Multi-strategy car-following control system based on vehicle-mounted edge cloud cooperative architecture

Abstract

This invention discloses a multi-strategy traffic following control system based on an onboard edge-cloud collaborative architecture, comprising: an onboard control unit, a roadside communication identification unit, and a cloud-based control management platform. The roadside communication identification unit is deployed at key road nodes to identify road types and communication status in real time, and issues control strategy switching commands based on these real-time identifications. The onboard control unit integrates an adaptive sliding mode controller and a nonlinear self-stabilizing backstepping controller, used for following control on straight and curved road sections, respectively. The cloud-based control management platform periodically updates controller parameters using historical data and pushes this information to the onboard control unit and the roadside communication identification unit. This invention achieves automatic switching of control strategies based on the road structure type where the vehicle is located, employing a hybrid weighted fusion control structure to ensure a smooth transition between different controller signals during switching, thereby improving the control accuracy and stability of mixed vehicle fleets in complex traffic environments.

Description

Technical Field

[0001] This invention belongs to the field of intelligent transportation systems and vehicle control technology, and particularly relates to a multi-strategy traffic following control system based on an on-board edge-cloud collaborative architecture. Background Technology

[0002] With the development of vehicle-to-everything (V2X) and autonomous driving technologies, connected vehicles (CAVs) are gradually operating in the same traffic environment as traditional human-driven vehicles (HDVs), forming a "mixed traffic flow." Against this backdrop, fleet following control systems have become an important research direction for maintaining traffic flow stability and improving traffic efficiency.

[0003] Currently, most mainstream vehicle following control systems are based on the following technical approaches:

[0004] Single control strategy models: such as adaptive cruise control (ACC) and cooperative adaptive cruise control (CACC) based on a fixed headway. These systems maintain platoon stability by setting a fixed distance or time interval. A representative technology, such as the CACC model, can obtain the status of the vehicle in front through V2V communication and dynamically adjust speed and acceleration to achieve longitudinal control of the vehicles.

[0005] Sliding mode control (SMC) and its variants: These exhibit good robustness against model uncertainties and external disturbances, such as adaptive sliding mode control and fuzzy sliding mode control. While these control methods can adapt to a certain degree of disturbance, they generally suffer from control output chattering, which affects ride comfort and shortens actuator lifespan.

[0006] Nonlinear control methods, such as backstepping, are suitable for the control design of traffic systems with strong nonlinear characteristics and can improve the stability of the system in complex dynamic environments. However, traditional backstepping methods are mostly used for closed trajectory tracking and are insufficient in response to situations such as sudden changes in road curvature and the influence of lateral forces.

[0007] Model Predictive Control (MPC): It has the ability to handle multivariate constraints and can plan control inputs in advance by predicting future states. However, it has high computational complexity, relies on powerful computing platforms, and is not suitable for real-time deployment in vehicle environments.

[0008] The structural flaws of the existing solution are as follows:

[0009] Lack of multi-strategy integration capability: Existing systems mostly adopt a single control strategy, which is designed only for specific road conditions (such as straight lines or gentle slopes). They cannot adaptively switch according to road type during vehicle operation, resulting in a significant decrease in control accuracy and stability in complex road scenarios (such as curves, uphill and downhill slopes).

[0010] The control structure lacks a distributed coordination mechanism: some technologies focus on vehicle-side processing of control logic, failing to fully utilize roadside (RSU) and cloud resources for collaborative computing and policy scheduling, resulting in slow response speed and performance degradation when communication is limited.

[0011] Incomplete signal judgment mechanism: Most current systems lack a real-time recognition mechanism for changes in road type, which leads to delays or misjudgments in control strategy switching, causing safety hazards such as sudden acceleration and deceleration of vehicles and violent queue fluctuations.

[0012] In summary, there is an urgent need for a multi-strategy traffic following control system based on an in-vehicle edge-cloud collaborative architecture to solve the problems of stable operation and expansion of intelligent fleets in mixed traffic environments. Summary of the Invention

[0013] To address the aforementioned technical problems, this invention proposes a multi-strategy traffic following control system based on an in-vehicle edge-cloud collaborative architecture, thereby resolving the issues present in the prior art.

[0014] To achieve the above objectives, the present invention provides a multi-strategy traffic following control system based on an in-vehicle edge-cloud collaborative architecture, comprising:

[0015] Vehicle-mounted control unit, roadside communication and identification unit, and cloud-based control and management platform;

[0016] The roadside communication identification unit is deployed at key road nodes to identify road type and communication status in real time, and issues control strategy switching commands based on the real-time identified road type and communication status.

[0017] The vehicle control unit integrates an adaptive sliding mode controller and a nonlinear self-stabilizing backstepping controller, which are used for car-following control on straight road sections and curved road sections, respectively.

[0018] The cloud-based control and management platform is used to periodically update controller parameters using historical data and push them to the vehicle control unit and the roadside communication identification unit.

[0019] Optionally, the roadside communication identification unit includes a road type identification module and a control strategy scheduling module;

[0020] The road type identification module is used to obtain current road type change information;

[0021] The control strategy scheduling module is used to calculate the probability of successful communication and generate a control strategy switching command based on road type change information, which is then sent to the vehicle control unit.

[0022] Optionally, the vehicle control unit further includes a data input interface, a control output interface, and a strategy switching module;

[0023] The data input interface is used to connect to the vehicle perception system to obtain the relative position, speed and acceleration of the vehicle in front;

[0024] The control output interface is used to connect to the vehicle drive actuator;

[0025] The strategy switching module is used to switch control strategies according to the control strategy switching command.

[0026] Optionally, the adaptive sliding mode controller and the nonlinear self-stabilizing backstepping controller are deployed in parallel and connected to the vehicle perception system through the data input interface, respectively.

[0027] Optionally, expressions for the car-following distance error and the speed error are constructed based on a micro-car-following model of the mixed flow in a straight lane. Based on these expressions, a sliding surface formula is constructed by linearly combining error terms. Using the micro-car-following model of the mixed flow in a straight lane and the sliding surface formula, the sliding surface dynamic equation is obtained by differentiating the sliding surface. Based on this dynamic equation, the adaptive sliding mode controller is obtained, as follows:

[0028]

[0029] Among them, v n Let p be the speed of the nth vehicle, V(·) be the optimal speed of the vehicle, and p be the speed of the nth vehicle. n p is the lateral clearance coefficient between two adjacent vehicles. c Δx represents the penetration rate of autonomous vehicles. n This is the distance between the front ends of two adjacent vehicles. Represents the Wiener process, Δv n,n+1 (t) and Δv n,n+2 (t) represents the speed difference between the nth car, the (n+1)th car, and the (n+2)th car, respectively; X1 is the following distance error; X2 is the speed error; 1 < ε < 2. k2 > 0, s is the sliding surface, b om α is a very small positive number, ξ1 and ξ2 are correlation coefficients, and 0 < ξ1 < ξ2 < 1, λ1 ∈ (0,1), λ2 ∈ (0,1), σ is the noise intensity, X1 is the following distance error, and X2 is the speed error.

[0030] Optionally, the nonlinear self-stabilizing backstepping controller is obtained based on the microscopic car-following model of the mixed flow on the curve and the desired car-following distance. The nonlinear self-stabilizing backstepping controller is as follows:

[0031]

[0032] Among them, v nLet p be the speed of the nth vehicle, V(·) be the optimal speed of the vehicle, and p be the speed of the nth vehicle. n p is the lateral clearance coefficient between two adjacent vehicles. c To increase the penetration rate of driverless cars, Representing the Wiener process, and These represent the speed differences between the nth car, the (n+1)th car, and the (n+2)th car, respectively, where r is the radius of curvature of the curve, and ω... n (t) represents the angular velocity of the vehicle. For the vehicle's acceleration, Z1 is the distance between two adjacent vehicles on the curve, Z2 is the vehicle position error, Z3 and K4 are both positive numbers, α is the sensitivity coefficient, and ξ is the distance between two adjacent vehicles on the curve. 1，2 and λ 1，2 Here, σ is the correlation coefficient, and σ is the noise intensity. It is the expected acceleration.

[0033] Optionally, the desired carousel distance is:

[0034]

[0035] In the formula, δ is the safety clearance; v i Indicates the speed of the i-th vehicle; Indicates the positive correlation coefficient; t h The headway is the distance between the front of the train, where, t0 is the base headway; m v and m a Sensitivity coefficient; t h,max The maximum headway; Δv i and Δa i These represent the speed difference and acceleration difference between two adjacent vehicles, respectively.

[0036] Optionally, the strategy switching module determines whether to start weight fusion based on the control strategy switching command. If started, it dynamically weights and fuses the outputs of the adaptive sliding mode controller and the nonlinear self-stabilizing backstepping controller through an exponential weight smoothing function based on a preset switching window period, and obtains a fused control command to control the vehicle drive actuator through the control output interface.

[0037] Optional, the exponential weighted smoothing function is as follows:

[0038]

[0039] in, u is the estimated probability of successful communication at present. ASMC For the output of the adaptive sliding mode controller, u NSBSC The output is for a nonlinear self-stabilizing backstepping controller.

[0040] Compared with the prior art, the present invention has the following advantages and technical effects:

[0041] This invention introduces two complementary controllers—Adaptive Sliding Mode Controller (ASMC) and Nonlinear Feedback Stability Controller (NSBSC)—and designs a mechanism for automatically identifying road types and switching controllers. Through real-time identification of road curvature and issuance of commands by the Roadside Unit (RSU), the Vehicle Control Unit (OBU) can quickly switch controllers. In straight-line scenarios, ASMC achieves stable distance and speed control, while in curved scenarios, NSBSC maintains lateral stability and prevents loss of following stability. This strategy switching mechanism allows the vehicle to maintain control continuity when entering different road types, avoiding sudden speed changes and discomfort caused by control jumps, and significantly improving driving smoothness in mixed traffic conditions.

[0042] During the strategy switching window, the output weight of the original controller is gradually reduced while the output weight of the target controller is increased. Dynamic stability is further enhanced through a linear unmeasurable disturbance suppression mechanism combined with output fusion. The control strategy switching employs an exponential weighted fusion curve to achieve a smooth output transition between ASMC and NSBSC, effectively avoiding drastic output changes caused by controller switching.

[0043] In real-world mixed traffic platoons, disturbances caused by the acceleration or deceleration of the vehicle in front are amplified vehicle by vehicle, forming "shock waves" and "sparse waves." The control strategy switching mechanism and control structure proposed in this invention effectively mitigate the propagation effect of this phenomenon. Attached Figure Description

[0044] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0045] Figure 1 This is a block diagram of the system structure of an embodiment of the present invention;

[0046] Figure 2 This is a flowchart of the controller policy scheduling process according to an embodiment of the present invention. Detailed Implementation

[0047] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0048] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0049] Example 1

[0050] like Figure 1 As shown, this embodiment provides a multi-strategy traffic following control system based on an in-vehicle edge-cloud collaborative architecture, including:

[0051] Vehicle-mounted control unit, roadside communication and identification unit, and cloud-based control and management platform;

[0052] The roadside communication identification unit is deployed at key road nodes to identify road type and communication status in real time, and issues control strategy switching commands based on the real-time identified road type and communication status.

[0053] The roadside communication identification unit includes a road type identification module and a control strategy scheduling module;

[0054] The road type identification module is used to obtain current road type change information;

[0055] The control strategy scheduling module is used to calculate the probability of successful communication and generate a control strategy switching command based on road type change information, which is then sent to the vehicle control unit.

[0056] Specifically, a) Road type recognition module: perceives the current road shape (such as curvature and slope) in real time through preset map data or sensors; b) Control strategy scheduling module: based on communication status and road type judgment, sends control strategy switching instructions to OBU; and connects with the cloud platform and vehicle OBU through a wireless communication module.

[0057] The vehicle control unit integrates an adaptive sliding mode controller and a nonlinear self-stabilizing backstepping controller, which are used for car-following control on straight and curved road sections, respectively; it also includes a data input interface, a control output interface, and a strategy switching module; the data input interface is used to connect to the vehicle perception system to obtain the relative position, speed, and acceleration of the vehicle in front; the control output interface is used to connect to the vehicle drive actuator; the strategy switching module is used to switch the control strategy according to the control strategy switching command.

[0058] In one specific implementation, the adaptive sliding mode controller and the nonlinear self-stabilizing backstepping controller are deployed in parallel and connected to the vehicle perception system through the data input interface, respectively.

[0059] As a specific implementation method, expressions for car-following distance error and speed error are constructed based on a micro-car-following model of mixed flow on a straight lane. Based on the expressions for car-following distance error and speed error, a sliding surface formula is constructed by linearly combining error terms. Based on the micro-car-following model of mixed flow on a straight lane and the sliding surface formula, the sliding surface dynamic equation is obtained by differentiating the sliding surface. The adaptive sliding controller is obtained based on the sliding surface dynamic equation.

[0060] Specifically, the longitudinal dynamics of each vehicle are simplified into a second-order system with position and velocity states. Through V2V communication, each vehicle can instantly obtain the relative position and velocity of the vehicle ahead. Bounded uncertainties are used to describe changes in external disturbances, such as road friction and sensor noise. The microscopic car-following model of the mixed flow in a straight lane can be described as follows:

[0061]

[0062] Among them, v n Let p be the speed of the nth vehicle, V(·) be the optimal speed of the vehicle, and p be the speed of the nth vehicle. n p is the lateral clearance coefficient between two adjacent vehicles. c Δx represents the penetration rate of autonomous vehicles. n This is the distance between the front ends of two adjacent vehicles. Represents the Wiener process, Δv n,n+1 (t) and Δv n,n+2 (t) represents the speed difference between the nth car, the (n+1)th car, and the (n+2)th car, respectively, and the other parameters represent coefficients greater than zero.

[0063] The microscopic car-following model of mixed flow on a curve can be described as follows:

[0064]

[0065] Where r is the radius of curvature of the curve, ω n (t) represents the angular velocity of the vehicle. For the vehicle's acceleration, This refers to the distance between two adjacent vehicles on a curve.

[0066] Controller design:

[0067] To more accurately describe real-world traffic scenarios, variable headway is used to further describe the driver's expected following distance d, expressed as:

[0068]

[0069] In the formula, δ is the safety clearance; v i Indicates the speed of the i-th vehicle; Indicates the positive correlation coefficient; t h The headway is expressed as:

[0070]

[0071] In the formula, t0 is the base headway; m v and m a Sensitivity coefficient; t h,max The maximum headway; Δv i and Δa i This represents the speed difference and acceleration difference between two adjacent vehicles.

[0072] In this embodiment, based on a heterogeneous traffic flow car-following model and the desired following distance, an adaptive sliding mode controller (ASMC) is designed for straight roads to enable the target vehicle to follow the vehicle in front at the desired car-following distance. The specific design steps include:

[0073] Step 1: To resist control system failure due to actuator malfunction, an adaptive sliding mode controller (ASMC) and a time-varying fault deviation function are implemented. In a car-following system, this is represented as:

[0074]

[0075] Step 2: Define the following distance error X1 and the speed error X2:

[0076]

[0077] In the formula, λ z It is the safety factor of the road environment, where ζ, ζ and ψ are all positive real numbers.

[0078] Step 3: Define the sliding surface s n (t):

[0079]

[0080] Where, 1 < ε < 2,

[0081] Combining formula (1) and differentiating it with respect to the sliding surface, we can obtain:

[0082]

[0083] in,

[0084]

[0085]

[0086] The following adaptive sliding mode controller (ASMC) is designed to address the problem of control system failure caused by actuator malfunction.

[0087]

[0088] Among them, b om It is a very small positive number.

[0089] In this embodiment, based on a heterogeneous traffic flow car-following model and the desired following distance, a nonlinear self-stabilizing backstepping controller (NSBSC) is designed on a curve to enable the target vehicle to follow the vehicle in front at the desired speed. The specific design steps include:

[0090] Step 1: Introduce a control law into the mixed traffic flow model system (2) under the sloping and curved road environment, which is expressed as:

[0091]

[0092] Where, τ F It refers to the driver's reaction time.

[0093] Step 2: Define the vehicle's position error:

[0094] Z1 = x n (t)-x d (14)

[0095] Define the first Lyapunov function The derivative of the Lyapunov function is expressed as:

[0096] Introducing virtual control variables And define virtual error Where k3 > 0, For x d The time derivative. When the vehicle position error Z1 = 0. equal to the expected speed Based on this, the derivative of L1 can be expressed as:

[0097] Step 3: To eliminate The Z3Z4 item makes Define another Lyapunov function L2:

[0098]

[0099] The derivative of L2 is:

[0100] To ensure the stability of heterogeneous traffic flow car-following systems, it is required that... Therefore, the expression for the designed nonlinear self-stabilizing backstepping controller (NSBSC) is:

[0101]

[0102] In this embodiment, the edge-cloud architecture based on distributed control includes an environment-aware policy switching mechanism. Specifically, the RSU smoothly switches between the ASMC (Straight Road) and NSBSC (Narrow Road Scaffolding) controllers by dynamically classifying road types and communication conditions. This overcomes the limitations of using fixed controllers in existing studies and greatly improves the system's adaptability to various road modes. The specific design steps include:

[0103] Step 1: Determine the communication probability of RSU:

[0104]

[0105] If Q n If the value is greater than 0.8, the vehicle is considered to be within the effective coverage area of ​​the RSU and undergoes road condition classification. The RSU then classifies the road curvature, and the vehicle switches control strategies accordingly.

[0106]

[0107] Smooth transition control mechanism design:

[0108]

[0109] in, This is an estimate of the probability of successful communication at present.

[0110] The cloud-based control and management platform is used to periodically update controller parameters using historical data and push them to the vehicle control unit and the roadside communication identification unit.

[0111] Specifically, the cloud-based control and management platform has remote learning and control optimization functions;

[0112] Collect historical traffic data and operating conditions, and update controller parameters regularly; communicate with each RSU via the public network.

[0113] Control process such as Figure 2 As shown, it includes:

[0114] When a vehicle enters a straight section of road: the RSU detects that the road curvature is less than a set threshold; it sends a "straight control" command to the OBU; the OBU activates the ASMC module, which dynamically adjusts the target headway based on the relative distance and speed difference between the vehicle and the vehicle in front; the control output is transmitted to the acceleration / deceleration module to achieve steady-state following.

[0115] When a vehicle enters a curved section: the RSU determines that the road curvature exceeds a set threshold; it sends a "curve control" command to the OBU; the OBU switches to the NSBSC module to build a feedback control chain with virtual control variables; the control output can automatically adjust the vehicle's angular velocity and longitudinal acceleration to smoothly pass through the curve and maintain queue stability.

[0116] Control strategy switching mechanism:

[0117] To avoid vehicle vibration caused by control jumps, this system designs an exponential weighted smoothing function to fuse the outputs between ASMC and NSBSC; within the switching window (approximately 1-2 seconds), the output is dynamically weighted and synthesized to achieve a continuous transition.

[0118] Example:

[0119] In a typical operational scenario, a mixed convoy consisting of manually driven vehicles numbered V1-V5, V6, V7, and V9, and autonomous vehicles numbered V2, V3, V4, V8, and V10, entered a curve section from a straight section. The RSU (Roadside Unit) identified the curvature change in advance and determined that the communication signal strength was acceptable. About 15 meters before V2 entered the curve, the OBU (On-Board Unit) received the strategy switching signal, and the controller smoothly transitioned from ASMC (Autonomous System Controlled Control) to NSBSC (Non-Standard Bus Controlled Control). Inside the curve, the headway of the vehicles was maintained within 20 ± 0.1 m, and the speed fluctuation was less than 0.1 m / s. After exiting the curve, the strategy automatically switched back to ASMC.

[0120] This embodiment introduces a road environment recognition module and control state scheduling mechanism from the roadside unit (RSU), enabling the vehicle control system to intelligently switch to the optimal control strategy based on the current road type, thereby improving car-following accuracy and queue stability. It constructs a multi-layered collaborative control architecture (vehicle-edge RSU-cloud), fully utilizing the computing resources of the roadside and cloud to achieve dynamic updates, global optimization, and rapid local response of controller parameters, overcoming the real-time bottleneck caused by the limitation of vehicle-side computing resources in traditional systems. A smooth controller switching mechanism is introduced, using a weighted function for transition fusion, eliminating abrupt shocks during control strategy switching, ensuring smooth speed changes and controllable distance fluctuations, and improving driving comfort and safety. It is suitable for environments where autonomous driving and human-driven vehicles coexist, enhancing robustness to sudden traffic disturbances and improving overall traffic flow stability and efficiency. This invention significantly improves the intelligent car-following control capability of vehicles in complex road and traffic scenarios, possessing strong engineering application prospects and promotional value.

[0121] The above are merely preferred embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A multi-strategy car-following control system based on a vehicle-mounted edge cloud cooperative architecture, characterized in that, Comprise: Vehicle-mounted control unit, road side communication identification unit and cloud control management platform; The road side communication identification unit is deployed at the key nodes of the road, for real-time identification of road type and communication state, and sending control strategy switching instructions according to the real-time identification of road type and communication state; The vehicle-mounted control unit integrates an adaptive sliding mode controller and a nonlinear self-stabilizing backstepping controller, respectively used for following control on straight road segments and curved road segments; The strategy switching module in the vehicle-mounted control unit judges whether to start weight fusion according to the control strategy switching instructions, and if so, based on a preset switching window period, dynamically weights and fuses the outputs of the adaptive sliding mode controller and the nonlinear self-stabilizing backstepping controller through an exponential weight smoothing function, to obtain a fusion control instruction to control the vehicle drive actuator through a control output interface; The exponential weight smoothing function is as follows: ； wherein, is a current communication success probability estimate value, is an adaptive sliding mode controller output, is a nonlinear self-stabilizing backstepping controller output; The cloud control management platform is used to update the controller parameters regularly through historical data, and push them to the vehicle-mounted control unit and the road side communication identification unit.

2. The multi-strategy traffic following control system based on the vehicle-mounted edge cloud collaborative architecture according to claim 1, wherein The road side communication identification unit comprises a road type identification module and a control strategy scheduling module; The road type identification module is used to obtain current road type change information; The control strategy scheduling module is used to calculate the communication success probability, and generate a control strategy switching instruction in combination with the road type change information, and send it to the vehicle-mounted control unit; The communication module interacts with the OBU through the V2X link, and connects with the cloud through the 4G / 5G network.

3. The multi-strategy traffic following control system based on the vehicle-mounted edge cloud collaborative architecture according to claim 2, wherein The vehicle-mounted control unit further comprises a data input interface, a control output interface and a strategy switching module; The data input interface is used to connect the vehicle perception system to obtain the relative position, speed and acceleration of the preceding vehicle; The control output interface is used to connect the vehicle drive actuator; The strategy switching module is used to switch the control strategy according to the control strategy switching instruction.

4. The multi-strategy traffic following control system based on the vehicle-mounted edge cloud collaborative architecture according to claim 3, wherein The adaptive sliding mode controller and the nonlinear self-stabilizing backstepping controller are deployed in parallel, and are connected with the vehicle perception system through the data input interface respectively.

5. The multi-strategy traffic following control system based on the vehicle-mounted edge cloud collaborative architecture according to claim 4, wherein Based on the microscopic following model of mixed flow on straight lanes, the following distance error expression and the speed error expression are constructed; Based on the following distance error expression and the speed error expression, the sliding mode surface formula is constructed by linearly combining the error terms, the sliding mode surface dynamic equation is obtained by deriving the sliding mode surface based on the microscopic following model of mixed flow on straight lanes and the sliding mode surface formula, and the adaptive sliding mode controller is obtained based on the sliding mode surface dynamic equation, which is as follows: ； wherein, is the speed of the i-th vehicle, is the optimal speed of the vehicle, is the lateral gap coefficient of the adjacent two vehicles, is the penetration rate of the autonomous vehicle, is the headway of the adjacent two vehicles, denotes a Wiener process, and denote the speed difference of the i-th vehicle and the j-th vehicle, the i-th vehicle and the k-th vehicle, respectively, , , , , is the car following headway error, is the speed error, , , , , is a sliding mode surface, is a small positive number, is a sensitivity coefficient, , is a correlation coefficient, and , , , is the intensity of the noise, is the car following headway error, is the speed error.

6. The multi-strategy traffic following control system based on the vehicle-mounted edge cloud collaborative architecture according to claim 1, wherein The nonlinear self-stabilizing backstepping controller is obtained based on a microscopic car-following model of mixed traffic flow on a curve and an expected car-following distance, and the nonlinear self-stabilizing backstepping controller is as follows: ； in, For the first The speed of the car The optimal speed for the vehicle. This is the lateral clearance coefficient between two adjacent vehicles. To increase the penetration rate of driverless cars, Representing the Wiener process, and They represent the first Car and the vehicle, number The speed difference of the vehicles Let be the radius of curvature of the curve. Let be the angular velocity of the vehicle. For the vehicle's acceleration, This represents the distance between two adjacent vehicles on a curve. This represents the vehicle's positional error. This is a virtual error. and All are positive numbers. This is the sensitivity coefficient. and The correlation coefficient, It is the intensity of the noise. It is the expected acceleration. 7.The multi-strategy car-following control system based on the vehicle-mounted edge cloud cooperative architecture according to claim 6, characterized in that, The expected car-following distance is: ； wherein is the safety distance; denotes the speed of the vehicle; denotes the positive correlation coefficient; is the headway at the time of the vehicle passing the , is the basic headway; and is the sensitivity coefficient; is the maximum headway; and denote the speed difference and the acceleration difference between the two adjacent vehicles, respectively.

Images