode switching is complete.

[0053] 3.5 Comparison of switching effects Compared with traditional methods (discrete state machine + corner transition):

[0054] Example 4: In-situ rotation mode control This embodiment demonstrates the control process of a vehicle rotating clockwise in place.

[0055] 4.1 Command Reception At t = 5.0 s, the upper-level planner issues the following instruction: Pattern: Rotate in place (clockwise) Desired rotational angular velocity: ω cmd = 1.5 rad / s 4.2 Calculation of Radius of Rotation Wheelbase L = 1.2 m, track width W = 0.9 m: ; 4.3 Calculation of Rotational Linear Velocity ; 4.4 Calculation of the natural frequency of each wheel

[0056] 4.5 Phase Offset Setting (Clockwise Rotation)

[0057] 4.6 Phase Evolution Results After a transition of approximately 0.3 seconds, the phases of all four wheels are locked to:

[0058] Physical meaning: The left front wheel and left rear wheel point directly forward (0°), and the right front wheel and right rear wheel point directly backward (180°), forming a rotational torque around the center.

[0059] 4.7 Turning command output With ρ i = 15 Calculation:

[0060] Note: The actual actuator has a limit of 85°, so the right front wheel and right rear wheel output 85° (approximately 180° effect, which is achieved by combining with the wheel speed difference to achieve rotation).

[0061] Example 5: Vehicle Speed ​​Adaptive Coupling Regulation This example demonstrates the coupling strength K. ij The mechanism of adaptive adjustment according to vehicle speed.

[0062] 5.1 Definition of Adaptive Function See Figure 5 Vehicle speed adaptive adjustment function: ; Parameter settings: Critical speed: v0 = 15 m / s; Transition width: Δv = 5 m / s.

[0063] 5.2 Adjustment factors at different vehicle speeds

[0064] 5.3 Control characteristics at different vehicle speeds Low-speed operation (v=1 m / s): σ≈0.94, strong coupling is maintained; It responds sensitively to maneuvers such as crab walks and spinning in place; Phase synchronization is established quickly.

[0065] Medium-speed operation (v=5 m / s): σ≈0.80, moderate coupling; Balancing mobility and stability; The mode switching is still relatively smooth.

[0066] High-speed operation (v=10 m / s): σ≈0.73, weak coupling; Prioritize linear stability; Reduce unnecessary phase synchronization and avoid oscillation.

[0067] Extreme high-speed operation (v=20 m / s): σ≈0.27, extremely weak coupling; The system degenerates into near-independent wheel control; To ensure safety at high speeds to the greatest extent possible.

[0068] 5.4 Dynamic Adjustment Effect During the process of accelerating from 0 to 10 m / s, K 12 (v) See the trajectory of change. Figure 2 .

[0069] Example 6: Fault-tolerant control (single-wheel steering actuator failure) This embodiment demonstrates the fault-tolerant control process when the steering actuator of the right front wheel (i=2) is stuck.

[0070] 6.1 Fault Detection and Identification At t = 8.0 s, the system detected an anomaly: δ 2，sensor = 6.9° (0.12 rad); Command response deviation > threshold (5°) for 0.12 consecutive seconds.

[0071] Diagnosis: Right front wheel actuator malfunction (stuck). Current vehicle speed v = 2 m / s.

[0072] 6.2 Fixed phase of the faulty wheel Fixed faulty wheel phase: .

[0073] 6.3 Reduction of the order of the coupling matrix Original 4×4 coupling matrix: ; Removing the second row and second column (right front wheel) yields a 3×3 matrix K' (corresponding to wheels 1, 3, and 4): .

[0074] 6.4 Reduced-order dynamic equations The dynamic equations for healthy wheels 1, 3, and 4 are as follows:

[0075]

[0076]

[0077] in , , This represents the influence coefficient of the faulty wheel on the healthy wheel in the original coupling.

[0078] 6.5 Resynchronization Process The evolution trajectory of each wheel phase after the fault occurred:

[0079] 6.6 Fault Tolerance Effect Evaluation

[0080] Conclusion: Resynchronization was completed within 0.5 seconds after the fault, and the vehicle drove stably with a small deviation, without loss of control.

[0081] Example 7: Adaptability under different road surface adhesion coefficients This embodiment verifies the robustness of the method under different road surface conditions.

[0082] 7.1 Test Conditions

[0083] 7.2 Control Parameter Adjustment Coupling strength K ij The excitation gain γ needs to be adjusted only, as it remains unchanged (already adapted via σ(v)).

[0084] Physical meaning: The lower the adhesion coefficient, the stronger the external excitation (larger γ) is required to overcome the response delay caused by insufficient friction.

[0085] 7.3 Test Results

[0086] 7.4 Comparison with traditional MPC method (wet road surface)

[0087] Example 8: Comparison of Numerical Integration Methods This embodiment compares the differences in control accuracy and computational burden between different numerical integration methods.

[0088] 8.1 Test Conditions Straight-line mode, v=3 m / s; Simulation duration: 10 seconds; Integral step size: T s = 0.005 s.

[0089] 8.2 Comparison of Integration Methods

[0090] 8.3 Real-time verification Actual test on the STM32F407 platform:

[0091] Conclusion: The method of this invention can meet the accuracy requirements using the Euler forward method, and the computational resource consumption is extremely low.

[0092] Example 9: Performance Summary This embodiment summarizes the overall performance of the method of the present invention across various dimensions.

[0093] 9.1 Core Performance Indicators

[0094] 9.2 Smoothness of switching between modes

[0095] 9.3 Fault Tolerance

[0096] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A method for controlling the drive-by-wire chassis of a four-wheel, four-rotation unmanned vehicle, characterized in that, Includes the following steps: S1: Treat the steering phase of the four wheels as a four-node coupled oscillator system and establish the dynamic equations for the phase evolution of each wheel. S2: Based on the desired motion pattern output by the upper-level planner, determine the corresponding desired phase offset matrix by looking up a table. The desired motion pattern is encoded as an external stimulus term. ; S3: Real-time acquisition of the actual steering angle and vehicle speed of each wheel, and calculation of the current natural frequency of each wheel. and current phase difference ; S4: Substitute the current state parameters obtained in step S3 into the dynamic equations of step S1, and update the steering phase of each wheel using a numerical integration method. ; S5: Update the steering phase Convert to target steering angle command for each wheel The output is sent to the steering-by-wire actuator.

2. The method for controlling a four-wheel, four-turn unmanned vehicle drive-by-wire chassis according to claim 1, characterized in that, The dynamic equation for the evolution of the wheel phase is expressed as: ,in, Let i be the steering phase of the i-th wheel; Let be the natural frequency of the i-th wheel; The coupling strength coefficient between wheels i and j; For the desired phase offset; For external instruction excitation terms; i,j=1,2,3,4.

3. The method for controlling a four-wheel, four-turn unmanned vehicle drive-by-wire chassis according to claim 1, characterized in that, The desired motion patterns include at least: straight-line mode, crab-like mode, stationary rotation mode, and small-radius turning mode; wherein: In straight-ahead mode, all ; Crab mode, ,the remaining ; In stationary rotation mode, , ; In small radius turning mode ,the remaining , Determined by the turning radius and wheelbase; When the desired motion mode is straight-line mode, crab mode, or small-radius turning mode, the natural frequency The calculation method is as follows: ,in The longitudinal velocity of the vehicle's center of gravity. The effective rolling radius of the wheel; When the desired motion pattern is a stationary rotation pattern, the natural frequency The calculation method is as follows: ,in is the rotational linear velocity.

4. The method for controlling a four-wheel, four-turn unmanned vehicle drive-by-wire chassis according to claim 1, characterized in that, The coupling strength coefficient For the adaptive function of vehicle speed, satisfying ,in It is a monotonically decreasing function. , To preset the critical speed, This is the transition width.

5. The method for controlling a four-wheel, four-turn unmanned vehicle drive-by-wire chassis according to claim 1, characterized in that, The external incentives The calculation method is as follows: ,in For excitation gain; Sensitivity coefficient; The desired phase of the i-th wheel under the current desired motion mode; This represents the actual phase of the current i-th wheel.

6. The method for controlling a four-wheel, four-turn unmanned vehicle drive-by-wire chassis according to claim 1, characterized in that, The numerical integration method is either the fourth-order Runge-Kutta method or the Euler forward method; Phase normalization is performed after each integration step: The phase value is mapped to the interval (-π, π].

7. The method for controlling a four-wheel, four-turn unmanned vehicle drive-by-wire chassis according to claim 1, characterized in that, It also includes a mode switching step: when the desired motion mode changes, the desired phase offset matrix is ​​adjusted. From the current mode Linear interpolation transition to the target mode and transition time No more than During this process, the phases of the four wheels are smoothly switched through the automatic evolution of coupled oscillator dynamics.

8. The method for controlling a four-wheel, four-turn unmanned vehicle drive-by-wire chassis according to claim 1, characterized in that, It also includes a fault-tolerant control step: when a steering actuator failure is detected in a single wheel, the phase of the faulty wheel is... The current value is fixed, and the row and column corresponding to the wheel are removed from the coupling matrix K. The remaining three healthy wheels form a three-node coupled oscillator system and are automatically resynchronized to the new phase-locked state.

9. The method for controlling a four-wheel, four-turn unmanned vehicle drive-by-wire chassis according to claim 1, characterized in that, The target turning command The calculation method is as follows: ,in Let be the steering ratio of the i-th wheel; Simultaneously apply safety limits: ,in , These are the minimum and maximum allowable turning angles, respectively.

10. A method for controlling the drive-by-wire chassis of a four-wheel, four-rotor unmanned vehicle according to any one of claims 1-9, characterized in that, The coupling strength coefficient Satisfying symmetry constraints Furthermore, they are categorized into three types based on the physical position of the wheels: coaxial coupling , The value range is 0.5~2.0 rad / s; ipsilateral coupling , The value range is 1.0~3.0 rad / s; diagonal coupling , The value range is 0.8~2.5 rad / s.

Images