Fault diagnosis and fault-tolerant control method for in-situ steering of distributed driving vehicle on wet and slippery road surface

Abstract

The invention discloses a fault diagnosis and fault-tolerant control method for in-situ steering of a distributed driving vehicle on a wet and slippery road surface, and aims to solve the problem of instability of the vehicle when a driving unit has faults such as driving failure, mechanical locking or excessive slipping. According to the method, the vehicle state is obtained in real time, the initial torque instruction is calculated, and online diagnosis and accurate recognition of the fault mode are achieved. Aiming at hardware faults such as driving failure, the system breaks and isolates a fault wheel; and aiming at excessive slipping, the torque is preferentially and actively reduced to recover the adhesive force, and temporary isolation is carried out after recovery failure. Once the fault unit is confirmed to be isolated, the system immediately starts a torque reconstruction module based on quadratic programming (QP), and the driving task is optimally redistributed to the remaining healthy wheels in real time so as to maintain the expected yaw moment of the whole vehicle. According to the strategy adopted by the invention, faults with different properties can be effectively distinguished and processed in a differentiated manner, and the control stability and active safety of the vehicle under the conditions of severe work and partial failure are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of distributed drive vehicle control, in particular to a fault diagnosis and fault-tolerant control method for a distributed drive vehicle to turn in place on a wet road. BACKGROUND

[0002] With the rapid development of intelligent driving and distributed drive technology, the maneuverability and fault tolerance of multi-axle independently driven vehicles in complex environments have attracted widespread attention. Turning in place as a high-maneuvering driving mode has become an important research direction for unmanned military vehicles, engineering rescue vehicles and special operation platforms.

[0003] In an 8x8 distributed independently driven unmanned vehicle, each wheel is driven by an independent motor, which has complete independent torque control capability. This driving architecture not only improves the driving redundancy of the vehicle, but also provides the possibility for differential control strategies in complex working conditions. However, in actual application, when the vehicle turns in place on a wet, icy, snowy, muddy and other low adhesion road surface, it will face new challenges.

[0004] On the one hand, due to the low adhesion coefficient of the wet road, the 8 wheels of the vehicle are in a large slip state, which easily leads to a serious lack of longitudinal and lateral force between the wheels and the ground, resulting in a turning center offset, a center of mass trajectory drift or even a whole attitude loss during the turning in place process. On the other hand, under the distributed drive architecture, although each drive unit has independent control capability, the torque distribution and coordinated control between the wheels are extremely complex in low adhesion conditions. If there is no effective control and fault-tolerant mechanism, it is easy to cause problems such as excessive wheel slip, drive unit performance degradation or failure, and thus reduce the overall maneuverability and safety of the vehicle. Therefore, the present application proposes a fault diagnosis and fault-tolerant control method for a distributed drive vehicle to turn in place on a wet road to solve the problems in the prior art. SUMMARY

[0005] In view of the above problems in the prior art, the present application aims to solve these problems. The method takes into account the stability of vehicle yaw motion and the control requirements of wheel slip rate, and realizes turning in place in low adhesion conditions and fault-tolerant control in abnormal drive unit conditions.

[0006] The technical solution of the present application is: a fault diagnosis and fault-tolerant control method for a distributed drive vehicle to turn in place on a wet road, comprising the following steps:

[0007] S01: Obtain the state of the drive unit, including the actual wheel speed w i of each wheel, the actual output current I i of the motor, and the actual yaw rate w z of the vehicle.Calculate the vehicle slip ratio λ i ;

[0008] S02: Calculate the expected value based on the throttle opening δ. th Calculate the desired yaw moment Expected wheel speed Desired motor current Desired yaw rate

[0009] S03: Calculate the initial torque command for each drive wheel using a proportional-integral-derivative (PID) speed tracker.

[0010] S04: Determine the fault mode, comprehensively judge the fault mode of the drive unit, and output the availability flag γ of each drive wheel. i ∈(0,1), where γ i =1 indicates that the wheel is available, γ i =0 indicates a fault;

[0011] S05: Reconstruct torque using a quadratic programming (QP) torque reconstructor based on different fault types;

[0012] A preferred method for fault diagnosis and fault-tolerant control of in-situ steering of distributed drive vehicles on slippery roads is characterized by,

[0013] In step S03, the initial torque command for each drive wheel is converted into the target yaw rate command for the vehicle and the initial drive torque T for each wheel. PID The obtained torque is then distributed to each wheel:

[0014]

[0015] A preferred method for fault diagnosis and fault-tolerant control of in-situ steering of distributed drive vehicles on slippery roads is characterized by,

[0016] The fault modes in step S04 include drive failure or performance degradation, mechanical jamming or lock-up, and excessive wheel slippage.

[0017] The drive failure or performance degradation indicates that the controller has issued a clear drive torque command, and the road surface adhesion conditions allow for the transmission of that torque, but the actual motor current is far lower than the command, resulting in insufficient wheel power. The determination is based on the duration of the failure. In the formula, ΔI th1 The first threshold is required. To eliminate misjudgments under low torque commands, and because the actual current of this wheel differs significantly from that of its symmetrical wheel;

[0018] The mechanical jam or lock means that the actual current of the motor is large, but the wheel speed is extremely low or zero, and the judgment basis is that I i >I stall ,w i <w min , where I stall is the locked-rotor current threshold, w min is the low speed threshold;

[0019] The wheel is excessively slipping, and the judgment basis is that the slip ratio λ i is greater than the threshold λ th , and the actual output current I i is required to be large, and the actual yaw rate w z of the vehicle is significantly smaller than the expected value, and the wheel speed is significantly higher than the speed of at least one other wheel;

[0020] The torque distribution optimization model in the step S05 adopts quadratic programming, and the optimization objective and the constraint condition include:

[0021] The objective function of the optimization objective is where T i is the reconstructed output torque of the i-th healthy driving unit, T i ref is the initial expected torque of the driving unit, and N is the number of healthy driving units;

[0022] The constraint condition is:

[0023]

[0024] In the formula, H is the total of the healthy driving units, F xi , and F yi is the longitudinal force and lateral force generated by each driving unit, M zi is the corresponding yaw moment;

[0025] The output form is obtained by solving the quadratic programming problem, and the corrected torque {T i} of each healthy driving unit that satisfies the constraint condition is obtained, which is used to replace the original control instruction, so as to realize the fault-tolerant stable control of the eight-wheel vehicle under the wet road turning condition;

[0026] Preferably, the fault diagnosis and fault-tolerant control method for the eight-wheel vehicle under the wet road turning condition is characterized in that,

[0027] The driving failure or performance degradation and the mechanical jam or lock hardware failure determine that a certain driving unit fails, and immediately stop sending driving instructions to the unit, and force the expected driving torque T i refZero, the original allocation to the drive task of the fault wheel, through the optimization algorithm is re-allocated to all remaining healthy wheels;

[0028] Preferably, the fault diagnosis and fault-tolerant control method for in-place turning of a distributed drive vehicle on a wet road surface is characterized in that,

[0029] The excessive slip requires the driving torque of the slipping wheel to be gradually reduced until the slip ratio returns to the optimal adhesion region, at which time the wheel can provide the maximum effective driving torque T i,max = μ i F z,i R, wherein μ i is the road adhesion coefficient, F z，i is the wheel vertical load, and R is the wheel radius. If the slip ratio of the wheel cannot return to the optimal adhesion region within a given recovery time, it is temporarily set as a failed unit, i.e. γ i = 0, and the torque distribution optimization model is triggered, with the driving task being assumed by other healthy wheels. BRIEF DESCRIPTION OF DRAWINGS

[0030] Figure 1 is a system block diagram of the present application;

[0031] Figure 2 is a workflow diagram of the present application. DETAILED DESCRIPTION

[0032] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0033] The present application provides a technical solution: a fault diagnosis and fault-tolerant control method for in-place turning of a distributed drive vehicle on a wet road surface, as shown in Figure 1 , comprising:

[0034] The state acquisition module: real-time acquisition of various driving wheel speeds, motor currents and vehicle yaw rate and other bottom layer state parameters, and calculation of slip rate and other derived variables, to provide accurate and dynamic data basis for the entire control system.

[0035] The expected value calculation module: analysis of the driver input, conversion of the intention into an overall vehicle level macro control target (such as the expected total yaw moment), and further decomposition into the expected speed of each driving wheel as the top input of the system control.

[0036] Initial torque calculation module: As the bottom layer of speed tracking controller, it calculates the initial driving torque command for each drive wheel in ideal state according to the deviation between expected and actual wheel speed, which is the benchmark for subsequent diagnosis and reconstruction.

[0037] Fault diagnosis module: It monitors and diagnoses the working state of each drive unit online, accurately identifies fault modes such as "drive failure", "mechanical lock", "excessive slip" by comparing the command and feedback information, and outputs the health status of each unit, which is the decision basis for triggering fault-tolerant control.

[0038] Torque reconstruction module: The core execution unit of fault-tolerant control. After the fault occurs, it uses optimization algorithms such as quadratic programming (QP) to real-time and optimally redistribute the driving tasks of the faulty wheel to the remaining healthy wheels under the premise of ensuring the vehicle motion target (total yaw moment) unchanged, realizing active failure safety control.

[0039] As shown in Figure 2 The fault diagnosis and fault-tolerant control method for distributed drive vehicle on wet road, comprising the following steps:

[0040] S01: Get the drive unit state through the controller area network (CAN) bus of each drive motor, including the actual wheel speed w i , the actual output current I i of the motor, and the vehicle driving state through the inertial measurement unit (IMU), including the actual yaw rate w z of the vehicle, and calculate the vehicle slip rate λ i Where v i is the linear speed of the hub center of the i-th wheel relative to the steering center;

[0041] S02: Calculate the expected value, calculate the expected yaw moment M , the expected yaw rate ω , and the expected wheel speed v The calculation formulas are as follows:

[0042]

[0043] Where K m is the throttle gain coefficient, δ th is the normalized throttle opening, M is the maximum on-site steering yaw moment that the vehicle can generate, C z is the yaw damping coefficient of the vehicle, d i is the distance from the i-th wheel to the vehicle steering center, and R

[0044] S03: Transform the vehicle target yaw rate command into initial driving torque T for each wheel by a proportional-integral-derivative (PID) based speed tracker PID , and the resulting torque is sent to each wheel, whose control law can be expressed as:

[0045]

[0046] where e v (t) is the wheel speed error of the i-th wheel, K p , K i , and K d are the proportional, integral, and derivative gains of the PID controller respectively

[0047] S04: Determine the fault mode, determine the fault mode of the drive unit comprehensively, and output the availability flag γ of each drive wheel i ∈(0, 1), where γ i = 1 indicates that the wheel is available, and γ i = 0 indicates a fault, the fault modes diagnosed in this embodiment include:

[0048] Drive failure or performance degradation: indicates that the controller has issued a clear driving torque command, and the road adhesion condition also allows the torque to be transmitted, but the actual motor current is much lower than the command, resulting in insufficient wheel power, which is determined by the fact that the actual current I is less than the first threshold value ΔI th1 for a certain duration, while requiring that the actual current I of the wheel is significantly different from that of the symmetric wheel (e.g., the difference is more than 50%), indicating a single wheel fault;

[0049] The mechanical jam or lock indicates that the actual motor current is large, but the wheel speed is very low or zero, which is determined by the fact that I i > I stall , w i < w min , where I stall is the stall current threshold, which is usually close to the maximum allowable current of the motor, and w min is the low speed threshold for determining whether the wheel is close to stationary;

[0050] The wheel is excessively slipping, which is determined by the fact that the slip ratio λ i is greater than the threshold value λ th for a certain duration, while requiring that the actual output current I i is large, indicating that the drive system is still trying to output torque, and the actual yaw rate w zfar less than expected, indicating that the driving force of the wheel has not been effectively converted into the steering motion of the vehicle, and the wheel speed is significantly higher than the speed of at least one other wheel;

[0051] S05: Reconstruct the torque according to different fault types using a quadratic programming (QP) torque reconstructor, and the optimization objective and constraint conditions include:

[0052] The objective function is where T i is the reconstructed output torque of the i-th healthy driving unit, T i ref is the initial desired torque of the driving unit, and N is the number of healthy driving units;

[0053] The constraint conditions are:

[0054]

[0055] In the formula, H is the total of healthy driving units, F xi , and F yi is the longitudinal force and lateral force generated by each driving unit, M zi is the corresponding yaw moment;

[0056] The output form is obtained by solving a quadratic programming problem, and the corrected torque {T i} of each healthy driving unit that satisfies the constraint condition is obtained, which is used to replace the original control instruction to realize fault-tolerant stable control of the eight-wheel vehicle under the wet road turning condition;

[0057] Further, the specific control strategies for different fault types are as follows:

[0058] Once the diagnosis module determines that a driving unit has driving failure or performance degradation and mechanical jamming or hardware failure, it immediately stops sending driving instructions to the unit and forces the desired driving torque T i ref to zero, γ i to zero, and the torque reconstruction module is activated. The module obtains two key inputs: the desired yaw moment of the entire vehicle and the updated list of healthy wheels. At this time, the constraint conditions of the quadratic programming (QP) optimization problem have changed: the task originally undertaken by N wheels must now be completed by N-1 (or fewer) healthy wheels. The QP solver recalculates an optimal torque distribution scheme on the new set of healthy wheels, with the goal of minimizing torque adjustment;

[0059] For excessive slip, it is essentially a matching problem between the wheel and the road, rather than the hardware itself being damaged. Therefore, the core of the strategy is to try to restore its effective driving ability by active adjustment first, and only after the restoration fails, it is temporarily regarded as a failed unit for processing. First, try active control, use a slip ratio error-based PI controller for nonlinear adjustment, gradually reduce the driving torque of the slipping wheel, until the slip ratio returns to the optimal adhesion region, at which time the wheel can provide the maximum effective driving torque T i,max = μ i F z，iR , where μ i is the road adhesion coefficient, F z，i is the wheel vertical load, and R is the wheel radius. If the slip ratio cannot return to the optimal adhesion region within a given recovery time, it is temporarily set as a failed unit, i.e. γ i = 0, and the torque distribution optimization model is triggered, with other healthy wheels taking over its driving task.

Claims

1. A fault diagnosis and fault-tolerant control method for in-situ steering of distributed drive vehicles on slippery roads, characterized in that, Includes the following steps: S01: Obtain the drive unit status, including the actual wheel speed w of each wheel. i The actual output current I of the motor i Simultaneously, the vehicle's driving status is acquired, including the vehicle's actual yaw rate w. z Calculate the vehicle slip ratio λ i ; S02: Calculate the expected value based on the throttle opening δ. th Calculate the desired yaw moment Expected wheel speed Desired motor current Desired yaw rate S03: Calculate the initial torque command for each drive wheel using a proportional-integral-derivative (PID) speed tracker. S04: Determine the fault mode, comprehensively judge the fault mode of the drive unit, and output the availability flag γ of each drive wheel. i ∈(0,1), where γ i =1 indicates that the wheel is available, γ i =0 indicates a fault; S05: Reconstruct torque using a quadratic programming (QP) torque reconstructor based on different fault types.

2. The fault diagnosis and fault-tolerant control method for in-situ steering of distributed drive vehicles on slippery roads according to claim 1, characterized in that... In step S03, the initial torque command for each drive wheel is converted into the target yaw rate command for the vehicle and the initial drive torque T for each wheel. PID The obtained torque is then distributed to each wheel:

3. The fault diagnosis and fault-tolerant control method for in-situ steering of distributed drive vehicles on slippery roads according to claim 1, characterized in that, The fault modes in step S04 include drive failure or performance degradation, mechanical jamming or lock-up, and excessive wheel slippage. The drive failure or performance degradation indicates that the controller has issued a clear drive torque command, and the road surface adhesion conditions allow for the transmission of that torque, but the actual motor current is far lower than the command, resulting in insufficient wheel power. The determination is based on the duration of the failure. In the formula, ΔI th1 The first threshold is required. To eliminate misjudgments under low torque commands, and because the actual current of this wheel differs significantly from that of its symmetrical wheel; The mechanical jamming or seizure indicates that the actual motor current is high, but the wheel speed is extremely low or zero. The criterion for this judgment is that within a certain duration... i >I stall ,w i <w min , where I stall For the stall current threshold, w min Low speed threshold; The determination of excessive wheel slippage is based on the slip ratio λ. i Continuously greater than the threshold λ th At the same time, the actual output current I is required. i Larger and the actual yaw rate of the vehicle w z It was much lower than expected, and the wheel speed was significantly higher than that of at least one other wheel.

4. The fault diagnosis and fault-tolerant control method for in-situ steering of distributed drive vehicles on slippery road surfaces according to claim 3, characterized in that, The torque distribution optimization model in step S05 adopts quadratic programming, and its optimization objective and constraints include: The objective function of the optimization objective is: Where T i T is the reconfigured output torque of the i-th healthy drive unit. i ref Let N be the initial desired torque of the drive unit, and N be the number of healthy drive units. The constraints are as follows: In the formula, H represents the total number of health-driving units, and F... xi F yi M represents the longitudinal and lateral forces generated by each drive unit. zi To correspond to the yaw moment; The output form is obtained by solving a quadratic programming problem to obtain the corrected torque {T} of each healthy drive unit that satisfies the constraints. i This is used to replace the original control commands and achieve fault-tolerant and stable control of distributed vehicles when turning in place on wet and slippery roads.

5. The fault diagnosis and fault-tolerant control method for in-situ steering of distributed drive vehicles on slippery roads according to claim 3, characterized in that, The drive failure or performance degradation is related to mechanical jamming or locked hardware malfunction. Once a drive unit is identified as faulty, drive commands to that unit are immediately stopped, and the desired drive torque T is forced to be applied. i ref The drive tasks originally assigned to the faulty wheel are reset to zero and then reassigned to all the remaining healthy wheels using an optimization algorithm.

6. The fault diagnosis and fault-tolerant control method for in-situ steering of distributed drive vehicles on slippery road surfaces according to claim 3, characterized in that, Excessive slippage requires gradually reducing the driving torque of the slipping wheel until the slip ratio returns to the optimal adhesion zone, at which point the wheel can provide the maximum effective driving torque of T. i,max =μ i F z,i R, where μ i F is the road surface adhesion coefficient. z，i Let R be the vertical load on the wheel and R be the wheel radius. If the wheel's slip ratio cannot return to the optimal adhesion zone within a given recovery time, it is temporarily set as a failure element, i.e., γ. i =0, and trigger the torque distribution optimization model, so that the other healthy wheels take over the driving task.

Images