CN-122018311-A - Vehicle formation MPC control method integrating asymmetric response model and delay compensation

CN122018311ACN 122018311 ACN122018311 ACN 122018311ACN-122018311-A

Abstract

The invention relates to the technical field of intelligent traffic and automatic driving control, and discloses a vehicle formation MPC control method integrating an asymmetric response model and delay compensation, which comprises the steps of firstly establishing an asymmetric vehicle longitudinal dynamics model for distinguishing acceleration and deceleration characteristics; and if the acceleration working condition is judged, adopting model parameters containing pure time delay, and utilizing a state augmentation technology to bring historical control input into a state vector to construct an augmented discrete state space model without delay. On the basis, the quadratic programming problem comprising multiple cost and soft constraint punishment is solved by combining acceleration disturbance information of the front vehicle and the head vehicle, and an optimal control instruction is obtained. In addition, a hierarchical safety redundancy strategy when solving failure is designed. The invention can accurately match the physical characteristic difference of the vehicles, effectively compensate the influence of braking delay and improve the tracking precision, stability and safety of formation control.

Inventors

GONG JIANWEI
CHEN XINRAN
JU ZHIYANG
QI JIANYONG
SUN BOFAN

Assignees

北京理工大学

Dates

Publication Date: 20260512
Application Date: 20251229

Claims (10)

1. A vehicle formation MPC control method integrating an asymmetric response model and delay compensation is characterized by comprising the following steps: S1, establishing a vehicle longitudinal dynamics model, wherein the vehicle longitudinal dynamics model comprises an acceleration longitudinal dynamics model describing acceleration characteristics and a deceleration longitudinal dynamics model describing deceleration characteristics; S2, constructing a formation control state space model integrating the longitudinal dynamics model of the vehicle, and representing dynamic association of formation error state quantity and control input; s3, acquiring real-time driving state data at least comprising the longitudinal acceleration of the vehicle; S4, comparing the longitudinal acceleration of the vehicle with a preset threshold value, determining a discretization basic model based on the formation control state space model, and selecting corresponding parameters of the acceleration longitudinal dynamics model if the acceleration working condition is judged, and selecting corresponding parameters of the deceleration longitudinal dynamics model if the deceleration working condition is judged; S5, carrying out delay processing on the discretization basic model, and introducing historical control input by using a state augmentation technology to construct an augmented discrete state space model without delay if the discretization basic model corresponds to a deceleration working condition; S6, substituting the augmented discrete state space model or the discretization basic model into a model prediction control solver, solving a quadratic programming problem comprising an objective function and constraint conditions, and sending the first element of the obtained optimal control sequence to an executor.
2. The vehicle formation MPC control method of claim 1, wherein the MPC control method is used for integrating an asymmetric response model and delay compensation, the method is characterized in that in the step S1: the acceleration longitudinal dynamics model is constructed as a first-order inertial system without input delay, and the characteristic parameters comprise an acceleration time constant; The deceleration longitudinal dynamics model is constructed as a first-order inertial system with pure time delay, and the characteristic parameters comprise a deceleration time constant and the pure time delay; the value of the deceleration time constant is greater than the acceleration time constant.
3. The vehicle formation MPC control method of claim 1, wherein said forming error state quantity characterized by said forming control state space model in step S2 comprises: the distance error is the difference between the actual distance between the vehicle and the front vehicle and the expected distance, and the expected distance is calculated based on a constant time distance strategy; the relative speed is the speed difference value between the vehicle and the front vehicle; the relative speed of the head car is the speed difference value between the head car and the head car; Actual acceleration of the vehicle.
4. The vehicle formation MPC control method of claim 1, wherein the determining logic in step S4 is: If the longitudinal acceleration of the vehicle is greater than or equal to the preset threshold value, judging an acceleration working condition, wherein the discretization basic model does not contain an input delay item; And if the longitudinal acceleration of the vehicle is smaller than the preset threshold value, judging a deceleration working condition, wherein the discretization basic model comprises an input delay item to be processed.
5. The vehicle formation MPC control method of claim 1, wherein said step S5 comprises using a state augmentation technique comprising: Quantifying the pure time delay in the deceleration longitudinal dynamics model to an integer multiple of a delay step number of a sampling period; Defining an augmented state vector, wherein the augmented state vector consists of the formation error state quantity and historical control inputs of a plurality of control periods in the past, and the number of the historical control inputs corresponds to the delay step number; reconstructing a system evolution matrix, introducing the history control input into a state update equation, and establishing a delay-free evolution relationship based on the augmented state vector.
6. The vehicle-formed MPC control method of claim 1 wherein said objective function in step S6 comprises the following weighted cost terms: the state tracking cost is used for punishing deviation between the predicted state and the reference state; the control quantity cost is used for punishing the magnitude of the control input; The control increment cost is used for punishing the change rate of control input between adjacent control periods; a soft constraint penalty for punishing the magnitude of a relaxation variable that is used to handle situations where hard constraints are not feasible.
7. The vehicle formation MPC control method of claim 6, wherein said constraint conditions include inter-vehicle soft constraints constructed by: assigning a non-negative relaxation variable to a prediction step in a prediction time domain, limiting the degree to which the inter-vehicle distance is allowed to be smaller than a safety distance lower bound to the relaxation variable, and applying a penalty to the relaxation variable through the soft constraint penalty cost in the objective function, thereby minimizing the degree of violation on the premise of ensuring solving feasibility.
8. The vehicle formation MPC control method of claim 1, further comprising a first level of safety redundancy strategy: When the quadratic programming problem fails to be solved in the current control period and the continuous failure times do not reach the preset upper limit, acquiring an optimal control sequence obtained by successfully solving the previous control period; and selecting a second element in the optimal control sequence of the previous control period as an instruction of the current control period and sending the instruction to an executor.
9. The vehicle formation MPC control method integrating an asymmetric response model and delay compensation as claimed in claim 1, wherein the method comprises a second-level safety redundancy strategy: when the number of continuous solving failures of the quadratic programming problem reaches or exceeds the preset upper limit, judging that the controller fails; and generating a preset emergency braking instruction, sending the emergency braking instruction to an actuator, and forcibly resetting the internal state of the model predictive control solver so as to reinitialize calculation in the next control period.
10. The vehicle formation MPC control method of claim 1, wherein the prediction equation according to which the quadratic programming problem is constructed in step S6 comprises disturbance response terms: The predictive equation represents the state sequence at a future time as a linear combination of the free response of the current state, the forced response of the future control input, and the forced response of the measurable disturbance; The measurable disturbance includes a lead vehicle acceleration and a head vehicle acceleration acquired through inter-vehicle communication, and is regarded as a known input quantity to participate in the operation in a prediction time domain of model predictive control.

Description

Vehicle formation MPC control method integrating asymmetric response model and delay compensation Technical Field The invention relates to the technical field of intelligent traffic and automatic driving control, in particular to a vehicle formation MPC control method integrating an asymmetric response model and delay compensation. Background The vehicle formation driving is used as an important component of the intelligent traffic system, and the wind resistance can be reduced and the road passing efficiency can be improved by shortening the vehicle distance. Among the formation control algorithms, model Predictive Control (MPC) is widely used for vehicle longitudinal follow control because it can explicitly handle multi-objective optimization and system constraints. The performance of the MPC controller is highly dependent on the accuracy of the predictive model, i.e., whether the model accurately reflects the dynamic response characteristics of the vehicle during actual travel. However, existing vehicle longitudinal control studies mostly use a uniform linear model to describe vehicle dynamics, assuming that the response characteristics of the vehicle during acceleration and deceleration are consistent. In fact, the longitudinal dynamics of the vehicle have an asymmetry. The acceleration process is mainly driven by power assemblies such as an engine, a motor, a gearbox and the like, the response characteristics of the acceleration process are mainly influenced by mechanical inertia, the deceleration process depends on a hydraulic or pneumatic braking system, and besides inertia hysteresis, pure time delay caused by pipeline transmission and pressure building processes exists. It is difficult for a conventional single model to simultaneously compromise these two distinct physical characteristics, resulting in model parameters that are typically only calibrated for a particular operating condition. Furthermore, conventional linear MPC methods typically ignore or reduce the pure time delay of the brake system as a first order inertial link when dealing with it. This simplification may allow the controller to misunderstand that the vehicle is immediately responsive to the braking command when the vehicle is in a deceleration condition, thereby producing a predicted deviation. Mismatch between this model and the actual physical object can cause the controller to overshoot in the actual application. Particularly in a formation driving scene with smaller inter-vehicle distance, uncompensated braking delay can reduce the phase margin of a system, cause post-vehicle reaction lag, further cause increase of tracking error and reduction of riding comfort, and even cause formation instability or rear-end collision accidents when serious. Disclosure of Invention Aiming at the defects of the prior art, the invention provides a vehicle formation MPC control method integrating an asymmetric response model and delay compensation, which solves the problems of large model prediction deviation, control instruction oscillation and poor formation running stability caused by that a single linear model cannot accurately describe the acceleration and deceleration dynamics asymmetric characteristic of a vehicle and the pure time delay of a brake system is ignored in the existing vehicle formation control. In order to achieve the above purpose, the invention provides a vehicle formation MPC control method integrating an asymmetric response model and delay compensation, which comprises the steps of firstly establishing a vehicle longitudinal dynamics model. Based on the physical characteristic difference between the vehicle power assembly and the braking system, the model adopts an asymmetric structure, and comprises an acceleration longitudinal dynamics model describing acceleration characteristics and a deceleration longitudinal dynamics model describing deceleration characteristics. The acceleration longitudinal dynamics model is constructed as a first-order inertial system without pure time delay, and the characteristic parameters comprise an acceleration time constant. The longitudinal dynamics model of deceleration is constructed as a first-order inertial system with pure time delay so as to represent the hysteresis characteristic in the brake pressure establishment process, wherein the characteristic parameters comprise a deceleration time constant and the pure time delay, and the value of the deceleration time constant is larger than the acceleration time constant. On the basis of establishing a dynamics model, the invention constructs a formation control state space model. The model integrates the vehicle longitudinal dynamics model into a system state evolution equation, and represents the dynamic correlation of formation error state quantity and control input. The formation error state quantity is used for selecting a distance error, a relative speed of a head car and an actual acceleration of the head car, wherein the di