CN-122008764-A - Active suspension pretightening control method, device and system based on signal game

Abstract

The invention provides an active suspension pretightening control method, device and system based on signal game, and relates to the technical field of vehicle suspension intelligent control, wherein the method comprises the steps of determining a first road grade index of a road based on pretightening sampling information; the method comprises the steps of determining a second road grade index of a road based on the road amplitude of the road, taking the smoothness index of a vehicle as a target, taking a signal game theory as a game theory guiding method, performing game decision on the first road grade index and the second road grade index to obtain an optimal road grade index of the road, determining an optimal weight coefficient combination corresponding to an LQR suspension controller based on the optimal road grade index, and determining optimal acting force based on the optimal weight coefficient combination. The method solves the problem that the apparent elevation value is different from the actual road excitation in the prior art and affects the suspension control effect, realizes the prediction of the elevation value of the front road of the vehicle, and effectively improves the smoothness of the vehicle.

Inventors

• WU XIAO
• WEI YINTAO
• LI ZHENGWEI
• TONG RUTING
• ZHANG JIAJUN

Assignees

• 清华大学

Dates

Publication Date

20260512

Application Date

20260323

Claims (10)

1. 1. An active suspension pretightening control method based on signal game is characterized by comprising the following steps: Determining a first road class index of the road based on the pre-aiming sampling information; Determining a second road class index for the road based on a road amplitude of the road, the road amplitude being used to characterize a vibrational response between a vehicle and the road; taking the smoothness index optimization of the vehicle as a target, taking a signal game theory as a game theory guiding method, and performing game decision on the first road grade index and the second road grade index to obtain an optimal road grade index of the road; determining an optimal weight coefficient combination corresponding to the LQR suspension controller based on the optimal road grade index; and determining optimal actuating force based on the optimal weight coefficient combination, wherein the optimal actuating force is used for determining state control parameters of a suspension pre-aiming control system of the vehicle.
2. 2. The method for controlling active suspension pretightening based on signal gaming according to claim 1, wherein said performing a gaming decision on said first road class index and said second road class index to obtain an optimal road class index for said road comprises: the method comprises the steps of determining a pre-aiming channel confusion matrix based on a first road grade observation signal, wherein the pre-aiming channel confusion matrix is obtained based on statistics and calibration of historical measurement errors of a laser radar under different environment illumination and road reflectivity conditions and is used for representing conditional probability distribution of observing the first road grade observation signal under the condition of determining a road real grade; The vibration channel confusion matrix is used for representing the conditional probability distribution that the second road class observation signal is observed to be the second road class index under the condition of determining the real road class and the environment state; determining a joint likelihood probability based on a first road class signal sequence, a second road class signal sequence, the pre-aiming channel confusion matrix and the vibration channel confusion matrix within a sliding time window; Based on Bayesian consistency, prior distribution and the joint likelihood probability, determining posterior distribution corresponding to the road real level, wherein the prior distribution is the prior distribution of the road real level and the environment state; And determining an optimal road grade index of the road based on a performance index function of the LQR algorithm and the posterior distribution, wherein the performance index function is used for representing expected cost of adopting actions in the real grade and environmental state of the road.
3. 3. The active suspension pre-aiming control method based on signal gaming according to claim 2, wherein determining joint likelihood probabilities based on a first road class signal sequence, a second road class signal sequence, the pre-aiming channel confusion matrix and the vibration channel confusion matrix within a sliding time window comprises: Determining target conditional probability distribution corresponding to each moment based on a pre-aiming channel confusion matrix corresponding to a first road class observation signal at each moment in the sliding time window and a vibration channel confusion matrix corresponding to a second road class observation signal at each moment; the target conditional probability distribution is used for representing the probability that the first road class observation signal is a first road class index and the second road class observation signal is a second road class index observed at each moment under the condition of determining the real road class and the environment state; And determining the joint likelihood probability based on the respective corresponding target conditional probability distribution of all the moments in the sliding time window.
4. 4. The method of claim 2, wherein determining the optimal road class index for the road based on the performance index function and the posterior distribution of the LQR algorithm comprises: determining a bayesian response based on the performance index function and the posterior distribution; And minimizing the Bayesian response to obtain the optimal road class index of the road.
5. 5. The method of active suspension pretightening control based on signal betting according to any one of claims 2 to 4, wherein said first road class observation signal and said second road class observation signal do not interfere with each other under the condition that a road true class and an environmental state are determined.
6. 6. The method for controlling active suspension pretightening based on signal gaming according to claim 1, wherein the determining the first road class index of the road based on the pretightening sampling information comprises: determining vertical coordinates and longitudinal coordinates of the road based on vehicle attitude information and the pre-aiming sampling information; carrying out frequency domain analysis on the vertical coordinate and the longitudinal coordinate to obtain a first reference power spectrum density corresponding to the road; Determining a first road class index for the road based on the first reference power spectral density; the determining a second road class index for the road based on the road amplitude for the road comprises: determining a road surface irregularity of the road based on the road amplitude and the longitudinal coordinates of the road; determining a second reference power spectral density corresponding to the road based on the road surface irregularities; a second road class index for the road is determined based on the second reference power spectral density.
7. 7. The method for controlling active suspension pretightening based on signal gaming according to claim 1, wherein the determining an optimal weight coefficient combination corresponding to the LQR suspension controller based on the optimal road class index comprises: The method comprises the steps of determining an optimal weight coefficient combination corresponding to an LQR suspension controller based on a weight combination incidence relation and the optimal road grade index, wherein the weight combination incidence relation is used for representing the incidence relation between a road grade and the weight coefficient combination, the weight combination incidence relation is determined based on power spectrum densities corresponding to a plurality of road grades, a two-degree-of-freedom suspension model and a multi-objective optimization algorithm, and the two-degree-of-freedom suspension model is a dynamics model of a vehicle suspension.
8. 8. The method of active suspension pretightening control based on signal gaming of claim 7, wherein said determining an optimal actuation force based on said optimal weight coefficient combination comprises: The method comprises the steps of inputting the optimal weight coefficient combination into an LQR suspension controller to obtain optimal acting force output by the LQR suspension controller, determining an LQR system matrix corresponding to a performance index function based on the optimal weight coefficient combination, determining a state feedback gain matrix based on a state space equation corresponding to the LQR system matrix and the two-degree-of-freedom suspension model, and determining the optimal acting force based on the state feedback gain matrix.
9. 9. An active suspension pretightening control device based on signal game, which is characterized by comprising: the first determining module is used for determining a first road grade index of the road based on the pre-aiming sampling information; A second determination module for determining a second road class index for the road based on a road amplitude of the road, the road amplitude being used to characterize a vibrational response between a vehicle and the road; The game decision module is used for performing game decision on the first road grade index and the second road grade index by taking the smoothness index of the vehicle as a target and taking a signal game theory as a game theory guiding method to obtain an optimal road grade index of the road; the suspension controller weight coefficient switching module is used for determining an optimal weight coefficient combination corresponding to the LQR suspension controller based on the optimal road grade index; And the suspension control algorithm module is used for determining optimal actuating force based on the optimal weight coefficient combination, and the optimal actuating force is used for determining state control parameters of a suspension pre-aiming control system of the vehicle.
10. 10. An active suspension pretightening control system based on signal gaming, comprising: The pre-aiming sensor is used for collecting pre-aiming sampling information corresponding to the road; a state observer for acquiring a road amplitude of the road; a processor for performing the active suspension pre-aiming control method based on signal gaming of any one of claims 1 to 8.

Description

Active suspension pretightening control method, device and system based on signal game Technical Field The invention relates to the technical field of intelligent control of vehicle suspensions, in particular to an active suspension pre-aiming control method, device and system based on signal gaming. Background The suspension is an important part for connecting wheels and a vehicle body of the vehicle, plays a role in buffering and damping, and influences the smoothness and the operation stability of the vehicle. In the prior art, on one hand, through high-precision sensors such as a laser radar or a binocular camera, a vehicle can accurately pretighten the elevation value of a front road surface of the vehicle, and the elevation value is used as the vertical excitation of the road, so as to try to solve the optimal control parameters of a suspension, and further effectively improve the smoothness and the operation stability of the vehicle. On the other hand, road excitation approaching a true value is determined in real time from the vibration response of the suspension by the state observer. However, the elevation value of the front road surface obtained by pre-aiming only reflects the apparent elevation value of the road, the road vertical excitation actually received by the tire is the result of the interaction between the road and the tire, the difference exists between the road vertical excitation and the apparent elevation value, and the apparent elevation value is directly used as the road vertical excitation to control the suspension, so that the actual control effect of the suspension can be influenced. In addition, the state observer is determined based on the historical vibration response, and is a reaction result of the suspension system to road excitation in time sequence, and the change of the elevation value of the front road of the vehicle cannot be predicted by calculating the vertical excitation of the road only through the state observation result. Therefore, the suspension control using the pre-aiming information has a prospective but low accuracy, and the suspension control using the state observation result is more accurate but has hysteresis, so that how to solve the two problems in the suspension control to improve the dynamics and the response performance of the vehicle system is a problem to be solved. Disclosure of Invention The invention provides an active suspension pretightening control method, device and system based on signal game, which are used for solving the problems that pretightening information is prospective but low in accuracy, and a state observation result is accurate but hysteresis exists in suspension control in the prior art. The invention provides an active suspension pretightening control method based on signal game, which comprises the following steps of. Determining a first road class index of the road based on the pre-aiming sampling information; Determining a second road class index for the road based on a road amplitude of the road, the road amplitude being used to characterize a vibrational response between a vehicle and the road; taking the smoothness index optimization of the vehicle as a target, taking a signal game theory as a game theory guiding method, and performing game decision on the first road grade index and the second road grade index to obtain an optimal road grade index of the road; determining an optimal weight coefficient combination corresponding to the LQR suspension controller based on the optimal road grade index; and determining an optimal actuating force based on the optimal weight coefficient combination, wherein the optimal actuating force is used for controlling the suspension of the vehicle. According to the active suspension pre-aiming control method based on the signal game provided by the invention, the game decision is carried out on the first road grade index and the second road grade index to obtain the optimal road grade index of the road, and the method comprises the following steps: the method comprises the steps of determining a pre-aiming channel confusion matrix based on a first road grade observation signal, wherein the pre-aiming channel confusion matrix is obtained based on statistics and calibration of historical measurement errors of a laser radar under different environment illumination and road reflectivity conditions and is used for representing conditional probability distribution of observing the first road grade observation signal under the condition of determining a road real grade; The vibration channel confusion matrix is used for representing the conditional probability distribution that the second road class observation signal is observed to be the second road class index under the condition of determining the real road class and the environment state; determining a joint likelihood probability based on a first road class signal sequence, a second road class signal sequence, the pre-aiming channel c