An adaptive switching control method for following and overtaking in autonomous vehicles

Abstract

This invention discloses an adaptive switching control method based on PD control. It constructs a basic car-following dynamics model of the vehicle based on a classical optimal speed model, introduces a switchable overtaking behavior switch, and establishes an integrated unified dynamic equation adapting to both car-following and overtaking modes. Addressing the differentiated control requirements of the two driving behaviors under emergency braking, this invention designs a hierarchical adaptive control strategy. P control is used to achieve steady-state regulation in car-following conditions, while PD adaptive control is used to achieve precise dynamic regulation in overtaking conditions. This invention combines linear stability analysis to derive the optimal feedback gain of the control system and the stability constraints of the traffic flow system, effectively suppressing traffic chaos caused by overtaking behavior. Simultaneously, Lyapunov analysis is introduced in the numerical simulation to assist in verifying system stability. This invention significantly improves the driving safety, ride smoothness, and road traffic efficiency of autonomous vehicles.

Description

Technical Field

[0001] This invention belongs to the field of traffic control technology, specifically relating to an adaptive switching control method for autonomous vehicles based on PD control and adapted to both following and overtaking driving behaviors. Background Technology

[0002] Following and overtaking are core driving behaviors for autonomous vehicles. Overtaking disturbances and sudden braking effects can easily cause traffic flow oscillations and driving fluctuations, seriously affecting traffic safety and stability. Existing autonomous driving following models are mostly based on a single steady-state assumption (such as the classic optimal speed model), failing to effectively distinguish and adapt to the essential dynamic differences between "following" and "overtaking." This leads to distorted model control of vehicle speed and spacing under dynamic conditions such as overtaking, easily causing traffic flow oscillations.

[0003] Meanwhile, existing control strategies generally adopt a uniform feedback framework, lacking differentiated control designs for different operating conditions. In particular, when dealing with the "sudden braking effect," existing models fail to consider the weakening effect of historical acceleration during overtaking, resulting in harsh control commands and exacerbating the jerking sensation of vehicle movement and traffic disturbance.

[0004] Furthermore, the tuning of control parameters in existing technologies often relies on experience or simple constraints, lacking theoretical guidance based on system stability. This makes it difficult to ensure the global stability of the control system under all operating conditions, thus limiting the safety and smoothness of autonomous driving in complex traffic environments. Summary of the Invention

[0005] This invention aims to address the problems of traditional car-following models being unable to adapt to both car-following and overtaking behaviors, and the poor stability of traffic flow caused by sudden braking and overtaking disturbances. It provides a PD-based adaptive switching control method for autonomous vehicles that enables car-following and overtaking. This method achieves precise and stable longitudinal vehicle control through a combination of a three-tiered technical architecture: "dual-behavior differentiated modeling - hierarchical adaptive control - linear stability constraints." Furthermore, the numerical simulation stage incorporates Lyapunov analysis to assist in verifying the system's operational characteristics, comprehensively improving the stability and traffic efficiency of autonomous vehicle traffic flow.

[0006] To achieve the above objectives, this invention adopts the following technical solution: First, by constructing a unified dual-behavior dynamic model based on the overtaking switching function, a precise mathematical description and smooth transition between following and overtaking modes are achieved at the underlying level. Second, based on this model, a condition-specific adaptive P / PD switching controller is designed, and the control law is simplified by introducing the speed proportional coefficient of adjacent lanes. Finally, the system stability constraints are derived through linear stability theory, achieving adaptive matching of control parameters and shock-free switching of operating conditions. These three levels are interconnected and progressively enhance each other: the dynamic model provides the controller with a precise controlled object, the controller provides the mathematical basis for stability analysis, and the stability constraints, in turn, guide the tuning of control parameters, together forming a complete technical closed loop.

[0007] The technical solution of this invention to solve the above-mentioned technical problems is as follows: A method for adaptive switching control of following and overtaking in an autonomous vehicle, comprising the following steps:

[0008] Step 1: Using onboard sensors, the position, speed, and acceleration of the vehicle and the vehicle in front are collected in real time, as well as the speed and acceleration of vehicles in adjacent lanes, to build a complete data base for condition identification and dynamic calculation.

[0009] Step 2: To address the issue that a single model cannot characterize the dynamic differences between the two behaviors, this step creatively introduces an "overtaking switching function" and an "emergency braking attenuation coefficient" based on the classic optimal speed model, and constructs an integrated model through the following sub-steps:

[0010] Step 2.1, Construct a basic car-following model: Based on the optimal speed function, describe the speed adjustment law of the vehicle under normal car-following conditions.

[0011] Step 2.2 defines the basic emergency braking term as the difference between the current acceleration and the historical acceleration. An overtaking switch is introduced.

[0012] and attenuation coefficient Build and improve emergency braking items Its core working principle is: when a vehicle follows another vehicle ( The emergency braking effect is fully preserved; when the vehicle is overtaking ( The influence of historical acceleration is attenuated. This aligns with the physical fact that drivers tend to focus more on the clearance ahead rather than their own past deceleration during overtaking.

[0013] Step 2.3, fuse the dual-mode dynamic equations:

[0014] ① Follow-the-car mode ( ): Based on the basic model, a dual-vehicle speed coordination correction term is introduced to simulate the perception of information about multiple vehicles ahead and the smooth following characteristics when following another vehicle.

[0015] ② Overtaking mode ( ): Introducing the overtaking expectation term This driving term combines the difference between the vehicle's speed and the optimal speed of the vehicle in front, representing the motivation for the vehicle to actively accelerate and seek lane change gaps during overtaking.

[0016] Finally, by using the overtaking switch The two model equations mentioned above are merged into a unified dynamic equation. The core of this technical feature combination lies in the fact that, for the first time, the two behaviors are decoupled and reconstructed at the underlying dynamic level, providing a precise controlled object for upper-level control.

[0017] Step 3: Based on the dual-behavior dynamic model obtained in Step 2, P control and PD composite control are designed to meet the differentiated control requirements of car following and overtaking conditions. The control terms are simplified by introducing the speed proportional coefficient of adjacent lanes, and the control logic is integrated with the overtaking switch to finally construct an integrated dual-mode feedback control model containing control terms.

[0018] Step 3.1, Design of control laws for different operating conditions:

[0019] ① Following the car's trajectory ( Proportional control (P) is used, and the control quantity is: ;in, This is a proportional control parameter;

[0020] ② Overtaking condition ( ): PD control is adopted, incorporating the acceleration of vehicles in adjacent lanes. The original PD control expression is: ;in, These are differential control parameters; It provides the acceleration for vehicles in adjacent lanes. The introduction of the derivative term allows it to adjust the control quantity in advance based on the rate of change of the vehicle's acceleration, effectively suppressing speed overshoot and oscillations during overtaking acceleration or lane changing.

[0021] Step 3.2, Define The acceleration information of vehicles in adjacent lanes is mapped to the ratio of the vehicle's own acceleration, thus matching the dimensions of the differential control term with the system. Finally, the aforementioned control term is embedded into the unified dynamic model of step 2, forming an integrated control equation containing control inputs. The core of this step lies in achieving adaptive matching between the control strategy and driving conditions through a combination of "P control + PD control + proportional coefficient" techniques.

[0022] Step 4: Based on the integrated control dynamics equations from Step 3, a linear stability analysis method is adopted. The system stability is deduced through time scaling, disturbance assumptions, series expansion, etc., and the neutral stability conditions and traffic flow stability constraints of the system are solved to achieve adaptive matching of control parameters.

[0023] Step 4.1: Time-scale the delay term in the integrated control equation of Step 3, and assume that a small perturbation is applied to the uniform traffic flow. .

[0024] Step 4.2: By performing a series expansion on the disturbance characteristic equation, the key coefficients that determine the system stability are obtained. According to the linear stability theory ( (System stability), and derive unified system stability constraints.

[0025] Step 4.3, based on the overtaking switch The value of differentiates the unified stability condition into:

[0026] ① Stability conditions for car-following mode: ;

[0027] ② Stable conditions for overtaking mode: .

[0028] The technical effect of this step is to increase the proportional gain. Differential gain Sensitivity coefficient These key parameters provide the theoretically optimal matching range, fundamentally suppressing the chaotic phenomena that may be caused by overtaking disturbances.

[0029] Step 5: During real-time operation, the system uses the overtaking switch... The value (determined by sensing data) automatically selects the corresponding dynamic model and control law, and calls the optimized parameters that satisfy the stability constraints in step four. Since both the underlying dynamic model and the upper-level control law are based on a unified switching function architecture, the control signal remains continuous at the switching moment, thereby achieving a smooth, shock-free switching between following and overtaking conditions.

[0030] In some possible implementations, step 1 specifically involves:

[0031] Step 1.1: Equip the vehicle with onboard perception sensors to collect three core status data: real-time vehicle position, speed, and acceleration.

[0032] Step 1.2: Simultaneously collect data on the distance between the front vehicles and the vehicle in front of the vehicle in this lane, as well as their driving status.

[0033] Step 1.3: Collect additional speed and acceleration data of vehicles in adjacent lanes;

[0034] Step 1.4: Summarize all collected data to form a complete road condition and vehicle status dataset, and input it into the subsequent modeling and control module.

[0035] Autonomous driving's following and overtaking behaviors rely on full-domain driving perception data. Multi-dimensional state variables are the foundation for dynamic modeling, condition judgment, and control parameter calculation. The beneficial effects of step 1 are: achieving full-scenario driving environment perception, supplementing all input variables required for the following and overtaking conditions, ensuring the normal operation of subsequent modeling and control logic, and improving control accuracy from the data source level.

[0036] In some possible implementations, step 2 specifically involves:

[0037] Step 2.1, Build the basic optimal speed following model:

[0038] ①Based on the data of vehicle frontage and speed, establish the basic dynamic expression:

[0039]

[0040] in, , and They represent The car is Position and velocity at any given moment Indicates the driver's sensitivity coefficient;

[0041] ② Define the optimal velocity function:

[0042]

[0043] in, and These are the maximum speed and the safe distance, respectively.

[0044] Step 2.2, design the emergency braking scenario and introduce overtaking-related parameters:

[0045] ① Introduce an emergency braking term and adapt it for dual-behavior optimization, defining the basic emergency braking term:

[0046]

[0047] ② Incorporate attenuation coefficients based on overtaking behavior characteristics. With overtaking switch Build an improved emergency braking item:

[0048]

[0049] ③ Definition of overtaking switch:

[0050]

[0051] Step 2.3: Construct independent dynamic models for following and overtaking respectively:

[0052] ① Car-following mode: Introducing a coordinated speed correction term for both preceding vehicles to construct a dynamic model for car-following mode:

[0053]

[0054] in, This is the weighting coefficient for emergency braking; Indicates the previous vehicle With the vehicles behind The headway between the trains; Indicates the previous vehicle With the vehicles behind The headway between the trains; For the preceding vehicle information coefficient;

[0055] ② Overtaking Mode: Based on the expectation of overtaking, a dynamic model of the overtaking mode is constructed:

[0056]

[0057] in, In anticipation of overtaking;

[0058] Step 2.4, integrate the switching functions and establish a unified dual-behavior dynamic model: based on the overtaking switch By integrating the two operating condition equations, an integrated basic model is obtained:

[0059]

[0060] Traditional models cannot distinguish the dynamic differences between car following and overtaking. Step 2, based on the optimal speed model and incorporating driving characteristics such as emergency braking effect, historical acceleration decay, multiple preceding vehicle perception, and overtaking expectation, uses a switching function to decouple and fuse the two driving conditions. The beneficial effects of Step 2 are: accurately reproducing the different driving patterns of car following and overtaking, weakening the influence of historical acceleration during overtaking, taking into account the characteristics of emergency braking, solving the shortcomings of traditional models that are too simplistic and cannot adapt to dual driving behaviors, and building a precise dynamic platform for hierarchical control strategies.

[0061] In some possible implementations, step 3 specifically involves:

[0062] Step 3.1. Design the basic control law for each working condition:

[0063] ① Following-car driving condition: Proportional P control is used, and the control variable is: ;in, This is a proportional control parameter;

[0064] ② Overtaking condition: PD control is used, incorporating the acceleration of vehicles in adjacent lanes. The original PD control expression is: ;in, These are differential control parameters; Acceleration of vehicles in adjacent lanes;

[0065] Step 3.2, introduce proportional coefficients to simplify control terms:

[0066] Define the speed ratio coefficient of adjacent lanes : According to safety constraints By combining the overtaking switch, a unified feedback control term is obtained:

[0067]

[0068] Step 3.3: Integrate the dynamic model and control terms to obtain the final integrated control equation: Embed the feedback control term into the unified dynamic model of Step 2 to obtain the complete integrated control model.

[0069]

[0070] Following the car represents steady-state driving, requiring only proportional control to smoothly adjust the speed difference; overtaking is a dynamic lane-changing process, requiring a differential element to predict acceleration changes and suppress dynamic deviations; introducing a proportional coefficient simplifies the dimensions of adjacent lane parameters, reducing computational complexity. The beneficial effects of step 3 are: enabling differentiated control under different operating conditions, ensuring smooth driving with following the car, and improving dynamic control accuracy with overtaking; simplifying the control algorithm and reducing the computational load on the onboard terminal; and simultaneously, merging the control layer and model layer by connecting the dynamic model through dual-mode control terms.

[0071] In some possible implementations, step 4 specifically involves:

[0072] Step 4.1, time scaling transformation, simplifying the dynamic formula:

[0073] ① Introducing time scaling transformation Simplify and improve the emergency braking item: ,in, This is the time scaling factor;

[0074] ② Order The original dynamic equations are simplified to:

[0075]

[0076] Step 4.2, Uniform Flow Assumption and Small Perturbation Modeling:

[0077] ① Determine the total length of the road Total number of vehicles The optimal speed is Evenly flow down the front spacing of the car The vehicle in time The initial position is represented as: ;

[0078] ②Introduction of the first The vehicle was slightly disturbed Actual vehicle location: ;

[0079] ③ Define the disturbance form ,right Perform a series expansion Solving for the problem yields:

[0080]

[0081] Step 4.3, Derive the neutral stability condition and system stability constraints:

[0082] ① According to the linear stability theory: The system is stable. The system becomes unstable, leading to the derivation of the neutral stability condition:

[0083]

[0084] ② Further, the stability constraints of the traffic flow system are obtained:

[0085]

[0086] Step 4.4, simplify the stability conditions for each operating condition:

[0087] The overtaking mode is:

[0088] ;

[0089] Follow mode is:

[0090]

[0091] Step 4 utilizes the uniform flow assumption and the theory of small disturbances, combined with the linear stability criterion, to solve for the stability boundaries of the control parameters and dynamic parameters, thus clarifying the parameter value ranges. The beneficial effect of Step 4 is that it obtains the optimal feedback gain and system stability constraints, achieving… , , , The adaptive matching of parameters theoretically suppresses traffic chaos and traffic flow vibration caused by overtaking and sudden braking, ensuring the stability of the control system across the entire domain.

[0092] In some possible implementations, step 5 specifically involves:

[0093] Step 5.1: Read the sensing data in real time, combine it with the driving status to determine the vehicle's operating condition, and update the overtaking switch. Values:

[0094] Vehicle was detected performing car-following behavior: Set ;

[0095] The vehicle was detected overtaking: ;

[0096] Step 5.2, according to The values ​​are matched to the corresponding control strategy, and the optimized stable control parameters from step four are invoked:

[0097] The system switches to P control and executes car-following steady-state driving control;

[0098] The system switches to PD control to perform dynamic and precise overtaking control.

[0099] Step 5.3: Relying on the integrated model and continuous control output, a seamless transition between the two control modes is achieved, and control commands are output to the vehicle actuators.

[0100] Step 5 uses the overtaking switch as the trigger signal for the operating condition switch. Relying on the integrated model and continuous control law, it avoids sudden changes in control quantities during the operating condition switch. At the same time, the parameters follow stability constraints to ensure system stability throughout the switch. The beneficial effects of Step 5 are: to achieve automatic recognition and smooth switching of control modes for following and overtaking behaviors, eliminating driving shocks and jerks caused by operating condition switches; and to comprehensively improve the safety, smoothness, and road traffic efficiency of autonomous driving by combining the modeling, control, and stability optimization results mentioned above.

[0101] The beneficial effects of this invention are as follows: This invention constructs a dual-mode differentiated dynamic model adapted to car-following and overtaking, introduces emergency braking attenuation and multi-vehicle perception mechanisms, effectively improving modeling accuracy and conforming to real autonomous driving driving conditions. Simultaneously, it adopts a P / PD hierarchical adaptive control strategy to ensure driving smoothness in steady-state car-following conditions and control accuracy in dynamic overtaking conditions, solving the industry problems of poor adaptability and significant traffic flow swaying associated with single control strategies. This invention innovatively introduces an adjacent lane speed ratio coefficient, effectively simplifying control parameters, reducing algorithm computational complexity, and significantly improving the real-time control efficiency of the onboard terminal. Furthermore, it relies on linear stability analysis to accurately identify the system's stability boundary, effectively suppressing traffic flow oscillations and driving disturbances by constraining the optimal feedback gain. During numerical simulation, Lyapunov analysis is introduced to further verify the system's stable state, ultimately achieving adaptive and smooth switching between car-following and overtaking conditions, eliminating control jerks, adapting to the needs of autonomous driving in all scenarios, and effectively stabilizing traffic flow and improving overall road traffic efficiency. Attached Figure Description

[0102] Figure 1 This is a schematic diagram of the vehicle position in the model of this invention;

[0103] Figure 2 Here are the neutral stability curves of the model of this invention for different parameters of car-following behavior: ;

[0104] Figure 3 Here are the neutral stability curves of the model of this invention for different parameters of overtaking behavior: ;

[0105] Figure 4 This is a spatiotemporal evolution diagram of the headway of the model of this invention under different parameters of car-following behavior: ;

[0106] Figure 5 This is a spatiotemporal evolution diagram of the headway of the model of this invention under different parameters of overtaking behavior: ;

[0107] Figure 6 These are hysteresis loop images of the model of this invention with different parameters in car-following behavior: ;

[0108] Figure 7 These are hysteresis loop images of the model of this invention for different parameters of overtaking behavior: ;

[0109] Figure 8 The difference in car-following behavior is due to the model of this invention. The Lyapunov exponent image and the front-end distance image are used to obtain the values;

[0110] Figure 9 The difference in car-following behavior is due to the model of this invention. The Lyapunov exponent image and the front-end distance image are used to obtain the values;

[0111] Figure 10 The different overtaking behaviors of the model of this invention The Lyapunov exponent image and the front-end distance image are fixed. ;

[0112] Figure 11 The different overtaking behaviors of the model of this invention The Lyapunov exponent image and the front-end distance image are fixed. ;

[0113] Figure 12 The different overtaking behaviors of the model of this invention The Lyapunov exponent image and the front-end distance image are fixed. . Detailed Implementation

[0114] The principles and features of the present invention are described below with reference to the accompanying drawings. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0115] like Figure 1 As shown, this invention provides an adaptive switching control method for following and overtaking in autonomous vehicles based on PD control, comprising the following steps:

[0116] Step 1: Real-time collection of vehicle status and road condition information. The vehicle's real-time position, speed, and acceleration, as well as the distance between the front and rear vehicles and their driving status, are obtained through onboard sensing sensors. At the same time, the speed and acceleration information of vehicles in adjacent lanes are collected to provide data support for dual-behavior recognition and control adjustment.

[0117] Step 2: Construct a basic car-following dynamics model for the vehicle based on the optimal speed model, introduce the overtaking switch function and the emergency braking attenuation coefficient, and establish differentiated dynamic equations for the car-following mode and the overtaking mode respectively to achieve accurate modeling of the two driving behaviors.

[0118] The basic optimal velocity model and dual-behavior dynamics modeling method include:

[0119] Establish the optimal speed-based car-following model, with the following dynamic expression:

[0120]

[0121] in: , and They represent The car is Position and velocity at any given moment The optimal speed function expression represents the driver's sensitivity coefficient:

[0122]

[0123] and These are the maximum speed and the safe distance, respectively.

[0124] Introducing an emergency braking term and adapting it for dual-behavior optimization, defining the basic emergency braking term:

[0125]

[0126] To address the weakening effect of historical acceleration during overtaking, a damping coefficient is introduced. With overtaking switch Build an improved emergency braking item: , The overtaking switch is defined as follows:

[0127]

[0128] In car-following mode, leveraging the advantages of multiple preceding vehicles' perception, a dual preceding vehicle speed coordination correction term is introduced to construct a dynamic model for car-following mode:

[0129]

[0130] in, This is the weighting coefficient for emergency braking. Indicates the previous vehicle With the vehicles behind The headway between the trains; Indicates the previous vehicle With the vehicles behind The headway between the trains; Information coefficient of the vehicle ahead

[0131] In overtaking mode, a dynamic model of overtaking mode is constructed based on the overtaking expectation:

[0132]

[0133] in, The expectation is to overtake.

[0134] By integrating the overtaking switch function, a unified dual-behavior-based dynamic model is constructed:

[0135]

[0136] Step 3: To address the control requirements of different driving behaviors under emergency braking, a hierarchical adaptive control strategy is designed. Proportional P control is used for following the car, and proportional-derivative PD composite control is used for overtaking. The speed proportional coefficient of adjacent lanes is introduced to simplify the control terms, and an integrated dual-mode feedback control model is constructed.

[0137] Methods for constructing dual-mode adaptive feedback control strategies include:

[0138] Differentiated control strategies are designed to address the control requirements of different driving behaviors under emergency braking: For car-following behavior, only steady-state adjustment of the speed difference is required, so proportional P control is adopted, and the control variable is: Overtaking maneuvers require high-precision dynamic control. PD (Power Generation) control is employed, incorporating vehicle acceleration information during lane-changing intervals in adjacent lanes to construct the original PD control expression: Introducing a speed ratio coefficient for adjacent lanes. Define the relation: Based on actual overtaking safety constraints, the following limits are imposed. Preferred As a baseline parameter for simulation and control, the dual-mode control logic is integrated with the overtaking switching function and simplified to obtain a unified feedback control term: .in This is a proportional control parameter. These are the differential control parameters.

[0139] The final integrated control model dynamic equations are as follows:

[0140]

[0141] Step 4: Combine linear stability analysis method to perform stability deduction on integrated dynamic control model, derive the optimal feedback gain of system and traffic flow stability constraints, and realize adaptive matching of control parameters.

[0142] The derivation is carried out using linear stability analysis, and the specific steps are as follows:

[0143] (1) To facilitate linear stability analysis, a time scaling transformation is introduced. .in, This is the time scaling factor, used to characterize the time scale of historical acceleration response. At this point... It can be simplified to: The dynamic formula then simplifies to the following equation:

[0144]

[0145] in: .

[0146] (2) Assume the distance between vehicles is And the corresponding optimal speed is To maintain a uniform flow. Therefore, each vehicle in time The initial position can be represented as: , This indicates the length of the road. This represents the number of vehicles. Assume the [number]th vehicle on the road... The vehicle was slightly disturbed The location of the vehicle can be described as follows, depending on the influence of the surrounding environment: This yields the following formula:

[0147]

[0148] in: , , .make ,get:

[0149]

[0150] make , ,get and coefficient:

[0151]

[0152] (3) According to the system linear stability theory, when When the system becomes unstable, it will become unstable; and when When the system remains in a stable state, the neutral stability condition can be solved as follows:

[0153]

[0154] Based on the stability conditions of a traffic system, the following conditions are obtained for system stabilization when a traffic system experiences disturbances:

[0155]

[0156] when The neutral stability condition simplifies to:

[0157]

[0158] when The neutral stability condition simplifies to:

[0159]

[0160] Step 5: Based on the overtaking switch function, identify the vehicle's driving conditions in real time, and adaptively switch between the following P control and overtaking PD control modes to achieve a smooth and shock-free switch between steady-state following driving and safe overtaking operations for autonomous vehicles.

[0161] Figure 2 (a)-(b) represent the fixed parameters when the car-following behavior is employed. And change the preceding vehicle information coefficient and fixed parameters Furthermore, the neutral stability curve of the proportional control parameter. From Figure 2 It can be seen that when the parameters are fixed At this time, the stability of traffic flow becomes more stable as the information coefficient of the preceding vehicle increases. Furthermore, under a fixed... Increasing the proportional control parameter will make traffic flow more stable.

[0162] Figure 3 (a)-(b) represent overtaking maneuvers with fixed parameters. And changing overtaking expectations and fixed parameters Furthermore, the neutral stability curves of the proportional control parameter and the derivative control parameter are obtained. From... Figure 3 It can be seen that when the parameters are fixed At that time, the stability of traffic flow deteriorates as the expectation of overtaking increases. Furthermore, in a fixed... At the same time, increasing the proportional control parameter can effectively smooth out speed fluctuations caused by overtaking and improve traffic flow stability. Furthermore, at a fixed... Increasing the differential control parameters can precisely regulate the driving deviation during the overtaking dynamic process, further optimize the traffic flow stability performance under overtaking conditions, and effectively suppress overtaking disturbances.

[0163] Experimental verification:

[0164] Numerical simulations were performed on the established autonomous vehicle car-following control model under open boundary conditions.

[0165] The initial traffic flow is set as follows:

[0166] In numerical simulations, in order to demonstrate the effectiveness of the model, Figures 4 to 12 The simulation results show the traffic evolution under the conditions of following and overtaking.

[0167] Figure 4 and Figure 5 This is a spatiotemporal evolution diagram of the headway of the model of the present invention under different parameters of following and overtaking behaviors.

[0168] Figure 4 This demonstrates how vehicle 55 performs under car-following behavior with different parameters. and The change in vehicle headway during the measurement process. Other parameters are: , and The time span is 29,700 to 29,800 time steps. Observations Figure 4 (a) It is evident that, with As the information coefficient of the vehicle increases, the fluctuation range of the distance between the vehicles gradually decreases, indicating that increasing the information coefficient of the vehicle in front is beneficial. This helps suppress fluctuations in the front-end distance and improve vehicle stability. (Observation) Figure 4 (b)(fixed) It can be seen that, with As the value increases, the fluctuation in the distance between the front ends of the vehicles also shows a decreasing trend, indicating that the proportional control parameters... The increase in headway can also effectively reduce headway oscillations and enhance traffic flow stability.

[0169] Figure 5 This demonstrates how vehicle 55 behaves with different parameters when overtaking occurs. , and The change in vehicle headway during the measurement process. Other parameters are set as follows: , , and The time span is 29,700 to 29,800 time steps. Observations Figure 5 (a)(fixed) , , It can be seen that, with As the distance between vehicles increases, the fluctuation range of the headway gradually increases, indicating that when overtaking is expected... Increasing the size can lead to system instability. Figure 5 (b)(fixed) , , It can be seen that increasing This reduces the fluctuation in headroom, indicating that proportional control can effectively mitigate the instability caused by overtaking when there is overtaking behavior and no differential control is available. Figure 5 (c)(fixed) , , It can be seen that increasing The headway fluctuation was further reduced, indicating that PD control can further enhance the stability of the system and suppress the adverse effects of overtaking when overtaking occurs.

[0170] Figure 6 and Figure 7 These are hysteresis loop images of different parameters of the model in the following and overtaking behaviors of this invention.

[0171] Figure 6 The fixed parameters are given in the car-following behavior. , and The evolution pattern of the hysteresis loop of the 55th vehicle. From Figure 6 (a) It can be seen that, with the parameter As the information coefficient of the preceding vehicle increases, the annular region enclosed by the hysteresis loop gradually shrinks, indicating that increasing the information coefficient of the preceding vehicle... It contributes to the stability of the transportation system. Figure 6 (b) It can be seen that, with As the value increases, the hysteresis loop region also shows a shrinking trend, indicating that increasing the proportional control parameter... It has a significant positive regulating effect on traffic flow stability and can effectively enhance the system's ability to suppress disturbances.

[0172] Figure 7 The following parameters are given for overtaking scenarios. , , and The evolution of the hysteresis loop of the 55th vehicle. (From...) Figure 7 (a)(fixed) , It can be seen that, with An increase in will significantly expand the hysteresis region, indicating that the overtaking expectation... An increase in [something] will exacerbate traffic flow instability. From Figure 7 (b) It can be seen that when , Increased time This reduces the hysteresis loop region, indicating that proportional control plays a positive role in mitigating instability caused by overtaking when overtaking occurs and differential control is absent. Furthermore, from Figure 7 (c) when , Increased time This further reduces the hysteresis loop region, indicating that the introduction of differential control can more precisely regulate the traffic system and suppress instability caused by overtaking. In summary, when overtaking occurs, PD control can further enhance system stability and effectively suppress the adverse effects of overtaking.

[0173] Figure 8 Under the condition of car-following behavior ( , , , , The image obtained by drawing. Figure 8 (a) indicates different The Lyapunov index calculated from the value ; Figure 8 (b) indicates a difference The evolution of the distance between vehicle heads under certain values. From Figure 8 (a) It can be seen that, When the Lyapunov exponent fluctuates, it indicates that the system has weak stability and a tendency towards instability or chaotic evolution; combined with Figure 8 (b) It can be found that, When moving left and right, the distance between the front and rear of the vehicle changes accordingly. The increase in shows a certain convergence trend, but the overall fluctuation is large, and a stable state has not yet been formed. When When around, the Lyapunov index Furthermore, the front-end distance eventually converges to the desired safe front-end distance. This indicates that the system is gradually entering a stable operating state. Overall, the parameters... An increase in the preceding vehicle information coefficient helps enhance the stability of the traffic flow system. The larger the value, the stronger the system's ability to suppress disturbances and the more stable the system's operating state.

[0174] Figure 9 Under the condition of car-following behavior ( , , , , The image obtained by drawing. Figure 9 (a) showcased different The Lyapunov index calculated from the value ; Figure 9 (b) indicates a difference The evolution of the vehicle headway under certain values. From Figure 9 (a) It can be seen that, When the Lyapunov exponent is in a state of alternating positive and negative oscillation, it indicates that the system has weak stability and a tendency towards instability or chaotic evolution; combined with Figure 9 (b) It can be observed that when When moving left and right, the distance between the front and rear of the vehicle changes accordingly. The increase in shows a certain convergence trend, but has not yet reached a stable state. When When around, the Lyapunov index Furthermore, the front-end distance eventually converges to the desired safe front-end distance. This indicates that the system is gradually entering a stable operating state. Overall, the parameters... The increase in the value helps to enhance the stability of the traffic flow system, further confirming the key role of proportional control in suppressing disturbances and improving system stability.

[0175] Figure 10 Under the condition that overtaking behavior exists ( , , , , , , The image obtained by drawing. Figure 10 (a) shows in and Different parameters The Lyapunov index calculated from the value ; Figure 10 (b) indicates that in and Different times The evolution of the vehicle headway under certain values. From Figure 10 (a) It can be seen that, with As the value increases, the Lyapunov exponent alternates between positive and negative values, and the system exhibits chaotic dynamic characteristics. Combined with... Figure 10 (b) It can be observed that, with As the distance between vehicles increases, the fluctuation range of the headway widens, and the traffic flow becomes highly unstable. This indicates that strong overtaking behavior induces complex nonlinear dynamic characteristics. Overall, in the presence of overtaking behavior, the expected overtaking speed... An increase in will weaken the stability of the traffic flow system and significantly enhance the chaotic characteristics of the system.

[0176] Figure 11 Under the condition that overtaking behavior exists ( , , , , , , The image obtained by drawing. Figure 11 (a) indicates that in and Different parameters The Lyapunov index calculated from the value ; Figure 11 (b) indicates that in and Different times The evolution of the vehicle headway under different values. Comparison. Figure 10 (a) and Figure 11 (a) It can be seen that when added Subsequently, the Lyapunov exponent value decreased significantly, and the alternating positive and negative oscillation characteristics weakened markedly. This indicates that the addition of proportional control can effectively suppress the unstable dynamic behavior caused by overtaking behavior and has a positive effect on improving system stability. Therefore, P control has a significant effect on mitigating the instability caused by overtaking behavior. Furthermore, a comparison... Figure 10 (b) and Figure 11 (b) It can be seen that the introduction of proportional control parameters Even in the future As the distance between vehicles increases, the fluctuation range of the headway also decreases significantly, indicating that P control can effectively suppress the oscillations caused by overtaking behavior and has a positive effect on improving traffic flow stability.

[0177] Figure 12 Under the condition that overtaking behavior exists ( , , , , , , The image obtained by drawing. Figure 12 (a) indicates that in and Different parameters The Lyapunov index calculated from the value ; Figure 12 (b) indicates that in and Different times The evolution of the vehicle headway under certain values. Comparison. Figure 11 (a) and Figure 12 (a) It can be seen that when adding differential control parameters Later, when At that time, the Lyapunov exponent no longer exhibited obvious alternating positive and negative oscillations, and the chaotic phenomenon was eliminated; when At that time, with As the value increases, the fluctuation range of the Lyapunov exponent further decreases. This indicates that introducing derivative control on top of proportional control can further enhance the stability of the system and suppress unstable dynamic behavior caused by overtaking. (Comparison) Figure 11 (b) and Figure 12 (b) It can be seen that when differential control parameters are introduced... Later, when At that time, the distance between the front ends of the vehicles can be stably converged to the desired safe distance between the front ends of the vehicles. .when At that time, with The increase in size has led to a slight increase in the front-end distance, but compared to... Figure 11 The fluctuation amplitude in (b) is significantly reduced. This further demonstrates that adding D control to P control can effectively suppress headway oscillations and improve system stability. In conclusion, the PD control strategy can make traffic flow more stable and effectively suppress the adverse effects of overtaking behavior.

[0178] comprehensive Figures 10 to 12 It can be seen that, with the expected parameters for overtaking... The increase, when uncontrolled ( Figure 10 The Lyapunov exponent of the system oscillates alternately between positive and negative, and the headway fluctuates wildly, exhibiting obvious chaotic characteristics and strong instability; after introducing proportional control ( Figure 11 The Lyapunov exponent value decreased significantly, the headway fluctuation amplitude decreased significantly, and the chaotic phenomenon was effectively suppressed; after further introducing differential control ( Figure 12 The Lyapunov index fluctuation range further weakened, and the headway was at It converges stably to the expected value at time. The time-varying fluctuations were also significantly suppressed. The above evolution process shows that the PD control strategy can effectively suppress the nonlinear instability caused by overtaking behavior and significantly improve the stability performance of the traffic flow system.

[0179] Technical effect

[0180] By organically combining the above-mentioned technical solutions, the present invention achieves the following significant technical effects:

[0181] 1. Significantly improves driving safety and stability: By constructing an adaptive dynamic model and hierarchical control strategy, the distortion problem of traditional models under overtaking conditions is effectively solved. The introduction of linear stability constraints ( Figure 2-3 The numerical simulation verified the control system's ability to suppress disturbances such as sudden braking and overtaking. Figure 4-7 The results showed that the fluctuations in the distance between vehicle heads and the area of ​​the hysteresis loop were significantly reduced, effectively curbing the chaotic phenomenon in traffic flow.

[0182] 2. Balancing driving smoothness and control precision: By employing a differentiated strategy of P control in car-following mode and PD control in overtaking mode, the system simultaneously meets the requirements for steady-state following comfort during car-following and the requirements for dynamic response precision during overtaking. Lyapunov index analysis ( Figure 8-12 The study verified that the strategy can effectively reduce the chaotic characteristics of the system and enable the headway to converge quickly to the desired value.

[0183] 3. Improve system real-time performance and engineering applicability: By introducing a speed ratio coefficient between adjacent lanes. Simplifying the control terms reduces the computational complexity of the algorithm, facilitating real-time deployment on in-vehicle terminals. Simultaneously, the theoretically derived stability conditions provide clear guidance for parameter tuning, avoiding trial-and-error and enhancing the system's robustness.

[0184] Invention Point

[0185] The core difference between this invention and existing publicly available technical solutions lies in:

[0186] 1. Difference in Modeling Methods: This invention proposes for the first time a "unified dynamic modeling method based on overtaking switches," which introduces an overtaking switch... With attenuation coefficient It achieves a precise mathematical description and smooth transition between the following and overtaking modes in a single equation, overcoming the shortcomings of traditional models that use two independent equations or a single averaging model, resulting in low accuracy or large switching shocks.

[0187] 2. Combination and Derivation of Control Strategies: This invention combines "follow-the-car P control, overtaking PD control" with "adjacent lane speed ratio coefficient". This combination forms a simple, adaptive hierarchical control scheme. This scheme is not a simple choice between P and PD, but rather cleverly integrates multi-dimensional environmental information into the control law through proportional coefficients, using the speed proportional coefficients of adjacent lanes. This method maps the speed information of adjacent lanes, which is difficult to measure directly, to a ratio of the vehicle's measurable speed, thereby avoiding the engineering difficulties of directly obtaining the acceleration of adjacent lanes and making PD control feasible in actual vehicle systems.

[0188] 3. Decoupling of Stability Constraints by Operating Condition: In linear stability analysis, this invention ultimately derives "explicit system stability constraints based on different operating conditions." This technical feature unifies the stability criteria of the model with specific driving behaviors (…). The values ​​of the parameters are correlated, enabling the control parameters to adaptively achieve the theoretical optimal match based on real-time operating conditions. This is a significant innovation that distinguishes it from traditional fixed parameter or global stability analysis.

[0189] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for adaptive switching control between following and overtaking in an autonomous vehicle, characterized in that, The steps include the following: Step 1: Collect operational data of the entire vehicle and surrounding vehicles and lanes through onboard perception sensors to obtain multi-dimensional driving status information, providing raw data support for subsequent dual-driving behavior modeling, operating condition identification and control adjustment; Step 2: Based on the vehicle status, spacing, and speed data collected in Step 1, a basic car-following dynamics model is built on the basis of the classic optimal speed model; overtaking switching function and emergency braking attenuation coefficient are introduced to establish exclusive dynamic equations for car-following and overtaking modes respectively, and finally merged to obtain a unified dual-behavior dynamics model, completing the accurate mathematical modeling of the two driving behaviors. Step 3: Based on the dual-behavior dynamic model obtained in Step 2, P control and PD composite control are designed to meet the differentiated control requirements of car following and overtaking conditions respectively; the control terms are simplified by introducing the speed proportional coefficient of adjacent lanes, and the control logic is integrated with the overtaking switch to finally construct an integrated dual-mode feedback control model containing control terms. Step 4: Based on the integrated control dynamics equations from Step 3, the linear stability analysis method is adopted. The system stability is deduced through time scaling, disturbance assumptions, series expansion, etc., and the neutral stability conditions and traffic flow stability constraints of the system are solved to achieve adaptive matching of control parameters. Step 5: Based on the real-time perception data from Step 1 and the overtaking switch function from Step 2, the current driving condition of the vehicle is identified in real time. Combined with the stability parameter constraints obtained in Step 4, the vehicle adaptively switches between two modes: following P control and overtaking PD control, so as to achieve a smooth and shock-free switching between following and overtaking conditions.

2. The adaptive switching control method for following and overtaking in autonomous vehicles according to claim 1, characterized in that, Step 1 is as follows: Step 1.1: Equip the vehicle with onboard perception sensors to collect three core status data: real-time vehicle position, speed, and acceleration. Step 1.2: Simultaneously collect data on the distance between the front vehicles and the vehicle in front of the vehicle in this lane, as well as their driving status. Step 1.3: Collect additional speed and acceleration data of vehicles in adjacent lanes; Step 1.4: Summarize all collected data to form a complete road condition and vehicle status dataset, and input it into the subsequent modeling and control module.

3. The adaptive switching control method for following and overtaking in autonomous vehicles according to claim 2, characterized in that, Step 2 is as follows: Step 2.1, Build the basic optimal speed following model: ①Based on the data of vehicle frontage and speed, establish the basic dynamic expression: in, , and They represent The car is Position and velocity at any given moment Indicates the driver's sensitivity coefficient; ② Define the optimal velocity function: in, and These are the maximum speed and the safe distance, respectively. Step 2.2, design the emergency braking scenario and introduce overtaking-related parameters: ① Introduce an emergency braking term and adapt it for dual-behavior optimization, defining the basic emergency braking term: ② Incorporate attenuation coefficients based on overtaking behavior characteristics. With overtaking switch Build an improved emergency braking item: ③ Definition of overtaking switch: Step 2.3: Construct independent dynamic models for following and overtaking respectively: ① Car-following mode: Introducing a coordinated speed correction term for both preceding vehicles to construct a dynamic model for car-following mode: in, This is the weighting coefficient for emergency braking; Indicates the previous vehicle With the vehicles behind The headway between the trains; Indicates the previous vehicle With the vehicles behind The headway between the trains; For the preceding vehicle information coefficient; ② Overtaking Mode: Based on the expectation of overtaking, a dynamic model of the overtaking mode is constructed: in, In anticipation of overtaking; Step 2.4, integrate the switching functions and establish a unified dual-behavior dynamic model: based on the overtaking switch By integrating the two operating condition equations, an integrated basic model is obtained:

4. The adaptive switching control method for following and overtaking in an autonomous vehicle according to claim 3, characterized in that, Step 3 specifically involves: Step 3.

1. Design the basic control law for each working condition: ① Following-car driving condition: Proportional P control is used, and the control variable is: ;in, This is a proportional control parameter; ② Overtaking condition: PD control is used, incorporating the acceleration of vehicles in adjacent lanes. The original PD control expression is: ;in, These are differential control parameters; Acceleration of vehicles in adjacent lanes; Step 3.2, introduce proportional coefficients to simplify control terms: Define the speed ratio coefficient of adjacent lanes : According to safety constraints By combining the overtaking switch, a unified feedback control term is obtained: Step 3.3: Integrate the dynamic model and control terms to obtain the final integrated control equation: Embed the feedback control term into the unified dynamic model of Step 2 to obtain the complete integrated control model.

5. The adaptive switching control method for following and overtaking in an autonomous vehicle according to claim 4, characterized in that, Step 4 specifically involves: Step 4.1, time scaling transformation, simplifying the dynamic formula: ① Introducing time scaling transformation Simplify and improve the emergency braking item: ,in, This is the time scaling factor; ② Order The original dynamic equations are simplified to: Step 4.2, Uniform Flow Assumption and Small Perturbation Modeling: ① Determine the total length of the road Total number of vehicles The optimal speed is Evenly flow down the front spacing of the car The vehicle in time The initial position is represented as: ; ②Introduction of the first The vehicle was slightly disturbed Actual vehicle location: ; ③ Define the disturbance form ,right Perform a series expansion Solving for the problem yields: Step 4.3, Derive the neutral stability condition and system stability constraints: ① According to the linear stability theory: The system is stable. The system becomes unstable, leading to the derivation of the neutral stability condition: ② Further, the stability constraints of the traffic flow system are obtained: Step 4.4, simplify the stability conditions for each operating condition: The overtaking mode is: ； Follow mode is:

6. The adaptive switching control method for following and overtaking in an autonomous vehicle according to claim 5, characterized in that, Step 5 specifically involves: Step 5.1: Read the sensing data in real time, combine it with the driving status to determine the vehicle's operating condition, and update the overtaking switch. Values: Vehicle was detected performing car-following behavior: Set ; The vehicle was detected overtaking: ; Step 5.2, according to The values ​​are matched to the corresponding control strategy, and the optimized stable control parameters from step four are invoked: The system switches to P control and executes car-following steady-state driving control; The system switches to PD control to perform dynamic and precise overtaking control. Step 5.3: Relying on the integrated model and continuous control output, a seamless transition between the two control modes is achieved, and control commands are output to the vehicle actuators.

Images