Active suspension preview control method, device and system based on signal game

Abstract

The invention provides an active suspension preview control method, device and system based on a signal game, and relates to the technical field of vehicle suspension intelligent control, and the method comprises the steps: determining a first road grade index of a road based on preview sampling information; determining a second road grade index of the road based on the road amplitude of the road; by taking the optimal smoothness index of the vehicle as a target and taking a signal game theory as a game theory guidance 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; based on the optimal road grade index, determining an optimal weight coefficient combination corresponding to the LQR suspension controller; and determining the optimal action force based on the optimal weight coefficient combination. The problem that in the prior art, the control effect of the suspension is affected when the apparent elevation value and actual road excitation are different is solved, prediction of the elevation value of the road in front of the vehicle is achieved, and the smoothness of the vehicle is effectively improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent vehicle suspension control technology, and in particular to an active suspension anti-aiming control method, device and system based on signal game theory. Background Technology

[0002] The suspension is a crucial component connecting the wheels and body of a car, acting as a buffer and damper to affect ride comfort and handling stability. Current technologies employ two main approaches: First, using high-precision sensors such as lidar or binocular cameras, the vehicle can accurately predict the elevation of the road surface ahead, using this as a vertical road excitation to determine the optimal control parameters for the suspension, thereby effectively improving ride comfort and handling stability. Second, a state observer can determine near-real-world road excitation in real time based on the suspension's vibration response.

[0003] However, the elevation value of the road surface in front of the vehicle obtained by pre-aiming only reflects the apparent elevation value of the road. The actual vertical excitation of the road to the tires is the result of the interaction between the road and the tires, which differs from the apparent elevation value. Directly using the apparent elevation value as the vertical road excitation to control the suspension will affect the actual control effect of the suspension. In addition, the state observer is determined based on historical vibration response, which is the result of the suspension system's response to road excitation in time. Calculating the vertical road excitation solely based on the state observation results cannot predict changes in the road elevation value in front of the vehicle.

[0004] Therefore, using preview information for suspension control is forward-looking but has low accuracy, while using state observation results for suspension control is more accurate but has lag. How to solve these two problems in suspension control to improve the dynamic characteristics and response performance of the vehicle system is an urgent problem to be solved. Summary of the Invention

[0005] This invention provides an active suspension anti-aiming control method, device, and system based on signal game theory, which solves the problems in the prior art where anti-aiming information is forward-looking but has low accuracy, and state observation results are more accurate but have lag.

[0006] This invention provides an active suspension anti-aiming control method based on signal game theory, comprising the following steps.

[0007] Based on the pre-sampling information, the first road grade index of the road is determined; A second road grade index is determined based on the road amplitude; the road amplitude is used to characterize the vibration response between vehicles and the road. With the goal of optimizing the ride comfort index of the vehicle, and using signal game theory as the game theory guidance method, a game decision is made on the first road grade index and the second road grade index to obtain the optimal road grade index of the road. Based on the optimal road grade index, determine the optimal weight coefficient combination corresponding to the LQR suspension controller; Based on the optimal combination of weighting coefficients, the optimal actuation force is determined, and the optimal actuation force is used to control the suspension of the vehicle.

[0008] According to the active suspension anti-aiming control method based on signal game theory provided by the present invention, the step of performing a game decision on the first road grade index and the second road grade index to obtain the optimal road grade index of the road includes: The aiming channel confusion matrix is ​​determined based on the first road level observation signal. The aiming channel confusion matrix is ​​obtained by statistical calibration of historical measurement errors of lidar under different ambient light and road surface reflectivity conditions. It is used to characterize the conditional probability distribution of the first road level observation signal under the condition of determining the true road level. The vibration channel confusion matrix is ​​determined based on the second road grade observation signal; the vibration channel confusion matrix is ​​used to characterize the conditional probability distribution of the observed second road grade observation signal being the second road grade index under the condition of determining the true road grade and environmental state; The joint likelihood probability is determined based on the first road level signal sequence, the second road level signal sequence, the pre-aiming channel confusion matrix, and the vibration channel confusion matrix within the sliding time window. Based on Bayesian consistency, prior distribution, and the joint likelihood probability, the posterior distribution corresponding to the true road level is determined; the prior distribution is the prior distribution of the true road level and the environmental state. Based on the performance index function of the LQR algorithm and the posterior distribution, the optimal road grade index of the road is determined; the performance index function is used to characterize the expected cost of taking an action under the actual road grade and environmental conditions.

[0009] According to the active suspension anti-aiming control method based on signal game theory provided by the present invention, the determination of joint likelihood probability based on the first road level signal sequence, the second road level signal sequence, the anti-aiming channel confusion matrix, and the vibration channel confusion matrix within a sliding time window includes: Based on the pre-aiming channel confusion matrix corresponding to the first road level observation signal at each time point within the sliding time window, and the vibration channel confusion matrix corresponding to the second road level observation signal at each time point, the target conditional probability distribution corresponding to each time point is determined; the target conditional probability distribution is used to characterize the probability that, under the condition of determining the true road level and environmental state, the first road level observation signal is observed to be the first road level index and the second road level observation signal is observed to be the second road level index at each time point. The joint likelihood probability is determined based on the target conditional probability distribution corresponding to each time point within the sliding window.

[0010] According to the active suspension anti-aiming control method based on signal game theory provided by the present invention, the determination of the optimal road grade index of the road based on the performance index function of the LQR algorithm and the posterior distribution includes: The Bayesian response is determined based on the performance index function and the posterior distribution. Minimize the Bayesian response to obtain the optimal road grade index for the road.

[0011] According to the active suspension anti-aiming control method based on signal game theory provided by the present invention, under the condition that the actual road level and environmental conditions are determined, the first road level observation signal and the second road level observation signal do not interfere with each other.

[0012] According to the active suspension anti-aiming control method based on signal game theory provided by the present invention, determining the first road grade index based on anti-aiming sampling information includes: Based on the vehicle attitude information and the pre-aiming sampling information, the vertical and longitudinal coordinates of the road are determined; Frequency domain analysis is performed on the vertical and longitudinal coordinates to obtain the first reference power spectral density corresponding to the road. Based on the first reference power spectral density, a first road grade index is determined for the road. The determination of the second road grade index based on the road amplitude includes: The road surface roughness is determined based on the road amplitude and the longitudinal coordinates of the road. Based on the road surface roughness, the second reference power spectral density corresponding to the road is determined; The second road grade index of the road is determined based on the second reference power spectral density.

[0013] According to the active suspension anti-aiming control method based on signal game theory provided by the present invention, determining the optimal weight coefficient combination corresponding to the LQR suspension controller based on the optimal road grade index includes: Based on the weight combination correlation and the optimal road grade index, the optimal weight coefficient combination corresponding to the LQR suspension controller is determined; the weight combination correlation is used to characterize the correlation between road grade and weight coefficient combination; and the weight combination correlation is determined based on the power spectral density corresponding to multiple road grades, the two-degree-of-freedom suspension model and the multi-objective optimization algorithm; the two-degree-of-freedom suspension model is the dynamic model of the vehicle suspension.

[0014] According to the active suspension anti-aiming control method based on signal game theory provided by the present invention, determining the optimal actuation force based on the optimal weight coefficient combination includes: The optimal weight coefficient combination is input into the LQR suspension controller to obtain the optimal force output by the LQR suspension controller; the LQR suspension controller is used to determine the LQR system matrix corresponding to the performance index function based on the optimal weight coefficient combination; based on the LQR system matrix and the state space equation corresponding to the two-degree-of-freedom suspension model, the state feedback gain matrix is ​​determined; based on the state feedback gain matrix, the optimal force is determined.

[0015] The present invention also provides an active suspension anti-aiming control device based on signal game theory, comprising the following modules.

[0016] The first determining module is used to determine the first road grade index of the road based on the pre-sampling information.

[0017] The second determining module is used to determine a second road grade index of the road based on the road amplitude; the road amplitude is used to characterize the vibration response between the vehicle and the road.

[0018] The game theory decision module is used to make game decisions on the first road grade index and the second road grade index with the goal of optimizing the smoothness index of the vehicle, and using signal game theory as the game theory guidance method, to obtain the optimal road grade index of the road.

[0019] The suspension controller weight coefficient switching module is used to determine the optimal weight coefficient combination corresponding to the LQR suspension controller based on the optimal road grade index.

[0020] The suspension control algorithm module is used to determine the optimal action force based on the optimal weight coefficient combination, and the optimal action force is used to control the suspension of the vehicle.

[0021] The present invention also provides an active suspension anti-aiming control system based on signal game theory, comprising: Preview sensor, used to collect preview sampling information corresponding to the road; A state observer is used to collect the road vibration amplitude of the road. A processor for executing the active suspension anti-aiming control method based on signal game theory as described above.

[0022] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the active suspension anti-aiming control method based on signal game theory as described above.

[0023] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the active suspension anti-aiming control method based on signal game theory as described above.

[0024] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the active suspension anti-aiming control method based on signal game theory as described above.

[0025] The present invention provides an active suspension preview control method, device, and system based on signal game theory. It determines a first road grade index by preview sampling information, and a second road grade index by road amplitude, which characterizes the vibration response between the vehicle and the road. With the optimal vehicle ride comfort index as the objective, and using signal game theory as the guiding method, it performs game-theoretic decision-making on the first and second road grade indices to obtain the optimal road grade index. Based on the optimal road grade index, it determines the optimal weight coefficient combination corresponding to the LQR suspension controller, and based on the optimal weight coefficient combination, determines the optimal action force for vehicle suspension control. In this invention, a first road grade index with forward-looking but low accuracy from a pre-aiming sensor and a second road grade index with accurate but lagging accuracy from a state observer are used as two independent observation signals. Using signal game theory as a guiding method, a game-theoretic decision is made between the two observation signals to obtain the optimal road grade index that optimizes vehicle ride comfort. This optimal road grade index serves as a reference for the vehicle suspension switching weight system. This solves the problem in the prior art where the apparent elevation value and the actual road excitation are not simultaneously affecting the suspension control effect, enabling the prediction of the road elevation value in front of the vehicle and effectively improving vehicle ride comfort. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0027] Figure 1This is a flowchart illustrating the active suspension anti-aiming control method based on signal game theory provided in an embodiment of the present invention.

[0028] Figure 2 This is a schematic diagram of a two-degree-of-freedom suspension model provided in an embodiment of the present invention.

[0029] Figure 3 This is a schematic diagram of the active suspension anti-aiming control device based on signal game theory provided in an embodiment of the present invention.

[0030] Figure 4 This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0031] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0032] To address the issues in existing suspension control technologies where pre-aiming information offers foresight but low accuracy, and state observation results provide greater accuracy but suffer from lag, this invention provides an active suspension pre-aiming control method based on signal game theory. Figure 1 This is a flowchart illustrating the active suspension anti-aiming control method based on signal game theory provided in an embodiment of the present invention, as shown below. Figure 1 As shown, the method includes steps 110 to 150.

[0033] Step 110: Based on the pre-sampling information, determine the first road grade index of the road.

[0034] Specifically, the preview sampling information is data collected by preview sensors, used to describe the apparent elevation value of the road surface ahead of the vehicle that has not yet been driven. Based on this preview sampling information, a first road grade index of the road ahead is calculated using a non-contact measurement method. This first road grade index is used to characterize the road surface unevenness of the road ahead as determined by the preview sampling information.

[0035] Optionally, the aiming sensor can be a lidar, and the corresponding aiming sampling information can be a depth value. This embodiment of the invention does not limit this.

[0036] Step 120: Determine the second road grade index of the road based on the road amplitude; the road amplitude is used to characterize the vibration response between the vehicle and the road.

[0037] Specifically, a suspension state observer is constructed to obtain the road amplitude by detecting the vibration response between the vehicle and the road. This road amplitude is used to characterize the vertical road excitation exerted by the road on the vehicle tires. Based on this road amplitude, a second road grade index is calculated, which is used to characterize the road surface roughness determined based on the vibration response.

[0038] Alternatively, a Kalman filter can be used as a suspension state observer, but this embodiment of the invention does not limit this.

[0039] Step 130: With the goal of optimizing the ride comfort index of the vehicle, and using signal game theory as the guiding method, perform game decision-making on the first road grade index and the second road grade index to obtain the optimal road grade index of the road.

[0040] Specifically, Signaling Game theory is a key model in game theory for studying information transmission mechanisms under conditions of information asymmetry, belonging to the category of dynamic games with incomplete information. The core of Signaling Game theory lies in the strategic interaction between the sender and receiver; that is, the sender possesses private information (such as product quality and individual capabilities) and influences the receiver's decisions by sending observable "signals"; the receiver then takes the optimal action based on the received signals. The suspension anti-signaling control process closely aligns with this, where road unevenness excitation is transmitted through the tires to the suspension system, causing vehicle vibration, forming a typical master-slave game structure. The road acts as the signal sender, and the suspension control system as the receiver. The suspension control system needs to make optimal judgments based on limited road perception information to improve the system's dynamic characteristics. Based on this correspondence, in this embodiment of the invention, signal game theory is introduced as a guide, and the control of the suspension is regarded as a game process between the vehicle and the road. With the optimal smoothness index of the vehicle as the goal, according to the actual vehicle control process, the feasible strategy set of the vehicle is determined as the first road grade index obtained by the pre-aiming sensor and the second road grade index obtained by the state observer. The feasible strategy set of the road is the commonly used road grade index, and the optimal road grade index of the road under game theory is obtained.

[0041] Step 140: Based on the optimal road grade index, determine the optimal weight coefficient combination corresponding to the LQR suspension controller.

[0042] Specifically, after determining the optimal road grade index, the optimal road grade index is input into the LQR suspension controller to calculate the optimal weight coefficient combination, thereby maintaining or improving the suspension ride comfort index.

[0043] Step 150: Based on the optimal weight coefficient combination, determine the optimal actuation force, which is used to control the suspension of the vehicle.

[0044] Specifically, after determining the optimal combination of weight coefficients, the optimal action force is calculated based on the optimal combination of weight coefficients. The optimal action force is the specific physical control force applied to the suspension. By changing the vibration output of the suspension control system through the optimal action force, the control of the suspension is completed.

[0045] The active suspension preview control method based on signal game theory provided in this invention determines a first road grade index by preview sampling information, and a second road grade index by road amplitude, which characterizes the vibration response between the vehicle and the road. With the optimal vehicle ride comfort index as the objective, and using signal game theory as the guiding method, a game decision is made between the first and second road grade indices to obtain the optimal road grade index. Based on the optimal road grade index, the optimal weight coefficient combination corresponding to the LQR suspension controller is determined, and the optimal action force is determined based on the optimal weight coefficient combination to control the vehicle suspension. In this invention, a first road grade index with forward-looking but low accuracy from a pre-aiming sensor and a second road grade index with accurate but lagging accuracy from a state observer are used as two independent observation signals. Using signal game theory as a guiding method, a game-theoretic decision is made between the two observation signals to obtain the optimal road grade index that optimizes vehicle ride comfort. This optimal road grade index serves as a reference for the vehicle suspension switching weight system. This solves the problem in the prior art where the apparent elevation value and the actual road excitation are not simultaneously affecting the suspension control effect, enabling the prediction of the road elevation value in front of the vehicle and effectively improving vehicle ride comfort.

[0046] In one embodiment, determining the first road grade index based on the pre-sampling information includes: Based on the vehicle attitude information and the pre-aiming sampling information, the vertical and longitudinal coordinates of the road are determined; Frequency domain analysis is performed on the vertical and longitudinal coordinates to obtain the first reference power spectral density corresponding to the road. Based on the first reference power spectral density, the first road grade index of the road is determined.

[0047] Specifically, taking a lidar as the pre-aiming sensor as an example, the depth value of the road sampling point in front of the vehicle is measured by the lidar. This depth value is the pre-aiming sampling information. Then, the vehicle attitude information can be obtained based on the vehicle's inertial measurement unit. Then, using a coordinate system transformation algorithm, the vertical and longitudinal coordinates of the road are calculated based on the depth value and the vehicle attitude information, thereby transforming the depth value of the road sampling point in front of the vehicle to a unified vehicle coordinate system or geodetic coordinate system. Then, frequency domain analysis is performed on the vertical and longitudinal coordinates to obtain the first reference power spectral density corresponding to the road. Using equation (1), the first road grade index of the road is calculated based on the first reference power spectral density. Equation (1) is: .

[0048] in, Represents the first reference power spectral density. The road grade index is calculated based on the first reference power spectral density. This is the first road grade index. .

[0049] Optionally, the vehicle attitude information may include pitch angle and elevation angle, etc., and the coordinate system transformation algorithm may be to construct a rotation matrix based on Euler angles or quaternions to perform coordinate transformation. This embodiment of the invention does not limit this.

[0050] It should be noted that if only the preview result is used to determine the optimal weight coefficient combination of the LQR suspension controller, if the first road grade index is large, that is, the road grade index of the actual road excitation is smaller than the first road grade index corresponding to the preview sensor, the root mean square value of the suspension body acceleration is smaller than the ideal control result, but the suspension dynamic deflection and power consumption are larger. Conversely, if the first road grade index is small, the root mean square value of the suspension body acceleration is larger, but the suspension dynamic deflection and power consumption are smaller.

[0051] In one embodiment, determining the second road grade index of the road based on the road amplitude includes: The road surface roughness is determined based on the road amplitude and the longitudinal coordinates of the road. Based on the road surface roughness, the second reference power spectral density corresponding to the road is determined; The second road grade index of the road is determined based on the second reference power spectral density.

[0052] Specifically, a Kalman filter is used as the suspension state observer. After the road and tires interact and generate a vibration response, this vibration response is input into the state observer. The state observer converts the vibration response into road amplitude. Combining this road amplitude with the longitudinal coordinates of the vehicle during travel, the road surface roughness can be calculated. Frequency domain analysis is performed on this road surface roughness to obtain the second reference power spectral density of the road. Substituting this second reference power spectral density into equation (1) yields the second road grade index of the road. .

[0053] In one embodiment, the step of performing a game-theoretic decision on the first road grade index and the second road grade index to obtain the optimal road grade index for the road includes: The aiming channel confusion matrix is ​​determined based on the first road level observation signal. The aiming channel confusion matrix is ​​obtained by statistical calibration of historical measurement errors of lidar under different ambient light and road surface reflectivity conditions. It is used to characterize the conditional probability distribution of the first road level observation signal under the condition of determining the true road level. The vibration channel confusion matrix is ​​determined based on the second road grade observation signal; the vibration channel confusion matrix is ​​used to characterize the conditional probability distribution of the observed second road grade observation signal being the second road grade index under the condition of determining the true road grade and environmental state; The joint likelihood probability is determined based on the first road level signal sequence, the second road level signal sequence, the pre-aiming channel confusion matrix, and the vibration channel confusion matrix within the sliding time window. Based on Bayesian consistency, prior distribution, and the joint likelihood probability, the posterior distribution corresponding to the true road level is determined; the prior distribution is the prior distribution of the true road level and the environmental state. Based on the performance index function of the LQR algorithm and the posterior distribution, the optimal road grade index of the road is determined; the performance index function is used to characterize the expected cost of taking an action under the actual road grade and environmental conditions.

[0054] Specifically, after determining the first road grade index and the second road grade index, the specific vehicle-road signal game process includes the following steps.

[0055] 1. Using signal game theory as the guiding method of game theory, a signal game framework is constructed by building a set of participants, types, signals and actions in a vehicle-road signal game.

[0056] (1) Set the road as the sender and the vehicle as the receiver.

[0057] (2) Assume the actual road grade is the sender type. The set of possible values ​​for the actual road grade is as follows: ,in, The set of values ​​representing the true road grade. This represents the total number of road grades, determined based on the range of road grades commonly used by the vehicle. For example, taking an urban passenger car as an example, the range of commonly used road grades for this vehicle could include [0,1,2,3,4,5,6,7,8,9,10], belonging to ISO 8608 Class A to Class D roads. Specifically, the road grade range [0,1,2,3,4,5] belongs to ISO 8608 Class A roads, the road grade range (5,6,7) belongs to ISO 8608 Class B roads, the road grade range (7,8,9) belongs to ISO 8608 Class C roads, and the road grade range (9,10) belongs to ISO 8608 Class D roads. In this case, N=11.

[0058] (3) To characterize the influence of the external environment, implicit environmental states are introduced. The set of values ​​for this environmental state is ,in, The number of environmental states is represented by, for example, the external environment is limited to four types of phenomena: ambient light, sensor imaging quality, tire envelope effect, and the inconsistency between apparent elevation and actual excitation caused by "sandwich road surface". Then W=4.

[0059] (4) Assume that the signal sent from the road to the vehicle consists of two types of observation signals, namely the first road level observation signal obtained by the aiming sensor. And the second road level observation signal obtained based on the suspension vibration response. ,in: , ,in, This represents the total number of road classes determined through preview observations. L represents the total number of road grades determined by vibration response. Taking the commonly used road grade range for vehicles as [0,1,2,3,4,5,6,7,8,9,10] as an example, then L V =L E =1,S V S represents the set of observation signals for the first road level. E This represents the set of observation signals for the second road level.

[0060] (5) Let the set of actions of the vehicle be Each action Indicating the weighted combination relationship, the first... Combination of weighting coefficients Or the corresponding LQR system matrix and state feedback gain matrix The vehicle's game objective is to optimize the ride comfort index under uncertain road grades.

[0061] 2. Given the actual type of road With environmental conditions Under these conditions, both the first-level and second-level road observation signals are random variables and can be represented by conditional probabilities. Preferably, a confusion matrix is ​​used to model the discrete-level observation signals. This pre-aiming channel confusion matrix can be expressed as: , Indicates the environmental state. , Each element in the pre-targeting channel confusion matrix is ​​used to characterize the environmental state. And the actual road grade is Under these conditions, the pre-aiming sensor outputs the observation signal d. V The conditional probability.

[0062] The confusion matrix of this vibration channel can be represented as: , Each element in this vibration channel confusion matrix is ​​used to characterize the environmental state. And the actual road grade is Under the given conditions, the state observer outputs the observation signal d. E The conditional probability.

[0063] It should be noted that the aiming channel confusion matrix and the vibration channel confusion matrix can be obtained through simulation data calibration, bench test calibration or actual vehicle data statistics, and in the two confusion matrices, the external environmental influencing factors are uniformly equivalent to statistical deviations.

[0064] 3. To avoid interference from noise at a single moment in decision-making, this embodiment of the invention utilizes historical observations to enhance road grade discrimination capabilities. Within the control period, a sliding time window of length T is used. All historical first road grade observation signals observed within the sliding time window are arranged in chronological order to obtain a first road grade signal sequence. This first road grade signal sequence is represented as follows: , This indicates the first-level road signal sequence. Let represent the first road level observation signal at time t, where t represents the upper limit of the sliding time window, and t-T+1 represents the lower limit of the sliding time window. Simultaneously, all historical second road level observation signals observed within the sliding time window are arranged in chronological order to obtain the second road level signal sequence, which is represented as follows: , This indicates the second-level road signal sequence. Let represent the second road level observation signal at time t. Then, the window-independent likelihood assumption is adopted, i.e., under environmental conditions of . And the actual road grade is Under these conditions, the observed signals at different times are approximately independent within the time window. Based on this, the joint likelihood probability is calculated using the first road level signal sequence, the second road level signal sequence, the aiming channel confusion matrix, and the vibration channel confusion matrix within the sliding time window.

[0065] It should be noted that the sliding time window can be adaptively set by vehicle speed, aiming distance and control update frequency, and the embodiments of the present invention do not limit this.

[0066] 4. After determining the joint likelihood probability and observing the first and second road level signal sequences at time t, the posterior distribution corresponding to the true road level is determined using equation (2) based on Bayesian consistency, prior distribution, and joint likelihood probability. This posterior distribution indicates that the true road level is determined after fusing the preview information and vibration response. External environmental influencing factors are The confidence level at that time is given by equation (2): .

[0067] in, Represents any value in the set of possible values ​​for the true road grade. Represents any value in the set of possible environmental states. Denotes the prior distribution, and .

[0068] 5. After determining the posterior distribution, the vehicle, as the receiver, selects an action based on the posterior probability. (That is, selecting the corresponding LQR weighting coefficient combination) to optimize the smoothness index. Therefore, it is defined based on real road types. With environmental conditions The following actions are adopted. The performance index function based on the LQR algorithm is derived. This performance index function can be composed of suspension performance indices, such as a weighted quadratic form of sprung acceleration, suspension dynamic deflection, and tire dynamic load, and can be pre-calculated through offline simulation or experimentation. Based on this posterior distribution and the performance index function, the optimal road grade index is determined.

[0069] It should be noted that before determining the optimal road grade index, based on the linear quadratic regulator (LQR), the LQR system matrix Q, N, R and the state feedback gain matrix K are designed to establish the LQR suspension controller, thereby constructing the calculation formula for the optimal action force.

[0070] A performance index function based on the LQR algorithm is constructed, as shown in equation (3). Equation (3) is: .

[0071] in, Represents a performance metric function. This represents the suspension dynamic deflection weighting coefficient. This represents the weighting coefficient for vehicle body acceleration. This represents the tire travel weighting coefficient.

[0072] The core strategy of the LQR algorithm is to minimize the performance index function. Since the scope is from zero to positive infinity, the suspension dynamic deflection, vehicle acceleration, and tire dynamic deformation of the suspension system are negatively correlated with their weight coefficients. By reasonably assigning weight coefficients, the performance of the car under specific road surface inputs can be improved.

[0073] In order to make the performance index function of the LQR algorithm applicable to the state space equations corresponding to the two-degree-of-freedom suspension model, the performance index function can be rewritten in matrix form, as shown in equation (4). Equation (4) is: .

[0074] in, , , .

[0075] In one embodiment, given that the true road grade and environmental conditions are determined, the first road grade observation signal and the second road grade observation signal do not interfere with each other.

[0076] Specifically, to simplify online calculations, the environment is set to... And the actual road grade is Under these conditions, the observation signals of the first road level and the second road level do not interfere with each other. Based on this, it can be determined that... That is, in the environmental state of And the actual road grade is Under these conditions, the pre-aiming sensor outputs the first road level observation signal as d. V And the state observer outputs the second road level observation signal as d. E The conditional probability is that the first road level observation signal output by the pre-aiming sensor under this condition is d. V The conditional probability and the second road level observation signal output by the state observer under this condition are d E The product of conditional probabilities.

[0077] In one embodiment, determining the joint likelihood probability based on the first road grade signal sequence, the second road grade signal sequence, the pre-aiming channel confusion matrix, and the vibration channel confusion matrix within a sliding time window includes: Based on the pre-aiming channel confusion matrix corresponding to the first road level observation signal at each time point within the sliding time window, and the vibration channel confusion matrix corresponding to the second road level observation signal at each time point, the target conditional probability distribution corresponding to each time point is determined; the target conditional probability distribution is used to characterize the probability that, under the condition of determining the true road level and environmental state, the first road level observation signal is observed to be the first road level index and the second road level observation signal is observed to be the second road level index at each time point. The joint likelihood probability is determined based on the target conditional probability distribution corresponding to each time point within the sliding window.

[0078] Specifically, under the window-independent likelihood assumption, in the environment state of And the actual road grade is Under the given conditions, the joint likelihood probability within the sliding window is the product of the target conditional probability distributions corresponding to the observed signals at each time step. This joint likelihood probability can be expressed as: , Let t denote the joint likelihood probability, t-T+1≤τ≤t. Since the environment state is... And the actual road grade is Under the condition that the observation signals of the first road level and the second road level do not interfere with each other, the joint likelihood probability can be expressed as follows: To improve computational stability, the joint likelihood probability can be converted into a log-likelihood form, i.e., .

[0079] In one embodiment, determining the optimal road grade index of the road based on the performance index function of the LQR algorithm and the posterior distribution includes: The Bayesian response is determined based on the performance index function and the posterior distribution. Minimize the Bayesian response to obtain the optimal road grade index for the road.

[0080] Specifically, after determining the performance index function and the posterior distribution, the Bayesian response of the vehicle at time t is constructed. Then, by minimizing this Bayesian response, the optimal road grade index can be obtained, as shown in equation (5). Equation (5) is: .

[0081] in, Indicates the environmental state as And the actual road grade is Under the given conditions, select the action. That is, selecting the performance index function for the m-th set of weight coefficient combinations. Indicates a Bayesian response. This represents the optimal road grade index.

[0082] In one embodiment, determining the optimal weight coefficient combination corresponding to the LQR suspension controller based on the optimal road grade index includes: Based on the weight combination correlation and the optimal road grade index, the optimal weight coefficient combination corresponding to the LQR suspension controller is determined; the weight combination correlation is used to characterize the correlation between road grade and weight coefficient combination; and the weight combination correlation is determined based on the power spectral density corresponding to multiple road grades, the two-degree-of-freedom suspension model and the multi-objective optimization algorithm; the two-degree-of-freedom suspension model is the dynamic model of the vehicle suspension.

[0083] Specifically, after determining the optimal road grade index and before determining the optimal weight coefficient combination, a road surface roughness function corresponding to each road grade is constructed based on the range of commonly used road grades for vehicles. The first derivative of this road surface roughness function is then obtained to acquire the road excitation corresponding to each road grade. The road excitation corresponding to each road grade is shown in equation (6), which is: .

[0084] Where q(l) represents the roughness function, Indicates road incentives, This indicates the vertical height or vertical coordinate of the road, representing the apparent elevation value of the road. This indicates the longitudinal length or longitudinal coordinate of the road. I represents the total number of spatial frequency intervals obtained after dividing a specific frequency interval as specified in ISO 8608. This represents the center frequency of the i-th spatial frequency interval. Let i represent the random variable whose spatial frequency interval is in [0, 2π]. express The power spectral density, and , Furthermore, road surface unevenness is classified using reference power spectral density.

[0085] Subsequently, based on vehicle dynamics theory, the actual quarter-vehicle model was theoretically abstracted and simplified to obtain the corresponding two-degree-of-freedom suspension model. This quarter-vehicle model is the smallest driving unit of the vehicle, including the suspension and tires. Figure 2 This is a schematic diagram of a two-degree-of-freedom suspension model provided in an embodiment of the present invention, as shown below. Figure 2 As shown, this two-degree-of-freedom suspension model includes sprung mass elements, unsprung mass elements, suspension elastic elements, actuator elements, and tire elastic elements. The mass of the sprung mass element is denoted as... The mass of the unsprung mass element is denoted as The stiffness coefficient of the suspension elastic element is denoted as The driving force of the actuator element is denoted as The stiffness coefficient of the tire's elastic element is denoted as... The displacement of the spring-loaded mass is denoted as The unsprung mass displacement is denoted as The road incentive value is denoted as .

[0086] According to the principle of vehicle vertical dynamics, the dynamic model corresponding to this two-degree-of-freedom suspension model is shown in equation (7), which is: .

[0087] in, This represents the second derivative of the displacement of the sprung mass. This indicates the second derivative with respect to the unsprung mass displacement.

[0088] The above dynamic model is generalized to a state-space equation, which is shown in equation (8). Equation (8) is: .

[0089] in, , This means taking the first derivative with respect to x. This indicates taking the first derivative with respect to the displacement of the sprung mass. This indicates taking the first derivative with respect to the unsprung mass displacement. , , , , , , , This indicates road incentives.

[0090] Subsequently, the road excitations corresponding to each road level were used as the road inputs for the two-degree-of-freedom suspension model. The NSGA-II algorithm was then used to combine the weight coefficients of the LQR suspension controller. , , Multi-objective optimization was performed to select the optimized weight coefficient combinations for each road level.

[0091] For example, optimizing the combination of weighting coefficients includes the following steps.

[0092] because , , The value is relative; therefore, in this embodiment of the invention, q3 is set to 1, and based on the NSGA-II algorithm, the optimization variable is set to... and The optimization function is the root mean square value of the vehicle acceleration and tire dynamic deformation response. The optimization objective is to minimize the optimization function. When the road grade is 4, the solution after multi-objective optimization based on the NSGA-II algorithm is a set of solutions in the spatial domain. Each solution in the solution set corresponds to different suspension performance and different combinations of controller weight coefficients. In order to obtain the most suitable controller weight coefficients, it is necessary to select the optimal solution in the multi-objective optimization solution set. In this embodiment of the invention, the multi-objective optimization solution set is normalized to offset the influence of units and values, and the x and y coordinates of all solutions are limited to the interval [0,1]. The normalized x and y are shown in equation (9). Equation (9) is: .

[0093] After normalization, the solution closest to the y=x line in the solution set is taken as the optimal solution, and the optimized weight coefficient combination under road excitation when road grade d=4 is obtained.

[0094] Based on this, the weight coefficient combinations corresponding to all road levels constitute the weight combination association relationship. This weight combination association relationship can be pre-stored in the controller in the form of a lookup table, mapping function, or neural network, achieving the effect of offline optimization to establish the controller and online mapping to obtain control parameters, thus reducing the real-time computing burden.

[0095] After determining the optimal road grade index, the road grade that matches the optimal road grade index is determined from the weight combination correlation, and the weight coefficient combination corresponding to the road grade is determined as the optimal weight coefficient combination corresponding to the LQR suspension controller.

[0096] In one embodiment, determining the optimal driving force based on the optimal combination of weight coefficients includes: The optimal weight coefficient combination is input into the LQR suspension controller to obtain the optimal force output by the LQR suspension controller; the LQR suspension controller is used to determine the LQR system matrix corresponding to the performance index function based on the optimal weight coefficient combination; based on the LQR system matrix and the state space equation corresponding to the two-degree-of-freedom suspension model, the state feedback gain matrix is ​​determined; based on the state feedback gain matrix, the optimal force is determined.

[0097] Specifically, after determining the optimal weight coefficient combination, the optimal weight coefficient combination is input into the LQR suspension controller, and the optimal weight coefficient combination is substituted into the performance index function corresponding to equation (3). After converting the performance index function into the matrix form corresponding to equation (4), the LQR system matrices Q, N, and R are determined. Then, based on the LQR system matrix and the state space equation corresponding to equation (8), the state feedback gain matrix K can be determined. The state feedback gain matrix K is shown in equation (10), which is: .

[0098] P can be solved by the Riccatti equation.

[0099] After determining the state feedback gain matrix K, the optimal action force corresponding to the LQR suspension controller is determined using equation (11) based on the state feedback gain matrix K. This optimal action force is then input into the suspension control system, which changes its vibration response under the influence of the optimal action force, thus completing the suspension control. Equation (11) is as follows: .

[0100] Furthermore, after determining the optimal action force, substituting this optimal action force into the state-space equation corresponding to equation (8) yields the LQR-based suspension continuous state-space equation, which is shown in equation (12). Equation (12) is: .

[0101] in, .

[0102] The continuous state-space equation of the suspension based on LQR is discretized to obtain the discrete state-space equation of the suspension based on LQR, which is shown in equation (13). Equation (13) is: .

[0103] Where k represents the sampling time, Representing discrete states , Representing discrete states , Discrete state .

[0104] Furthermore, the control model (i.e., vehicle suspension control model) in this embodiment of the invention is compared with existing control models, and the superiority of the active suspension preview control method based on signal game theory provided in this embodiment of the invention is verified through the following data. The road model includes four road conditions: Road 1, Road 2, Road 3, and Road 4, ensuring a difference between the apparent elevation obtained from visual detection and the actual excitation elevation. The simulated road conditions are constructed based on Class A roads (d=4) and Class B roads (d=6) as specified in ISO 8608 standard, and are divided into four road conditions. Specifically, Road 1: visually detected as a Class A road, but the actual excitation is a Class B road; Road 2: visually detected as a Class B road, but the actual excitation is a Class A road; Road 3: a discrete impact surface with an amplitude of 0.1 meters is superimposed on Road 1; Road 4: a discrete impact surface with an amplitude of 0.1 meters is superimposed on Road 2. Under road conditions 1 and 2, the root mean square (RMS) value of vehicle acceleration or tire dynamic deformation is used as the evaluation index; under road conditions 3 and 4, the peak value of vehicle acceleration or tire dynamic deformation is used as the evaluation index. In addition to the control model corresponding to the embodiments of this invention, the control group models selected include passive suspension, first LQR suspension (controller fixed at road level A, denoted as LQR-A), second LQR suspension (controller fixed at road level B, denoted as LQR-B), suspension controller based solely on pre-aiming information (using the LQR algorithm, calculating the optimal action force solely based on visual detection results, denoted as LQR-V), and suspension controller based solely on state information perceiving the road (using the LQR algorithm, calculating the optimal action force solely based on state observation results, denoted as LQR-E). Specific comparison results are shown in Table 1.

[0105] Table 1

[0106] In Table 1, the unit of vehicle acceleration is meters per second squared, and the unit of tire dynamic deformation is millimeters. As shown in the simulation data in Table 1, compared to the control group that relies solely on preview sampling information or only uses state observation results, the active suspension preview control method based on signal game theory provided in this embodiment of the invention can significantly reduce the root mean square value of vehicle acceleration. This method constructs a dual-source signal game model of preview and vibration channels, performs fusion decision-making on the first road grade index and the second road grade index based on Bayesian posterior probability, selects the optimal road grade index that best approximates the real road excitation, and adaptively adjusts the weight coefficient combination of the LQR suspension controller accordingly, thereby effectively improving the vehicle's ride comfort under different road surface excitation conditions.

[0107] Optionally, in the prior art, the control model may include passive suspension or ordinary anti-sight control suspension, etc., and the embodiments of the present invention do not limit this.

[0108] The active suspension anti-aiming control device based on signal game theory provided by the present invention will be described below. The active suspension anti-aiming control device based on signal game theory described below can be referred to in correspondence with the active suspension anti-aiming control method based on signal game theory described above.

[0109] This invention also provides an active suspension anti-aiming control device based on signal game theory. Figure 3 This is a schematic diagram of the active suspension anti-aiming control device based on signal game theory provided in an embodiment of the present invention, as shown below. Figure 3 As shown, the active suspension anti-aiming control device 300 based on signal game theory includes: a first determination module 310, a second determination module 320, a game decision module 330, a suspension controller weight coefficient switching module 340, and a suspension control algorithm module 350.

[0110] The first determining module 310 is used to determine the first road grade index of the road based on the pre-sampling information.

[0111] The second determining module 320 is used to determine a second road grade index of the road based on the road amplitude of the road; the road amplitude is used to characterize the vibration response between the vehicle and the road.

[0112] The game decision module 330 is used to make game decisions on the first road grade index and the second road grade index with the goal of optimizing the smoothness index of the vehicle, and using signal game theory as the game theory guidance method, to obtain the optimal road grade index of the road.

[0113] The suspension controller weight coefficient switching module 340 is used to determine the optimal weight coefficient combination corresponding to the LQR suspension controller based on the optimal road grade index.

[0114] The suspension control algorithm module 350 is used to determine the optimal action force based on the optimal weight coefficient combination, and the optimal action force is used to perform suspension control on the vehicle.

[0115] The active suspension preview control device based on signal game theory provided in this invention determines a first road grade index by preview sampling information, and a second road grade index by road amplitude, which characterizes the vibration response between the vehicle and the road. With the optimal vehicle ride comfort index as the objective, and using signal game theory as the guiding method, a game decision is made between the first and second road grade indices to obtain the optimal road grade index. Based on the optimal road grade index, the optimal weight coefficient combination corresponding to the LQR suspension controller is determined, and the optimal action force is determined based on the optimal weight coefficient combination to control the vehicle's suspension. In this invention, a first road grade index with forward-looking but low accuracy from a pre-aiming sensor and a second road grade index with accurate but lagging accuracy from a state observer are used as two independent observation signals. Using signal game theory as a guiding method, a game-theoretic decision is made between the two observation signals to obtain the optimal road grade index that optimizes vehicle ride comfort. This optimal road grade index serves as a reference for the vehicle suspension switching weight system. This solves the problem in the prior art where the apparent elevation value and the actual road excitation are not simultaneously affecting the suspension control effect, enabling the prediction of the road elevation value in front of the vehicle and effectively improving vehicle ride comfort.

[0116] Optionally, the first determining module 310 is specifically used for: Based on the vehicle attitude information and the pre-aiming sampling information, the vertical and longitudinal coordinates of the road are determined; Frequency domain analysis is performed on the vertical and longitudinal coordinates to obtain the first reference power spectral density corresponding to the road. Based on the first reference power spectral density, the first road grade index of the road is determined.

[0117] Optionally, the second determining module 320 is specifically used for: The road surface roughness is determined based on the road amplitude and the longitudinal coordinates of the road. Based on the road surface roughness, the second reference power spectral density corresponding to the road is determined; The second road grade index of the road is determined based on the second reference power spectral density.

[0118] Optionally, the game decision module 330 is specifically used for: The aiming channel confusion matrix is ​​determined based on the first road level observation signal. The aiming channel confusion matrix is ​​obtained by statistical calibration of historical measurement errors of lidar under different ambient light and road surface reflectivity conditions. It is used to characterize the conditional probability distribution of the first road level observation signal under the condition of determining the true road level. The vibration channel confusion matrix is ​​determined based on the second road grade observation signal; the vibration channel confusion matrix is ​​used to characterize the conditional probability distribution of the observed second road grade observation signal being the second road grade index under the condition of determining the true road grade and environmental state; The joint likelihood probability is determined based on the first road level signal sequence, the second road level signal sequence, the pre-aiming channel confusion matrix, and the vibration channel confusion matrix within the sliding time window. Based on Bayesian consistency, prior distribution, and the joint likelihood probability, the posterior distribution corresponding to the true road level is determined; the prior distribution is the prior distribution of the true road level and the environmental state. Based on the performance index function of the LQR algorithm and the posterior distribution, the optimal road grade index of the road is determined; the performance index function is used to characterize the expected cost of taking an action under the actual road grade and environmental conditions.

[0119] Optionally, the game decision module 330 is specifically used for: Based on the pre-aiming channel confusion matrix corresponding to the first road level observation signal at each time point within the sliding time window, and the vibration channel confusion matrix corresponding to the second road level observation signal at each time point, the target conditional probability distribution corresponding to each time point is determined; the target conditional probability distribution is used to characterize the probability that, under the condition of determining the true road level and environmental state, the first road level observation signal is observed to be the first road level index and the second road level observation signal is observed to be the second road level index at each time point. The joint likelihood probability is determined based on the target conditional probability distribution corresponding to each moment within the sliding time window.

[0120] Optionally, the game decision module 330 is specifically used for: The Bayesian response is determined based on the performance index function and the posterior distribution. Minimize the Bayesian response to obtain the optimal road grade index for the road.

[0121] Optionally, given that the actual road grade and environmental conditions are determined, the first road grade observation signal and the second road grade observation signal do not interfere with each other.

[0122] Optionally, the suspension controller weight coefficient switching module 340 is specifically used for: Based on the weight combination correlation and the optimal road grade index, the optimal weight coefficient combination corresponding to the LQR suspension controller is determined; the weight combination correlation is used to characterize the correlation between road grade and weight coefficient combination.

[0123] Optionally, the weight combination association is determined based on the following steps: Based on the power spectral density corresponding to multiple road levels, the road excitation corresponding to each road level is determined. A two-degree-of-freedom suspension model is constructed; the two-degree-of-freedom suspension model is a dynamic model of the vehicle suspension. The road excitation corresponding to each road grade is input into the two-degree-of-freedom suspension model to obtain the weight coefficient combination corresponding to each road grade.

[0124] Optionally, the suspension control algorithm module 350 is specifically used for: The optimal weight coefficient combination is input into the LQR suspension controller to obtain the optimal force output by the LQR suspension controller; the LQR suspension controller is used to determine the LQR system matrix corresponding to the performance index function based on the optimal weight coefficient combination; based on the LQR system matrix and the state space equation corresponding to the two-degree-of-freedom suspension model, the state feedback gain matrix is ​​determined; based on the state feedback gain matrix, the optimal force is determined.

[0125] This invention also provides an active suspension anti-aiming control system based on signal game theory, the system comprising: Preview sensor, used to collect preview sampling information corresponding to the road; A state observer is used to collect the road vibration amplitude of the road. The processor is used to execute the active suspension anti-aiming control method based on signal game theory as described in any of the above embodiments.

[0126] Figure 4 This is a schematic diagram of the structure of the electronic device provided in the embodiment of the present invention, such as... Figure 4As shown, the electronic device may include: a processor 410, a communication interface 420, a memory 430, and a communication bus 440, wherein the processor 410, the communication interface 420, and the memory 430 communicate with each other through the communication bus 440. The processor 410 can call logical instructions in the memory 430 to execute an active suspension anti-aiming control method based on signal game theory. This method includes: determining a first road grade index based on anti-aiming sampling information; determining a second road grade index based on the road amplitude; the road amplitude is used to characterize the vibration response between the vehicle and the road; with the optimal ride comfort index of the vehicle as the objective, and using signal game theory as the game theory guidance method, performing game decisions on the first road grade index and the second road grade index to obtain the optimal road grade index; determining the optimal weight coefficient combination corresponding to the LQR suspension controller based on the optimal road grade index; and determining the optimal actuation force based on the optimal weight coefficient combination, the optimal actuation force being used to determine the state control parameters of the vehicle's suspension anti-aiming control system.

[0127] Furthermore, the logical instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0128] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the active suspension preview control method based on signal game theory provided by the above methods. The method includes: determining a first road grade index based on preview sampling information; determining a second road grade index based on the road amplitude; the road amplitude being used to characterize the vibration response between the vehicle and the road; taking the optimal ride comfort index of the vehicle as the objective and using signal game theory as the game theory guidance method, performing game decision-making on the first road grade index and the second road grade index to obtain the optimal road grade index of the road; determining the optimal weight coefficient combination corresponding to the LQR suspension controller based on the optimal road grade index; and determining the optimal actuation force based on the optimal weight coefficient combination, the optimal actuation force being used to determine the state control parameters of the vehicle's suspension preview control system.

[0129] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the active suspension anti-aiming control method based on signal game theory provided by the above methods. This method includes: determining a first road grade index based on anti-aiming sampling information; determining a second road grade index based on the road amplitude; the road amplitude being used to characterize the vibration response between the vehicle and the road; aiming at optimizing the ride comfort index of the vehicle, and using signal game theory as a game theory guidance method, performing game decisions on the first road grade index and the second road grade index to obtain an optimal road grade index; determining an optimal weight coefficient combination corresponding to the LQR suspension controller based on the optimal road grade index; and determining an optimal actuation force based on the optimal weight coefficient combination, the optimal actuation force being used to determine the state control parameters of the vehicle's suspension anti-aiming control system.

[0130] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0131] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0132] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An active suspension anti-aiming control method based on signal game theory, characterized in that, include: Based on the pre-sampling information, the first road grade index of the road is determined; Based on the road amplitude, a second road grade index is determined for the road; The road amplitude is used to characterize the vibration response between the vehicle and the road; With the goal of optimizing the ride comfort index of the vehicle, and using signal game theory as the game theory guidance method, a game decision is made on the first road grade index and the second road grade index to obtain the optimal road grade index of the road. Based on the optimal road grade index, determine the optimal weight coefficient combination corresponding to the LQR suspension controller; Based on the optimal combination of weighting coefficients, the optimal action force is determined, and the optimal action force is used to determine the state control parameters of the vehicle's suspension anti-aiming control system.

2. The active suspension anti-aiming control method based on signal game theory according to claim 1, characterized in that, The step of performing a game-theoretic decision on the first road grade index and the second road grade index to obtain the optimal road grade index for the road includes: The aiming channel confusion matrix is ​​determined based on the first road level observation signal. The aiming channel confusion matrix is ​​obtained by statistical calibration of historical measurement errors of lidar under different ambient light and road surface reflectivity conditions. It is used to characterize the conditional probability distribution of the first road level observation signal under the condition of determining the true road level. The vibration channel confusion matrix is ​​determined based on the second road grade observation signal; the vibration channel confusion matrix is ​​used to characterize the conditional probability distribution of the observed second road grade observation signal being the second road grade index under the condition of determining the true road grade and environmental state; The joint likelihood probability is determined based on the first road level signal sequence, the second road level signal sequence, the pre-aiming channel confusion matrix, and the vibration channel confusion matrix within the sliding time window. Based on Bayesian consistency, prior distribution, and the joint likelihood probability, the posterior distribution corresponding to the true road level is determined; the prior distribution is the prior distribution of the true road level and the environmental state. Based on the performance index function of the LQR algorithm and the posterior distribution, the optimal road grade index of the road is determined; the performance index function is used to characterize the expected cost of taking an action under the actual road grade and environmental conditions.

3. The active suspension anti-aiming control method based on signal game theory according to claim 2, characterized in that, The determination of the joint likelihood probability based on the first road level signal sequence, the second road level signal sequence, the pre-aiming channel confusion matrix, and the vibration channel confusion matrix within the sliding time window includes: Based on the pre-aiming channel confusion matrix corresponding to the first road level observation signal at each time point within the sliding time window, and the vibration channel confusion matrix corresponding to the second road level observation signal at each time point, the target conditional probability distribution corresponding to each time point is determined; the target conditional probability distribution is used to characterize the probability that, under the condition of determining the true road level and environmental state, the first road level observation signal is observed to be the first road level index and the second road level observation signal is observed to be the second road level index at each time point. The joint likelihood probability is determined based on the target conditional probability distribution corresponding to each moment within the sliding time window.

4. The active suspension anti-aiming control method based on signal game theory according to claim 2, characterized in that, The optimal road grade index for the road is determined by the performance index function based on the LQR algorithm and the posterior distribution, including: The Bayesian response is determined based on the performance index function and the posterior distribution. Minimize the Bayesian response to obtain the optimal road grade index for the road.

5. The active suspension anti-aiming control method based on signal game theory according to any one of claims 2-4, characterized in that, Under the condition that the actual road grade and environmental conditions are determined, the first road grade observation signal and the second road grade observation signal do not interfere with each other.

6. The active suspension anti-aiming control method based on signal game theory according to claim 1, characterized in that, The determination of the first road grade index based on pre-sampling information includes: Based on the vehicle attitude information and the pre-aiming sampling information, the vertical and longitudinal coordinates of the road are determined; Frequency domain analysis is performed on the vertical and longitudinal coordinates to obtain the first reference power spectral density corresponding to the road. Based on the first reference power spectral density, a first road grade index is determined for the road. The determination of the second road grade index based on the road amplitude includes: The road surface roughness is determined based on the road amplitude and the longitudinal coordinates of the road. Based on the road surface roughness, the second reference power spectral density corresponding to the road is determined; The second road grade index of the road is determined based on the second reference power spectral density.

7. The active suspension anti-aiming control method based on signal game theory according to claim 1, characterized in that, The process of determining the optimal weight coefficient combination corresponding to the LQR suspension controller based on the optimal road grade index includes: Based on the weight combination correlation and the optimal road grade index, the optimal weight coefficient combination corresponding to the LQR suspension controller is determined; the weight combination correlation is used to characterize the correlation between road grade and weight coefficient combination; and the weight combination correlation is determined based on the power spectral density corresponding to multiple road grades, the two-degree-of-freedom suspension model and the multi-objective optimization algorithm; the two-degree-of-freedom suspension model is the dynamic model of the vehicle suspension.

8. The active suspension anti-aiming control method based on signal game theory according to claim 7, characterized in that, The determination of the optimal action force based on the optimal weight coefficient combination includes: The optimal weight coefficient combination is input into the LQR suspension controller to obtain the optimal force output by the LQR suspension controller; the LQR suspension controller is used to determine the LQR system matrix corresponding to the performance index function based on the optimal weight coefficient combination; based on the LQR system matrix and the state space equation corresponding to the two-degree-of-freedom suspension model, the state feedback gain matrix is ​​determined; based on the state feedback gain matrix, the optimal force is determined.

9. An active suspension anti-aiming control device based on signal game theory, characterized in that, include: The first determining module is used to determine the first road grade index of the road based on the pre-sampling information; The second determining module is used to determine a second road grade index of the road based on the road amplitude; the road amplitude is used to characterize the vibration response between the vehicle and the road; The game decision module is used to make game decisions on the first road grade index and the second road grade index with the goal of optimizing the smoothness index of the vehicle, and using signal game theory as the game theory guidance method, to obtain the optimal road grade index of the road. The suspension controller weight coefficient switching module is used to determine the optimal weight coefficient combination corresponding to the LQR suspension controller based on the optimal road grade index. The suspension control algorithm module is used to determine the optimal action force based on the optimal weight coefficient combination. The optimal action force is used to determine the state control parameters of the vehicle's suspension anti-aiming control system.

10. An active suspension anti-aiming control system based on signal game theory, characterized in that, include: Preview sensor, used to collect preview sampling information corresponding to the road; A state observer is used to collect the road vibration amplitude of the road. A processor for executing the active suspension anti-aiming control method based on signal game theory as described in any one of claims 1 to 8.

Images