Real-time vehicle formation automatic control method based on zero neural network

Abstract

The invention discloses a real-time vehicle formation automatic control method based on a zero neural network, and belongs to the field of automatic driving. According to the method, a real-time vehicle formation control problem is converted into a nonlinear equation set solving problem. By using the model, the challenge of real-time vehicle formation control can be effectively solved. In addition, the research also develops a distributed vehicle driving parameter monitoring device and a controller, and constructs a novel automatic control system. Particularly, the invention further provides a zero neural network convergent in preset time, the network not only can realize rapid convergence of the system, but also can optimize the use of computing resources by adjusting time-varying parameters, and the computing efficiency of the system is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent driving, and particularly to a real-time vehicle platoon automatic control method based on a nullifying neural network. Background Art

[0002] With the progress of technology, the number of vehicles in many countries and regions has increased significantly, which has undoubtedly brought a series of complex traffic challenges. Vehicle platoon control technology, as a key part of intelligent transportation systems, has attracted extensive research attention. This technology has many advantages, including enhancing driving safety, optimizing traffic flow, improving fuel efficiency, and reducing emissions. The core goal of vehicle platoon control is to ensure that each vehicle in the platoon maintains a set speed and keeps a safe distance between vehicles. Through vehicle-to-vehicle communication technology, the perception range of vehicles has been significantly extended, enabling them to receive and exchange key information such as position, speed, and acceleration, thus achieving more efficient traffic management.

[0003] To achieve vehicle platoon control, researchers have proposed various control strategies. However, the effectiveness of these methods is usually based on the assumption that the system state gradually approaches infinity over time. To address this challenge, developing strategies that can achieve control objectives within a finite time has become a hot research area in academia. In summary, it is crucial to propose an efficient and reliable control strategy for real-time vehicle platoons. Summary of the Invention

[0004] In view of the above deficiencies in the prior art, the present invention provides a real-time vehicle platoon automatic control method based on a nullifying neural network.

[0005] To achieve the above invention objective, the technical solution adopted by the present invention is as follows:

[0006] A real-time vehicle platoon automatic control method based on a nullifying neural network, comprising the following steps:

[0007] S1. Input the real-time vehicle platoon control problem to be solved and design a nullifying neural network model;

[0008] S2. Establish a mathematical model based on the input real-time vehicle platoon control problem;

[0009] S3. Establish an automatic control system based on distributed vehicle driving monitoring data and control data, and use the mathematical model established in S2 to iteratively solve the real-time vehicle platoon control problem under the interference of environmental noise;

[0010] S4. Control the vehicle distance, vehicle speed, and acceleration of the tracking vehicle based on the result of the iterative solution in S3.

[0011] Furthermore, the real-time vehicle formation control problem to be solved in S1 is expressed as:

[0012] ;

[0013] In the formula, represents the position where vehicle i travels to, represents the speed of vehicle i, represents the mass of vehicle i, the mechanical efficiency of the transmission system of vehicle i, represents the actual driving or braking torque of vehicle i, is the wheel radius of vehicle i; is the comprehensive aerodynamic drag coefficient of vehicle i; is the acceleration due to gravity, is the rolling resistance coefficient, is the longitudinal dynamic inertia delay of vehicle i, is the driving or braking torque required by vehicle i.

[0014] Furthermore, the nullifying neural network model in S1 is expressed as:

[0015] ;

[0016] In the formula, is the nullifying neural network model; is the i-th error variable and ; are all model coefficients and among them , .

[0017] Furthermore, establishing an automatic control system based on distributed vehicle driving monitoring data and control data in S3 specifically includes the following steps:

[0018] S31. Monitor the state of the i-th vehicle;

[0019] S32. Use the error variable to perform quantization performance calculation on the monitored state;

[0020] S33. Design a new controller for the monitored vehicle.

[0021] Furthermore, the specific method of quantization performance calculation in S32 is:

[0022] ;

[0023] In the formula, is the position error variable of the i-th vehicle, is the speed error variable of the i-th vehicle, is the acceleration error variable of the i-th vehicle, is the position of the i-th vehicle, , and represent the position, speed and acceleration of the leading vehicle monitored by vehicle i, is the speed of vehicle i, is the acceleration of vehicle i, is the expected distance between vehicle i and the leading vehicle.

[0024] Furthermore, the new controller in S33 is expressed as:

[0025] ;

[0026] In the formula, , and are fixed parameters, is a variable regarding the residual, and and represent the designed non-linear variables regarding vehicle i and vehicle j, is the Laplacian matrix element related to vehicle i and vehicle j, is the weight matrix element between vehicle i and the virtual leading vehicle.

[0027] The present invention has the following beneficial effects:

[0028] The present invention utilizes its predetermined time convergence and anti-interference ability to design an automatic control system in combination with the dynamic system of vehicle formation. The advantages of the present invention lie in its real-time performance, anti-interference ability, high precision, high efficiency and saving of computing resources. The present invention solves the real-time vehicle formation control problem with a neural network and has quite high novelty and practicability. Description of the Drawings

[0029] Figure 1 is the flowchart of the real-time vehicle formation automatic control method based on the annihilation neural network of the present invention.

[0030] Figure 2 is the information diagram of the longitudinal dynamics of the vehicle in the embodiment of the present invention.

[0031] Figure 3 is the information topology diagram of the undirected communication of the vehicle formation in the embodiment of the present invention.

[0032] Figure 4 is the acceleration information diagram of vehicle i in the formation in the embodiment of the present invention.

[0033] Figure 5 is the speed information diagram of vehicle i in the formation in the embodiment of the present invention.

[0034] Figure 6 This is the position information diagram of vehicle i in the formation of the embodiment of the present invention.

[0035] Figure 7 This is the spacing information diagram of the front and rear vehicles in the formation of the embodiment of the present invention.

[0036] Figure 8 This is the spacing information diagram between vehicle i and the virtual leading vehicle in the formation of the embodiment of the present invention.

[0037] Figure 9 This is the speed difference information diagram between vehicle i and the virtual leading vehicle in the formation of the embodiment of the present invention.

[0038] Figure 10 This is the acceleration difference information diagram between vehicle i and the virtual leading vehicle in the formation of the embodiment of the present invention. Detailed implementation manners

[0039] The following describes the detailed implementation manners of the present invention to facilitate those skilled in the art of the present technology to understand the present invention. However, it should be clear that the present invention is not limited to the scope of the detailed implementation manners. For those of ordinary skill in the art of the present technology, as long as various changes are within the spirit and scope of the present invention defined and determined by the appended claims, these changes are obvious, and all inventions created using the concept of the present invention are within the scope of protection.

[0040] A real-time vehicle formation automatic control method based on a nullifying neural network, as Figure 1 shown, includes the following steps:

[0041] S1. Input the real-time vehicle formation control problem to be solved and design a nullifying neural network model;

[0042] To simplify the analysis process of the control system, the longitudinal slip of the tires can be ignored, the vehicle body is regarded as a rigid body symmetric structure, the flexible deformation and asymmetric effects of the vehicle body are ignored, and only the longitudinal dynamics of the vehicle are concerned to reduce the complexity of the control system, as Figure 2 shown. The longitudinal dynamics of the vehicle can be expressed as:

[0043] ;

[0044] where represents the position where vehicle i travels to, represents the speed of vehicle i, represents the mass of vehicle i, the mechanical efficiency of the transmission system of vehicle i, represents the actual driving or braking torque of vehicle i, is the wheel radius of vehicle i; is the comprehensive aerodynamic drag coefficient of vehicle i; is the gravitational acceleration, is the rolling resistance coefficient, is the longitudinal dynamic inertia delay of vehicle i, is the driving or braking torque required by vehicle i. Then the non - linear model is converted into a linear representation for controller design:

[0045]

[0046] where is the input after linearization, and then a new linear model of vehicle longitudinal dynamics can be obtained:

[0047] ;

[0048] where is the acceleration of vehicle i, is the input of vehicle i, is the interference of environmental factors on vehicle i. However, for the virtual lead vehicle, it is not affected by external factors, so its dynamic model can be expressed as:

[0049] ;

[0050] i.e., the case of i = 0. 、 、 and represent the position, speed, acceleration and input of the virtual lead vehicle respectively

[0051] S2. Establish a mathematical model for the real - time vehicle formation control problem based on the input;

[0052] Establish a basic non - linear equation set, expressed as;

[0053]

[0054] where represents time, and represent the start time and end time respectively. is a non - linear mapping. By constructing the annihilating neural network, can be designed first, with the error being , and then the model of the annihilating neural network can be obtained:

[0055] ;

[0056] where is an activation function. The activation function designed in the present invention can make the error converge to zero rapidly within a predetermined time. The specific model is as follows: ;

[0057] Among them , , and .

[0058] Then design a time-varying parameter that is efficient and can save computing resources , specifically expressed as:

[0059] .

[0060] S3. Establish an automatic control system based on distributed vehicle driving monitoring data and control data, and use the mathematical model established in S2 to iteratively solve the real-time vehicle formation control problem under the interference of environmental noise;

[0061] For monitoring the state of vehicle i, it can be expressed as:

[0062]

[0063] Among them , and represent the position, speed, and acceleration of vehicle i monitoring the leading vehicle. Among them , and are defined as follows:

[0064]

[0065] Among them represents the sign function. Our expected effect of vehicle formation control is as follows:

[0066]

[0067] Among them represents the expected inter-vehicle distance, represents the predetermined time. The effect of the predetermined time convergence can be achieved through the designed activation function, and the effect is expressed as follows:

[0068] ;

[0069] To analyze vehicle performance, a series of error variables are introduced to quantify the performance:

[0070] ;

[0071] Among them , represents the expected distance between vehicle i and the leading vehicle.

[0072] For vehicle i, design a new controller, expressed as follows:

[0073] ;

[0074] Among them 、 and are all fixed parameters, is a positive variable parameter regarding the residual, where is expressed as .

[0075] Among them ;

[0076] S4. Control the distance, speed, and acceleration of the following vehicle based on the result obtained by iterative solution in S3.

[0077] Using the designed model, the specific steps for iterative solution of the real-time vehicle formation control problem under the interference of environmental noise are as follows:

[0078] Give the communication topology diagram of the vehicle formation, as Figure 3 shown.

[0079] Give the relevant data of the virtual leading vehicle's travel:

[0080] Give the relevant environmental noise , give the expected adjacent vehicle distance . Then bring the parameters into the system for solution.

[0081] The final result Figure 4 It can be seen that the vehicles in the formation can well track the acceleration information of the virtual leading vehicle, Figure 5 It can be seen that the vehicles in the formation can well track the speed information of the virtual leading vehicle, Figure 6 It can be seen that the vehicles in the formation can well track the position information of the virtual leading vehicle and there will be no overtaking situation. Figure 7 and Figure 8 It can be seen that the vehicles in the formation can well maintain the distance between the front and rear vehicles, and the vehicle distance has always been at the ideal distance of 8m, and there is no overtaking situation during the tracking process. Figure 9 It can be seen that the vehicles in the formation can well track the speed of the virtual leading vehicle and can maintain the same speed as the leading vehicle. Figure 10 It can be seen that the vehicles in the formation can well track the acceleration of the virtual leading vehicle and can quickly reach the state where the acceleration is the same as that of the leading vehicle. Through this example, it can be concluded that an automatic control system based on neural network proposed by the present invention for realizing real-time vehicle formation has high accuracy and robustness, and its efficiency and excellent performance are further verified by its practical application under environmental interference.

[0082] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable data processing devices to generate a machine, such that the instructions executed by the processors of the computer or other programmable data processing devices produce means for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0083] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including instruction means that implement the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0084] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operating steps are executed on the computer or other programmable device to generate a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0085] Specific embodiments are applied in the present invention to elaborate on the principles and implementation manners of the present invention. The descriptions of the above embodiments are only used to help understand the method of the present invention and its core idea; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present invention.

[0086] Those of ordinary skill in the art will realize that the embodiments described herein are for helping readers understand the principles of the present invention, and it should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those of ordinary skill in the art can make various other specific deformations and combinations that do not depart from the essence of the present invention based on the technical revelations disclosed in the present invention, and these deformations and combinations are still within the protection scope of the present invention.

Claims

1. A real-time vehicle formation automatic control method based on zeroing neural network, characterized in that: The steps include: S1. Input the real-time vehicle platoon control problem to be solved and design a zeroing neural network model; S2. Establish a mathematical model for the real-time vehicle platoon control problem based on the input; S3, establishing an automatic control system based on distributed vehicle driving monitoring data and control data, and using the mathematical model established in S2 to iteratively solve the real-time vehicle formation control problem under the interference of environmental noise; S4. Based on the result of the iterative solution of S3, the distance, speed and acceleration of the tracked vehicle are controlled.

2. The real-time vehicle formation automatic control method based on annihilation neural network according to claim 1 is characterized in that: The real-time vehicle platoon control problem to be solved in S1 is expressed as: ； In the formula, represents the location to which vehicle i travels, represents the speed of vehicle i, represents the mass of vehicle i, The mechanical efficiency of the transmission system of vehicle i, represents the actual driving or braking torque of vehicle i, is the wheel radius of vehicle i; is the comprehensive aerodynamic drag coefficient of vehicle i; is the acceleration due to gravity, is the rolling resistance coefficient, is the longitudinal dynamic inertia delay of vehicle i, is the driving or braking torque required by vehicle i.

3. The real-time vehicle formation automatic control method based on annihilation neural network according to claim 1 is characterized in that: The zeroing neural network model in S1 is expressed as: ； In the formula, It is a zeroing neural network model; is the i-th error variable and ; are model coefficients and , .

4. The real-time vehicle formation automatic control method based on annihilation neural network according to claim 1 is characterized in that: The establishment of an automatic control system based on distributed vehicle driving monitoring data and control data in S3 specifically includes the following steps: S31, monitoring the status of the i-th vehicle; S32, using the error variable to perform quantitative performance calculation on the monitored state; S33. Design a new controller for the monitored vehicle.

5. The real-time vehicle formation automatic control method based on annihilation neural network according to claim 4 is characterized in that: The specific method of calculating the quantitative performance in S32 is: ； In the formula, is the position error variable of the i-th vehicle, is the speed error variable of the i-th vehicle, is the acceleration error variable of the i-th vehicle, is the position of the i-th vehicle, , and Indicates that vehicle i monitors the position, speed and acceleration of the leading vehicle, is the speed of vehicle i, is the acceleration of vehicle i, is the expected distance between vehicle i and the leading vehicle.

6. The real-time vehicle formation automatic control method based on annihilation neural network according to claim 4 is characterized in that: The new controller in S33 is represented as: ； In the formula, , and is a fixed parameter, is the variable about the residual, and represents the designed nonlinear variables about vehicle i and vehicle j, is the Laplace matrix element related to vehicle i and vehicle j, is the weight matrix element between vehicle i and the virtual lead vehicle.

Images