A pure electric tractor train multi-constraint cooperative ACC control method

Abstract

The present application belongs to the field of road vehicle control system, and relates to a kind of pure electric towing vehicle train multi-constraint collaborative ACC control method, the method constructs the unified input of pure electric towing vehicle train ACC control multi-source information, comprehensively utilizes road preview information, mixed traffic state, transportation task constraint and double-sided electric drive / energy storage execution ability, and plans reference speed and reference acceleration in advance;Construct conservative-enthusiastic double-strategy ACC control mechanism for mixed traffic environment, and adaptively adjust the following strategy according to traffic risk, task state, double-sided energy storage and hinged stability, finally, the expected acceleration is coordinated and distributed as traction drive, trailer auxiliary drive, traction regenerative braking, trailer regenerative braking and mechanical braking control quantity, to realize the safety, energy saving and task constraint collaborative control of pure electric towing vehicle train, so as to reduce the electric energy consumption while ensuring the safety of following and the stability of vehicle, improve the brake energy recovery efficiency and transportation task adaptability.

Description

Technical Field

[0001] This invention belongs to the field of road vehicle control systems, and relates to intelligent driving and longitudinal control of commercial vehicles, specifically to a multi-constraint cooperative ACC control method for pure electric trailer trains. Background Technology

[0002] With the continuous improvement of electrification and intelligence in long-haul logistics transportation, pure electric trailer trucks, formed by combining pure electric tractors and semi-trailers, are increasingly being applied to medium- and long-distance freight, port collection and distribution, and fixed-route transportation. Compared with passenger cars, trailer trucks are characterized by larger overall mass, greater inertia, longer body length, longer braking distance, and more pronounced coupled motion between the tractor and semi-trailer. Compared with traditional fuel-powered commercial vehicles, the power output, energy recovery, driving range, and refueling plans of pure electric trailer trucks are influenced by factors such as the state of charge (SOC) of the power battery, battery charging and discharging power, motor torque capacity, and regenerative braking capacity. Therefore, directly adopting the ACC (Adaptive Cruise Control) control strategy used for ordinary vehicles or traditional fuel-powered commercial vehicles on pure electric trailer trucks makes it difficult to simultaneously consider following safety, traffic efficiency, energy consumption control, and the requirements for completing transportation tasks.

[0003] Existing ACC systems typically perform follow-the-car control based on the distance to the vehicle ahead, relative speed, and the vehicle's own speed, focusing on maintaining a safe following distance and reducing speed errors. For ACC control of semi-trailers or tractor-trailers, existing technologies have begun to consider the impact of vehicle configuration on longitudinal control. For example, Chinese patent CN118124568B achieves longitudinal cruise control for semi-trailers by acquiring information such as the target vehicle's real-time speed, vehicle structural state, and desired acceleration, combined with control relationships under different vehicle structural states and PID control outputting engine torque. This method can improve the acceleration response of semi-trailers under different structural states, but it primarily targets traditional power systems, focusing on engine torque control and not fully considering the SOC (state of charge) of the power battery, battery charging and discharging power, regenerative braking capability, and road-oriented energy-saving control requirements of pure electric tractor-trailers.

[0004] For example, Chinese patent CN121019663A calculates the risk factor based on information such as the average speed of the tractor wheels, the average speed of the trailer wheels, and the angle between the tractor and trailer, and implements deceleration, collision avoidance, or alarm strategies according to the risk level to improve the safety of the tractor train during cruising. This method focuses on the impact of the combined motion state of the tractor and trailer on safety risks, but it mainly addresses the problem of judging the risk level of the vehicle combination and graded deceleration, without considering the pre-emptive information such as road slope, curvature, speed limit changes, and traffic flow speed changes in conjunction with the energy consumption of pure electric vehicles, regenerative braking, and transportation task constraints.

[0005] With the development of electric semi-trailers, electric axle trailers, and trailers with energy storage, pure electric trailer trains are no longer necessarily structured with the tractor as the sole power source and the trailer passively following. Some vehicles can be equipped with power batteries, energy storage systems, or electric axles on the trailer side, enabling the trailer to not only participate in regenerative braking but also potentially participate in driving or auxiliary traction under specific operating conditions. In this case, vehicle energy management is no longer just a one-sided constraint problem of the tractor's power battery SOC, maximum discharge power, and maximum charging power; it requires simultaneous consideration of the energy storage states on both the tractor and trailer sides, charge / discharge power boundaries, electric axle torque capacity, and regenerative braking capability. If a single power battery power model is still used, treating the trailer as a passive load, it is difficult to accurately describe the energy flow relationship between the tractor and trailer, and it is also difficult to fully utilize the driving and energy recovery capabilities provided by the trailer-side energy storage system and electric axle.

[0006] Existing ACC (Adaptive Cruise Control) methods for pure electric commercial vehicles or trailers typically assume that regenerative braking energy is recovered by the tractor's battery, or simply differentiate between regenerative and mechanical braking at the initial braking execution stage. There are few methods that establish a unified power distribution and SOC (State of Charge) recursive model for tractor-trailer dual energy storage structures. For trailers with energy storage systems or electric drive axles, existing methods do not fully consider the coordinated distribution of regenerative braking force between the tractor and trailer, and lack a control mechanism to dynamically adjust the regenerative braking ratio based on the tractor's SOC, trailer SOC, allowable charging power on both sides, regenerative braking torque capacity, and articulation angle. Therefore, under braking conditions such as downhill, reduced speed limits, and deceleration of the preceding vehicle, problems such as the tractor's battery solely handling energy recovery, underutilization of trailer-side braking energy, excessive mechanical braking intervention, and unreasonable regenerative braking distribution can easily occur, reducing energy recovery efficiency and potentially adversely affecting the longitudinal and articulated stability of the trailer-trailer train.

[0007] Meanwhile, pure electric trailer trucks typically operate in mixed traffic environments with complex vehicle types and behaviors. Existing ACC (Adaptive Cruise Control) systems do not adequately consider factors such as cut-in risks, communication reliability, and traffic disturbances. Fixed-following strategies struggle to simultaneously ensure safety, traffic efficiency, and energy recovery efficiency. Furthermore, existing ACC systems usually serve short-term single-vehicle follow-along control, rarely considering the vehicle's overall transport mission. They fail to incorporate factors such as estimated arrival time, remaining mileage, State of Charge (SOC), target energy consumption, and refueling plans into cruise decisions, easily leading to inconsistencies between local follow-along control and the overall transport mission objectives.

[0008] In summary, while existing technologies have addressed longitudinal ACC control, vehicle combination motion risks, and graded deceleration control for semi-trailers or tractor-trailers, they still suffer from several issues: insufficient utilization of road preview information, inadequate consideration of energy constraints in pure electric power systems, fixed car-following strategies in mixed traffic environments, insufficient involvement of transportation task constraints, and insufficient coordination between upper-level desired acceleration and lower-level electric / regenerative / mechanical braking execution. In particular, for pure electric trailer-trailers with energy storage systems or electric drive axles, existing methods have not yet developed a unified modeling and collaborative control scheme for the dual energy sources of the tractor and trailer, lacking comprehensive consideration of SOC recursion, charging and discharging power boundaries, regenerative braking capability, driving force distribution, regenerative braking force distribution, and articulated stability constraints on both the tractor and trailer sides. Therefore, it is necessary to propose an ACC control method for pure electric trailer-trailers. Summary of the Invention

[0009] In view of the shortcomings and deficiencies of the existing technology, the purpose of this invention is to provide a multi-constraint cooperative ACC control method for pure electric trailer-trailer trains. This method comprehensively utilizes road preview information, mixed traffic conditions, transportation task constraints, and the electric drive / energy storage execution capabilities of both the tractor and trailer sides to plan reference speed and reference acceleration in advance. Based on traffic risks, task status, dual-side energy storage status, and articulated stability, the car-following strategy is adaptively adjusted. Finally, the desired acceleration is coordinated and distributed as control quantities for tractor drive, trailer auxiliary drive, tractor regenerative braking, trailer regenerative braking, and mechanical braking. This achieves safe, energy-saving, and task-constraint cooperative control of pure electric trailer-trailer trains, thereby reducing energy consumption, improving braking energy recovery efficiency, and enhancing transportation task adaptability while ensuring car-following safety and vehicle stability.

[0010] To achieve the above objectives, the present invention adopts the following technical solution: A multi-constraint cooperative ACC control method for pure electric trailer-trailer trains includes the following steps: Step S1: Construct a unified multi-source information input for the ACC control of pure electric trailer truck trains, including the status of the vehicle and surrounding vehicles, road preview information, energy status of the tractor and trailer, and transportation task constraint information; Step S2: Based on the data obtained in Step S1, traffic flow status, and vehicle combination mass, predict the future longitudinal state. Then, introduce acceleration correction to establish a phase-aware longitudinal resistance and tractor-trailer dual energy storage power model. At the same time, construct the road pre-aiming energy-saving planning objective function and plan the energy-saving ACC reference speed and reference acceleration in advance in the prediction time domain through rolling optimization. Step S3: Construct a conservative-active dual-strategy ACC control mechanism for mixed traffic environments; identify the primary car-following target and potential cutting-in targets, calculate the cutting-in risk, the cooperative stability of the primary car-following target and the intensity of mixed traffic disturbance, and determine the fusion coefficient of the conservative-active dual strategy accordingly, thereby determining the expected ACC distance at the current moment and calculating the safe distance corresponding to the primary car-following target. Step S4: Generate the basic car-following acceleration and fuse it with the energy-efficient ACC reference acceleration; use the current time's desired ACC acceleration. To optimize the variables, an ACC acceleration objective function is constructed, consisting of a term ensuring car-following safety, a term meeting the speed requirements of the transportation task, a term inheriting the energy-saving results of road pre-aiming, a term for tracking and fusing the initial acceleration, and a term limiting acceleration mutations. The expected acceleration of the ACC at the current moment is obtained by solving a single-step constraint optimization problem. Step S5: Drive, regenerative braking and mechanical braking are executed in coordination according to the ACC desired acceleration.

[0011] As a preferred embodiment of the present invention, the vehicle status in step S1 includes the vehicle's longitudinal position, speed, longitudinal acceleration, yaw angle, yaw rate, articulation angle between the tractor and the semi-trailer, equivalent mass of the vehicle, and state of charge of the power battery; the surrounding vehicle status includes the longitudinal position, lateral position, longitudinal speed, longitudinal acceleration, lateral speed, and communication reliability of the vehicles in front of the vehicle and vehicles in adjacent lanes; the road prediction information includes the location of the prediction point and the road slope, road curvature, road speed limit, traffic flow reference speed, and curve safety speed at each prediction point. The information includes: speed and road speed limits; tractor-side energy status, including the tractor's battery state of charge, maximum allowable discharge power, maximum allowable charging power, maximum drive torque, and maximum regenerative braking torque; trailer-side energy status, including the trailer's energy storage system state of charge, maximum allowable discharge power, maximum allowable charging power, maximum drive torque of the trailer's electric drive axle, and maximum regenerative braking torque; and transportation task constraint information, including remaining mileage, required arrival time, estimated arrival time, cargo type, target unit mileage energy consumption, available charging station or battery swapping station locations, and refueling time.

[0012] As a preferred embodiment of the present invention, in step S2, a phase-sensing macro-micro fusion longitudinal state prediction model based on Transformer is established, which predicts the current state at the current moment. forward The micro-tracking state sequence within each sampling time point, constructed based on road preview information, is the first... The macroscopic road-traffic feature vector corresponding to each prediction step, along with the energy states of the tractor and trailer sides, the electric drive axle capacity, and the vehicle combination mass, are used as prediction inputs to output the future... The initial longitudinal state sequence for the prediction step includes the initial predicted velocity, initial predicted acceleration, and the pure electric vehicle's operating phase probability. Based on the pure electric vehicle's operating phase probability, the vehicle's position in the prediction step is determined in advance. The prediction step is in the driving, coasting, regenerative braking or mechanical braking condition; then, the upper and lower boundaries of the prediction acceleration are determined according to the current comprehensive available discharge power and comprehensive available charging power of the tractor and trailer, the initial prediction acceleration is corrected by boundary correction, and smoothing correction is performed in combination with the acceleration change rate constraint to obtain the corrected acceleration, i.e., the prediction acceleration; then, the velocity and longitudinal position sequence in the prediction time domain are generated according to the corrected acceleration sequence.

[0013] As a preferred embodiment of the present invention, the road pre-aiming energy-saving planning objective function in step S2 includes a prediction prior term. Road speed constraints Energy Item Smoothness item The optimal acceleration correction sequence is obtained by minimizing this objective function. Furthermore, during the solution process, constraints are set for the planned speed, planned acceleration, rate of change of acceleration, battery power on the tractor side and trailer side, SOC on the tractor side and trailer side, and acceleration correction amount in the prediction time domain. By combining the forward slope, speed limit, curvature, battery status, and regenerative braking capability, the obtained predicted acceleration is subject to limited correction to obtain the road pre-aiming reference acceleration, i.e., the energy-saving ACC reference acceleration. Then, the speed correction amount is recursively obtained according to the optimal acceleration correction sequence to obtain the road pre-aiming reference speed, i.e., the energy-saving ACC reference speed. Predicting priors ; in, and These are the predicted velocity and predicted acceleration, respectively. and These are the planned speed and planned acceleration after road energy-saving corrections. ; This refers to the acceleration correction amount to be optimized in the road energy-saving planning. , It is the velocity correction amount, calculated based on the acceleration correction amount. To predict the step size; Road speed constraints ;in, For the first The permissible speed at each pre-aiming point; Energy Item ; Among them, the The discharge power of the tractor and trailer sides at each pre-aiming point are respectively , The regenerative braking power of the tractor and trailer sides is respectively , The equivalent power of mechanical braking is tractor and trailer sides They are respectively All of these are calculated based on the phase-sensing longitudinal resistance and the dual energy storage power model of the tractor-trailer; , , and They are respectively used to characterize discharge power penalty, recovered energy utilization, mechanical braking penalty, and bilateral. Equilibrium constraints and Determined based on the operating phase probability of pure electric vehicles and Offline calibration; Smoothness term ;in, The predicted acceleration of the previous prediction step. The sampling period.

[0014] As a preferred embodiment of the present invention, in step S3, based on the positional relationship between the vehicle and surrounding vehicles, each surrounding vehicle is divided into candidate following targets or potential cutting targets. For vehicles in the current lane of the vehicle and located in front of the vehicle, the vehicle with the closest longitudinal distance is selected as the primary following target; for the first vehicle in the set of potential cutting targets... For each vehicle, construct an entry risk feature vector. The offline calibrated logic function is used to calculate the first... Risk of entering the market for individual vehicles : ; in, The intercept parameter of the risk logic function is used to input the risk. This is the entry risk feature weight vector, used to describe the degree of influence of longitudinal distance, relative speed, lateral distance, lateral speed, lateral approach time, and steering / trajectory deflection indicators on entry risk. It is the transpose symbol; The intensity of mixed traffic disturbance is calculated based on the number of surrounding vehicles, average relative speed, and acceleration fluctuations within the sensing range. And identify the maximum entry risk among all potential entry targets. ; A cooperative stability evaluation quantity is constructed based on the communication reliability, velocity fluctuation, and acceleration fluctuation of the primary and secondary targets. ; Current ACC expected time interval for: ; in, The conservative-aggressive dual-strategy fusion coefficient is given by the expected time interval under the aggressive strategy. The expected time interval under the conservative strategy is ; The safe distance corresponding to the main following target for: ; in, For minimum stationary safe distance, This represents the current speed of the vehicle. The main focus is on the target speed. Assuming the vehicle in front has the maximum deceleration, This is a correction factor for the distance to the entry risk. This is the effective braking deceleration of the vehicle.

[0015] As a preferred embodiment of the present invention, in step S4, the distance error between the main following target and the vehicle is considered. and speed error Generate basic car-following acceleration ; ; in, , and To control the gain in the case of car-following, Accelerate towards the primary target. For communication reliability; Then, based on the risk of main follow-up. Determine the fusion coefficient between the energy-saving ACC reference acceleration and the base car-following acceleration. ; Initial ACC acceleration after fusion for: ; in, The road aiming reference acceleration output in step S2 is the energy-saving ACC reference acceleration. Adjustment factor for task speed; For task reference speed.

[0016] As a preferred embodiment of the present invention, the safety term for ensuring car-following in the ACC acceleration objective function in step S4 ; Meet the speed requirements for transportation tasks Inheriting the results of road pre-planning energy conservation ; Tracking and fusing the initial acceleration term Limiting acceleration mutation terms in, The longitudinal distance between the main target vehicle and the vehicle itself. The acceleration of this vehicle at the current moment. , and These are respectively: safety weight, task time weight, and energy-saving weight; and To integrate acceleration tracking and smoothness weights.

[0017] As a preferred embodiment of the present invention, in step S5, the required target wheel-end longitudinal force is first calculated based on the desired ACC acceleration, taking into account the vehicle's longitudinal dynamics. Then determine whether the vehicle is currently in a driving or braking condition; When the vehicle requires motor drive, the target driving force is allocated according to the electric drive system capacity of the tractor and trailer to obtain the target driving torque on the tractor and trailer sides. After combining the maximum driving torque and maximum discharge power on the tractor and trailer sides, the actual driving torque is obtained. When the vehicle needs to decelerate and brake, the required total braking force is determined; combined with the current maximum available regenerative braking force on the tractor and trailer sides, the actual regenerative braking force and corresponding regenerative braking torque on the tractor and trailer sides are calculated; when the regenerative braking capacity of the tractor and trailer is insufficient to meet the total braking force requirement, the remaining part is supplemented by mechanical braking.

[0018] As a further preferred embodiment of the present invention, the regenerative braking power on the tractor side, the regenerative braking power on the trailer side, and the equivalent power of mechanical braking are respectively: ; in, and The regenerative braking energy recovery efficiency is respectively for the tractor side and the trailer side; The actual regenerative braking forces at the pre-aiming points on the tractor side and trailer side are respectively Mechanical braking force is The regenerative braking distribution coefficient of the tractor is It is based on the tractor and trailer The difference, hinge angle, and hinge angular velocity are dynamically adjusted; At each pre-aiming point, the maximum available regenerative braking force on the tractor and trailer sides is respectively... The calculations were based on the maximum regenerative braking torque on the tractor side and trailer side, the maximum allowable charging power on the tractor side and trailer side, the transmission ratio on the tractor side and trailer side, the transmission efficiency, the regenerative braking energy recovery efficiency, the wheel rolling radius, and the planned speed after road pre-aiming energy saving correction; The total braking force required at each pre-aiming point is ; Discharge power on the tractor side and trailer side ; in, These refer to the driving forces on the tractor side and trailer side when the vehicle is in driving mode. and These refer to the drive efficiency of the electric drive systems on the tractor side and the trailer side, respectively.

[0019] As a further preferred embodiment of the present invention, in step S4, the average speed required to meet the transportation task at the current moment is calculated based on the remaining driving mileage and the required arrival time of the task, and the task reference speed is obtained by combining the current road speed limit. ; Calculate the urgency of the task based on the expected arrival time deviation. Then, based on the current road forecasting information and historical power consumption per unit mileage, the remaining power required for the route is estimated and calculated, thereby determining the amount of electricity needed to complete the remaining transportation task. At the same time, determine whether the current battery status is sufficient to complete the remaining tasks, and based on... Energy demand level is calculated using state difference. Then, the risk of main follow-the-leader is determined based on the fusion coefficient of the conservative and aggressive dual strategies. And based on the urgency of the task Energy demand level and the risk of following the main trend The safety weight, energy-saving weight, and task time weight in the ACC acceleration solution are corrected.

[0020] The advantages and beneficial effects of this invention are as follows: (1) In response to the large vehicle mass, strong inertia, and power response constraints imposed by the electric drive system and power battery status during ACC following of pure electric trailer trains, and the fact that existing ACC systems mostly rely on the distance to the preceding vehicle, relative speed, and the vehicle's own speed for local following control, usually only making passive adjustments after the vehicle's movement has been affected by slopes, curves, or speed limit changes, lacking advance utilization of the road's slope, curvature, speed limit changes, traffic flow speed, and the pure electric power system status, this invention introduces road preview information, surrounding traffic conditions, and power battery status into the ACC speed planning of pure electric trailer trains, and combines it with Transformer-based longitudinal state prediction and vehicle operation phase recognition to predict in advance whether the vehicle may be in a driving, coasting, regenerative braking, or mechanical braking state in the future. Compared with existing ACC methods that only output acceleration based on the current distance error and speed error, this invention can reasonably maintain kinetic energy before the vehicle enters an uphill section, avoiding short-term high-power drive in the slope; reduce drive output in advance before going downhill, on curves, or before the speed limit is reduced, and make full use of coasting and regenerative braking to complete speed adjustment. This reduces the need for rapid acceleration, deceleration, and mechanical braking, thereby lowering the energy consumption during the ACC control process of pure electric trailer trains, increasing driving range and energy utilization efficiency, and making the speed planning results more consistent with the operating characteristics of pure electric trailer trains, which have large mass, strong inertia, and limited power response.

[0021] (2) To address the issues of long braking distances, significant target switching impacts, and greater sensitivity to adjacent vehicle cut-offs in mixed traffic environments for pure electric trailer trucks, and the fact that existing ACC systems mostly rely on the current state of the preceding vehicle for car-following control, with car-following time and control parameters typically changing little with operating conditions, and insufficient utilization of mixed traffic characteristics such as adjacent vehicle cut-off risks, preceding vehicle coordination, motion stability, and traffic disturbance intensity, this invention constructs a conservative-active dual-strategy ACC control mechanism for mixed traffic environments. By identifying the primary car-following target and potential cut-off targets, the cut-off risk, communication reliability, preceding vehicle motion fluctuations, and surrounding traffic disturbance intensity are calculated, and the fusion coefficient of the conservative strategy and the active strategy, as well as the adaptive expected headway, are dynamically determined. Compared to existing ACC methods that apply a single car-following strategy to all traffic conditions, this invention employs a more aggressive car-following strategy when communication with the preceding vehicle is reliable, the vehicle is moving stably, and the risk of cutting in is low. This appropriately shortens the headway, improves traffic efficiency, and reduces unnecessary conservative deceleration. When there are non-cooperative vehicles, adjacent vehicles cutting in, the preceding vehicle decelerating suddenly, or traffic disturbances are strong, the system switches to a more conservative car-following strategy, increasing the safe headway and limiting acceleration changes. This reduces the problems of excessively close following, frequent braking, or low traffic efficiency caused by fixed strategies in complex mixed traffic environments, improving the car-following safety, ride comfort, and traffic adaptability of pure electric trailer trains.

[0022] (3) When pure electric trailer trucks undertake medium- and long-distance transportation tasks, they are simultaneously constrained by on-time arrival, remaining power, target power consumption, and refueling plans. Existing ACC systems primarily rely on the current following status for control, focusing on ensuring short-term vehicle distance safety and smooth speed, while neglecting to consider estimated arrival time, remaining mileage, and other factors. Insufficient utilization of transportation task information such as status, target power consumption, and refueling requirements can lead to inconsistencies between local car-following targets and overall transportation task objectives. This invention introduces transportation task constraints into the ACC control process. By calculating task urgency and energy demand, it dynamically adjusts safety, energy conservation, and task time-related control objectives. This allows the vehicle to prioritize energy-saving driving when time requirements are more relaxed or power pressure is higher, while appropriately increasing speed tracking requirements when there is a risk of late arrival, while meeting safety distance, battery power, and vehicle stability constraints. Compared to existing ACC methods that only pursue current distance safety or minimum speed error, this invention links local car-following control with the overall transportation task requirements, enabling the vehicle to exhibit different control tendencies under different transportation states. Therefore, the vehicle can not only maintain current car-following safety but also balance on-time arrival, energy consumption control, and refueling plans according to transportation task requirements, avoiding problems such as insufficient power, overly conservative driving, or excessive arrival time deviation caused by purely local optimal control.

[0023] (4) To address the potential conflicts between road preview energy-saving speed planning, mixed traffic car-following safety, and transportation task constraints, this invention integrates road preview reference acceleration, basic car-following acceleration, and task reference speed in the ACC expected acceleration solution, and adjusts the control objective based on the primary car-following risk, task urgency, and energy demand. Compared with existing methods where road preview control, car-following safety control, and task scheduling are independent or simply superimposed, this invention can coordinate energy saving, safety, and task requirements in a unified acceleration solution process: when the car-following risk is low, the road preview energy-saving planning results are used more, allowing vehicles to fully utilize gradient, speed limit, and traffic flow information; when the risk of cutting in, sudden deceleration of the preceding vehicle, or distance risk increases, car-following safety is prioritized to avoid weakening the safety margin for energy-saving objectives; when the expected arrival time deviation or When the state changes, the speed and acceleration outputs are adjusted in a timely manner. This improves the overall adaptability of the ACC control of the pure electric trailer truck to complex roads, mixed traffic, and changes in transportation tasks, avoiding problems such as increased energy consumption, insufficient safety, or task delays caused by a single control objective.

[0024] (5) To address the issues of insufficient coordination between the tractor and trailer energy systems, regenerative braking system, and mechanical braking system during ACC operation of pure electric trailer trains, which can easily lead to problems such as the tractor's power battery solely handling energy recovery, wasted trailer braking energy, unreasonable regenerative braking distribution, and compromised articulation stability, this invention incorporates the tractor's power battery, trailer energy storage system, and trailer electric drive axle into a unified energy management model. This is achieved by establishing a tractor-trailer dual... The recursive model, the regenerative braking capacity constraint model, and the braking force distribution model based on the articulation angle state are used to determine the braking force distribution based on the tractor and trailer under braking conditions. This invention allows for dynamic distribution of regenerative braking force, including charging power, regenerative braking torque capacity, and articulation angle and angular velocity. When the trailer has an energy storage system and an electric drive axle, it prioritizes the use of both the tractor and trailer's regenerative braking systems to recover braking energy. When the trailer lacks an electric drive axle or its energy storage system is limited, it coordinates the tractor's regenerative braking with the trailer's mechanical braking. Compared to existing methods that only consider the tractor's single-sided regenerative braking or simple mechanical braking distribution, this invention can increase the proportion of brake energy recovery in trailer-trailer trains, reduce mechanical braking energy loss, and simultaneously reduce the impact of trailer braking lag and unreasonable braking force distribution on the train's articulation stability. This results in a more suitable coordinated execution process of drive, regenerative braking, and mechanical braking for pure electric trailer-trailer trains. Attached Figure Description

[0025] Figure 1 The present invention provides a flowchart of a multi-constraint cooperative ACC control method for a pure electric trailer train. Detailed Implementation

[0026] To enable those skilled in the art to better understand the technical solutions and advantages of the present invention, the present application will be described in detail below with reference to the accompanying drawings, but this is not intended to limit the scope of protection of the present invention.

[0027] like Figure 1 As shown in the figure, this embodiment provides a multi-constraint cooperative ACC control method for pure electric trailer-trailer trains. The method includes the following steps: Step S1: Construct a unified multi-source information input for the ACC control of pure electric trailer trucks, including the vehicle status, surrounding vehicle status, road preview information, energy status of the tractor and trailer sides, and transportation task constraint information, to support subsequent preview energy-saving ACC speed planning, mixed traffic dual strategy selection, and acceleration solution under task constraints.

[0028] Step S2: Generate energy-saving ACC reference speed and reference acceleration based on road preview information and longitudinal power consumption model; specifically, based on the road slope, curvature, speed limit, traffic flow speed and longitudinal power consumption model of pure electric trailer train, plan energy-saving ACC reference speed and reference acceleration in advance in the prediction time domain.

[0029] Step S3: Construct a conservative-active dual-strategy ACC control mechanism for mixed traffic environments; identify the main following target and potential cutting targets, calculate the cutting risk, the stability of the preceding vehicle's motion, the communication reliability and the intensity of surrounding traffic disturbances, and determine the fusion coefficient of the conservative strategy and the active strategy and the adaptive expected headway based on these factors.

[0030] Step S4: Introduce transportation task constraints and solve for the expected ACC acceleration; specifically, based on the estimated arrival time, remaining mileage, The system calculates the urgency and energy demand of the task based on the status, energy replenishment requirements, and target power consumption. It dynamically and adaptively adjusts the weights of safety, energy saving, and task time, and integrates road preview reference acceleration and car-following acceleration to solve for the current expected acceleration of ACC.

[0031] Step S5: Based on the desired acceleration of ACC, coordinate the execution of drive, regenerative braking, and mechanical braking, and update online. The system calculates the target wheel end longitudinal force based on the desired acceleration, outputs the target motor torque under driving conditions, prioritizes the allocation of regenerative braking force under braking conditions and supplements it with mechanical braking, and feeds the execution results back to the next control cycle to achieve rolling optimization control.

[0032] Furthermore, in this embodiment, the multi-source information acquisition and state initialization process in step S1 is as follows: Step S11: Obtain the vehicle's operating status and basic parameters of the pure electric trailer train. At each sampling moment, the vehicle's position, speed, longitudinal acceleration, yaw angle, yaw rate, articulation angle between the tractor and semi-trailer, equivalent vehicle mass, and power battery information are obtained through integrated navigation, wheel speed sensors, vehicle controller, battery management system, and articulation angle sensor. This constitutes the vehicle's state vector:

[0033] in, This is the longitudinal position of the vehicle. For the speed of this vehicle, This is the longitudinal acceleration of the vehicle. For the horizontal swing angle, The yaw rate is angular velocity. The articulation angle between the tractor and the semi-trailer. For the current vehicle combination equivalent mass, The state of charge of the power battery. This is the transpose symbol.

[0034] In addition, this embodiment also obtains parameters related to the vehicle wheel end and driving resistance, including the wheel rolling radius. Transmission ratio Transmission efficiency Rolling resistance coefficient air drag coefficient Windward area and the maximum permissible lateral acceleration This parameter is used as the basis for subsequent longitudinal dynamics modeling, road curvature speed limit calculation, and driving / regenerative braking control quantity solution.

[0035] Step S12: Obtain surrounding vehicle status information and establish the relative motion relationship between the vehicle and surrounding vehicles. Obtain the longitudinal position, lateral position, longitudinal velocity, longitudinal acceleration, lateral velocity, and communication reliability of vehicles in front of the vehicle and in adjacent lanes through millimeter-wave radar, cameras, lidar, or V2X communication. For the... The state vectors of the surrounding vehicles are represented as follows:

[0036] in, For the first The longitudinal position of a vehicle along the road direction. Its horizontal position, For longitudinal velocity, For longitudinal acceleration, For lateral velocity, As an indicator of communication reliability or trustworthiness, and ∈[0,1]. Let the equivalent safety envelope length of the pure electric trailer train be... Then this car and the first The longitudinal distance and relative speed between the surrounding vehicles are as follows:

[0037]

[0038] when >0 and the first When a vehicle is in the current lane or is predicted to enter the current lane, it is included in the ACC candidate following target set; when the first vehicle is in the lane, it is included in the ACC candidate following target set. When a vehicle is located in an adjacent lane and shows a tendency to approach laterally, it is included in the set of potential entry targets.

[0039] Step S13: Select the aiming distance in advance and according to the spatial step size Discretize to obtain the first The position of the pre-aiming point for:

[0040] in, Pre-aiming distance for the road For the space to be far from the walking distance, This represents the number of pre-aiming points. Road slope, road curvature, speed limit, and traffic flow reference speed at each pre-aiming point are obtained through high-precision maps, navigation systems, roadside units, or cloud-based traffic information to form a road pre-aiming vector:

[0041] in, The road slope angle, For road curvature, Speed ​​limits for roads, This is the reference speed for the traffic flow ahead. The speed is calculated based on the road curvature. Safe speed at each pre-aiming point on the curve :

[0042] Where ε is a very small positive number to prevent the denominator from being zero. To allow the maximum lateral acceleration. Taking into account speed limits, curvature, and traffic flow conditions, the first... The permissible speed at each pre-aiming point for:

[0043] in, This represents the permissible speed margin relative to the traffic flow reference speed. Through the above processing, subsequent ACC speed planning can respond in advance to changes in ramps, curves, speed limits, and traffic flow speeds.

[0044] This forms a road constraint speed sequence. :

[0045] This sequence is used for subsequent ACC speed planning, enabling pure electric trailer trains to respond in advance to changes in slopes, curves, speed limits, and traffic flow speeds.

[0046] Step S14: Obtain the energy system, electric drive system, and transportation task constraint information of the tractor-trailer. The energy states of the tractor and trailer sides are obtained through the battery management system, vehicle controller, and trailer electric drive axle controller. The tractor-side energy state includes the tractor's battery state of charge, maximum allowable discharge power, maximum allowable charging power, maximum drive torque, and maximum regenerative braking torque. The trailer-side energy state includes the trailer energy storage system state of charge, maximum allowable discharge power, maximum allowable charging power, maximum drive torque of the trailer electric drive axle, and maximum regenerative braking torque.

[0047]

[0048] in, For the first State vectors of the energy system and electric drive system of the pure electric trailer-trailer at each sampling time; and These represent the state of charge of the energy storage systems on the tractor side and the trailer side, respectively. , , and These represent the maximum allowable discharge power on the tractor side, the maximum allowable charging power on the tractor side, the maximum allowable discharge power on the trailer side, and the maximum allowable charging power on the trailer side, respectively. , , and These represent the maximum drive torque on the tractor side, the maximum regenerative braking torque on the tractor side, the maximum drive torque on the trailer side, and the maximum regenerative braking torque on the trailer side, respectively. These parameters are synchronously acquired by the corresponding controllers and participate in the subsequent distribution of drive and regenerative braking.

[0049]

[0050]

[0051] The above formulas give the maximum available discharge power of current pure electric trailer truck trains. and combined maximum available charging power When the trailer is not equipped with a power battery, energy storage system, or electric drive axle, =0、 =0、 =0、 =0, at which point the control method automatically degenerates into a mode where only the tractor participates in driving and regenerative braking.

[0052] Simultaneously, current transportation task information is acquired, including remaining mileage, required arrival time, estimated arrival time, cargo type, target unit mileage energy consumption, available charging or battery swapping station locations, and refueling time constraints. Additional refueling time is determined based on the available charging or battery swapping station locations and refueling time constraints. If the current transport mission does not require refueling en route, then set it to... Conversely, the time required for charging and battery swapping is determined based on the location of available charging stations, the estimated travel time for the vehicle to reach the charging station, and the time needed for charging and battery swapping. And used for correction :

[0053] in, To account for the estimated arrival time after refueling en route, To disregard the estimated arrival time when refueling en route, The additional charging time is determined based on the set of available charging or battery swapping station locations and charging time constraints.

[0054] Then, the task time deviation is defined as:

[0055] in, For the first Task time deviation at each sampling moment The required arrival time for the task; when When, it indicates that the vehicle is at risk of being late; when When the time requirement for the current transportation task is met.

[0056] Furthermore, in this embodiment, step S2 establishes a longitudinal motion model and a power consumption model of the pure electric trailer train in the prediction time domain based on the road slope, curvature, speed limit, traffic flow speed and power battery status. It also generates energy-saving ACC reference speed and reference acceleration through rolling optimization, so that the vehicle can respond to changes in the road ahead in advance while meeting the requirements of safe following and road constraints, thereby reducing high-power drive, unnecessary rapid acceleration, rapid deceleration and mechanical braking intervention.

[0057] Specifically, the process of generating the energy-saving ACC reference speed and reference acceleration in step S2 is as follows: Step S21: Establish a phase-sensing macro-micro fusion longitudinal state prediction model based on Transformer; At the present moment Based on historical observation length Sampling period and prediction step size A rolling prediction time domain is constructed. Unlike constant acceleration extrapolation based solely on current speed and acceleration, this invention employs a Transformer-based longitudinal state prediction model. It uses historical following states, road preview information, surrounding vehicle interaction states, and the pure electric power system state as prediction inputs. Furthermore, vehicle operation phase recognition is incorporated into the prediction results, enabling the prediction model to simultaneously reflect changes in the motion of the vehicle in front, changes in road gradient and speed limits, traffic flow disturbances, and the driving, coasting, and regenerative braking characteristics of the pure electric trailer train.

[0058] First, construct the current moment. forward The sequence of micro-car-following states within each sampling time point. For the candidate car-following target vehicle At a historical moment The micro-carrier eigenvector is defined as follows:

[0059] in, and These are the vehicle's speed and acceleration, respectively. For this vehicle and the target vehicle The longitudinal spacing between them For relative velocity, and These are the target vehicle's speed and acceleration, respectively. This is the target vehicle's interaction state feature vector, used to characterize vehicle type, communication capability, engagement trend, and degree of motion fluctuation. This is a vehicle type identifier, obtained by the perception system. This is a communication capability identifier, and it can be either 0 or 1 depending on the V2X communication status. To determine the entry trend indicator, it is based on lateral speed, lateral lane approach time, and turn signal or trajectory deflection status; The motion fluctuation indicator is determined based on the standard deviation of the target vehicle's speed and acceleration within a short time window.

[0060] Simultaneously, based on the road preview information obtained in step S13, the first... The macroscopic road-traffic feature vector corresponding to each prediction step :

[0061] in, The road slope angle, For road curvature, Speed ​​limits for roads, For traffic flow reference speed, These are the traffic flow status parameters ahead.

[0062] Further, a dynamic state feature vector for the pure electric trailer train is constructed. Unlike using only a single power battery state, this invention combines the energy storage states on both the tractor and trailer sides, the electric drive axle capacity, and the combined vehicle mass as the dynamic state. enter:

[0063] Among them, the dynamic state feature vector , , , , , , , , , and These represent the tractor-side state of charge, trailer-side state of charge, charging and discharging power boundaries on both sides, driving and regenerative braking torque boundaries on both sides, and the equivalent mass of the current vehicle combination, respectively. The combined power boundary is obtained from step S14, i.e. = + , = + When the trailer does not have energy storage or an electric drive axle, the corresponding term on the trailer side is set to zero. Based on this dynamic state characteristic, the prediction model can distinguish between different structures such as single-power tractor units, trailers with energy storage, and trailers with electric drive axles.

[0064] The aforementioned micro-level car-following state sequences, macro-level road-traffic feature sequences, and pure electric power state features are input into a Transformer-based spatiotemporal prediction network. This network leverages the Transformer's ability to model long-term temporal information and multi-source feature dependencies to extract the coupling features between historical car-following states, surrounding vehicle interactions, road preview changes, and battery power constraints, and outputs future... The initial longitudinal state sequence for each prediction step:

[0065] in, For the first The initial prediction speed for each prediction step. For the first The initial prediction acceleration for each prediction step. The operating phase probability vector for pure electric vehicles:

[0066] in, , , and These respectively indicate the vehicle's position in the [missing information]. Each prediction step is in the driving, coasting, regenerative braking, or mechanical braking phase. By introducing the operating phase probability, the prediction model can not only output future speed and acceleration, but also determine whether the vehicle is more likely to be in a driving, coasting, regenerative braking, or mechanical braking state in the future, providing a basis for subsequent energy consumption calculation and regenerative braking priority control.

[0067] Furthermore, to avoid the prediction network output from exceeding the actual power boundary of the pure electric trailer train, the initial predicted acceleration is physically consistent with the road gradient and battery power constraints.

[0068] In this embodiment, the upper and lower boundaries of the predicted acceleration are determined based on the current combined available discharge power and combined available charging power of the tractor and trailer. :

[0069]

[0070] in, To allow maximum acceleration, To allow maximum deceleration, The maximum available discharge power for both the tractor and trailer. To comprehensively consider the maximum available charging power, To drive efficiency, To improve regenerative braking energy recovery efficiency, To prevent extremely small positive numbers with a denominator of zero, For the first Equivalent driving resistance per predicted step The expression is:

[0071] in, For rolling resistance coefficient, For air drag coefficient, For windward area, For gravitational acceleration, the above variables together determine the dynamic acceleration boundary of the current prediction step.

[0072] Boundary corrections are applied to the initial predicted acceleration:

[0073] Further smoothing corrections are made by incorporating acceleration rate of change constraints:

[0074] in, To allow for the maximum rate of change of acceleration, the above-mentioned physical consistency correction ensures that the prediction results simultaneously satisfy the power battery power constraint, road slope resistance constraint, and acceleration ride comfort constraint, thus avoiding the problem that the neural network prediction results cannot be actually executed by the vehicle.

[0075] Finally, the velocity and longitudinal position sequences in the predicted time domain are generated based on the corrected acceleration sequence. Let the initial state satisfy:

[0076] but:

[0077]

[0078] The final predicted sequence of longitudinal position, velocity, and acceleration in the time domain is obtained:

[0079]

[0080]

[0081] in, , and The first Predicted longitudinal position, predicted velocity, and predicted acceleration for each prediction step. , and Together, these constitute the longitudinal motion prediction prior sequence within the current prediction time domain. This prediction prior sequence is not directly used as the final road-aiming energy-saving planning result, but rather as the basis for subsequent acceleration correction and speed planning, enabling step S22 to make limited corrections to the future motion state of the vehicle based on the Transformer prediction results. This model, through the fusion of macroscopic road prediction, microscopic car-following interaction, and pure electric power state, improves the adaptability of the prediction results to changes in gradient, speed limit, traffic flow disturbances, and the energy state changes of pure electric trailer trains.

[0082] Step S22: Establish a phase-aware longitudinal resistance and tractor-trailer dual energy storage power model; To avoid relying entirely on neural network prediction results, this invention introduces an acceleration correction factor based on the prediction prior output in step S21, expressing the actual planning acceleration in road pre-aiming energy-saving planning as:

[0083] in, The acceleration correction amount to be optimized in the road pre-aiming energy-saving planning is obtained through optimization solution in step S25; This is the revised planned acceleration. To ensure that the velocity sequence obtained recursively in step S21 can participate in subsequent energy-saving planning, this step no longer starts from the current velocity. Instead of recalculating the planning speed separately, the predicted speed obtained in step S21 is used. and predicted location Based on this, the velocity correction and position correction caused by the acceleration correction are superimposed.

[0084]

[0085]

[0086]

[0087] This leads to the planned speed in road energy-saving planning. and planned location :

[0088] in, and The predicted velocity and predicted longitudinal position are obtained recursively from step S21. and This refers to the planned speed and longitudinal position after road pre-planning and energy-saving corrections. When When the value is 0, the planned speed and planned position degenerate into the predicted prior results obtained in step S21; when road gradient, road speed limit, traffic flow status, or power battery power constraints require adjustment of the vehicle's motion state, then... Limited corrections are made to the predicted prior sequence.

[0089] In the j-th prediction step (the j-th prediction step) (at each pre-aiming point), according to the planned location The road gradient, permissible road speed, and traffic flow reference speed obtained in step S13 are matched. The vehicle is subject to the combined effects of rolling resistance, air resistance, and gradient resistance, resulting in its equivalent driving resistance. Represented as:

[0090] in, Let g be the equivalent mass of the current vehicle assembly, and g be the acceleration due to gravity. The rolling resistance coefficient, air density, Where A is the air resistance coefficient and A is the frontal area. For planning location The corresponding road slope angle.

[0091] The longitudinal force at the wheel end required to achieve the planned acceleration and wheel end mechanical power They are respectively:

[0092]

[0093] Further consideration is given to the impact of trailer energy storage systems and trailer electric drive axles on energy flow. For pure electric trailer trains, the longitudinal force at the wheel ends is no longer mapped solely to the power of a single power battery, but is instead distributed as tractor-side power and trailer-side power based on the tractor and trailer-side execution capabilities.

[0094] when When the value is ≥0, the vehicle is in driving condition, and the driving force on the tractor and trailer sides is ≥0. The allocation is as follows:

[0095]

[0096] in, This is the drive force distribution coefficient for the tractor. If the trailer is not equipped with an electric drive axle, then... =1; if the trailer is equipped with an electric drive axle, then According to the tractor and trailer The torque capacity and articulation stability of the electric drive axle are dynamically determined. Discharge power on the tractor side and trailer side. They are respectively:

[0097]

[0098] in, and These refer to the drive efficiency of the electric drive systems on the tractor side and the trailer side, respectively.

[0099] when When the speed is less than 0, the vehicle is in braking condition, and the required total braking force is:

[0100] No. At each pre-aiming point, the tractor and trailer sides can utilize maximum regenerative braking force. They are respectively:

[0101]

[0102] in, and These are the maximum regenerative braking torques on the tractor side and the trailer side, respectively. and These are the transmission ratios for the tractor and trailer sides, respectively. and To correspond to the transmission efficiency. and To improve regenerative braking energy recovery efficiency, This is the wheel's rolling radius.

[0103] Define the regenerative braking distribution coefficient of the tractor And according to the tractor and trailer The difference, hinge angle, and hinge angular velocity are dynamically adjusted:

[0104] in, and These are the articulation angle and articulation angular velocity between the tractor and the trailer, respectively. , and The calibration coefficient; the first At each pre-aiming point, the tractor and trailer sides They are respectively When the trailer When the charging power is lower and a higher permissible charging power is allowed, the speed can be reduced. This allows more braking energy to be recovered to the trailer side; when the articulation angle or articulation velocity is large, it improves... Limit the proportion of regenerative braking on the trailer side to reduce the impact of trailer braking disturbances on train stability.

[0105] No. Actual regenerative braking force at each pre-aiming point on the tractor side and trailer side for:

[0106]

[0107] The remaining portion not covered by the regenerative braking of the tractor and trailer is supplemented by mechanical braking:

[0108] The corresponding regenerative braking power on the tractor side, the regenerative braking power on the trailer side, and the equivalent power of mechanical braking are as follows:

[0109] in, and The first The predicted step involves regenerative braking power recovery on both the tractor side and trailer side. For mechanical braking equivalent power, and These are the regenerative braking energy recovery efficiencies on the tractor side and the trailer side, respectively.

[0110] Predicting tractor and trailer energy storage systems in the time domain The recurrence relation is:

[0111]

[0112] in, and The first The predicted state of charge of the tractor-side and trailer-side energy storage systems is as follows: and These are the energy storage system capacities on the tractor and trailer sides, respectively. When the trailer is not equipped with an energy storage system or electric drive axle, [the following is added]: =0、 =0、 =0, at which point the above model degenerates into a single-sided tractor drive and regenerative braking model. Through the above model, road gradient changes, vehicle operating phase, dual-sided battery power constraints, and articulated stability can be uniformly mapped into energy consumption and energy recovery evaluation quantities in subsequent speed planning.

[0113] Step S23: Based on the planned speed, planned acceleration, tractor-trailer battery-side power, and operating phase probability obtained in Step S22, construct the road pre-aiming energy-saving planning objective function. This step makes limited corrections near the Transformer prediction prior given in Step S21, so that the planning results inherit historical car-following and traffic interaction prediction information, and can also make energy-saving adjustments based on road gradient, speed limit, dual-side battery power, and regenerative braking capability.

[0114] First, construct the predictive priors. :

[0115] in, and These are the predicted velocity and predicted acceleration output in step S21, respectively. and These are the planned speed and planned acceleration after road pre-aiming energy saving correction, respectively. This term is used to constrain the optimization results from deviating excessively from the longitudinal motion prediction prior given in step S21, so that the road pre-aiming energy saving planning can be modified to a limited extent based on the traffic interaction prediction results.

[0116] Constructing road speed constraints :

[0117] in, The first one obtained in step S13 Preview of the permitted speed at certain road points. This feature allows vehicles to respond in advance to changes in speed limits, curvature, and traffic flow speeds.

[0118] Constructing energy terms :

[0119] Among them, the energy term takes into account the discharge power of both the tractor and trailer sides. , Regenerative braking power recovery of tractor and trailer sides , Mechanical braking equivalent power and both sides Degree of balance; , , and They are respectively used to characterize discharge power penalty, recovered energy utilization, mechanical braking penalty, and bilateral. Equilibrium constraints and This is the offline calibration coefficient. This item is used to reduce total discharge power and mechanical braking losses, while avoiding long-term excessive consumption of energy storage systems on one side of the tractor or trailer.

[0120] further, and Determine based on the running phase probability output in step S21:

[0121]

[0122] in, and This is the base coefficient. To prevent unnecessary braking by the vehicle in order to recover energy, it is set... The upper limit of the value; to suppress premature intervention of mechanical braking, set The lower limit of the value. When the predictive model determines that the vehicle is more likely to enter the regenerative braking or mechanical braking phase, increase the value. This encourages prioritizing regenerative braking to recover energy; when the model determines that the vehicle is more likely to be in a coasting, regenerative braking, or mechanical braking phase, the [measurement / increase] is [increased / stretched]. This reduces the need for mechanical braking, allowing the deceleration process to be completed primarily through coasting and regenerative braking.

[0123] Constructing smoothness terms :

[0124] This feature is used to suppress rapid acceleration, rapid deceleration, and sudden acceleration changes, improving the smoothness of speed planning for pure electric trailer trains. Specifically, for ,Pick .

[0125] To avoid the impact of differences in the dimensions and orders of magnitude of different objective items on the optimization results, the corresponding benchmark values ​​are first used to compare the results. , , and Dimensionless transformation yields , , and The benchmark value can be determined from vehicle design parameters, regulatory limits, or historical operating data statistics.

[0126] Finally, the objective function for road pre-aiming energy-saving planning in the prediction time domain is:

[0127] in, , , and These are the weights for the prediction prior, road speed constraint, energy, and ride comfort. These weights are based on factors such as flat road, uphill, downhill, speed limit changes, traffic congestion, and low... and high Typical operating condition data are used for offline calibration using a multi-objective genetic algorithm, and the optimal weight combination under different operating conditions is stored in the vehicle controller; when the vehicle is running, the system calculates the weight based on the current road slope, traffic flow status, etc. The status and the corresponding risk of following the lead time are assigned to different operating conditions, and the appropriate weight combination is applied.

[0128] Using the objective function described above, a comprehensive evaluation of the vehicle motion state and energy state corresponding to different acceleration correction sequences can be performed, building upon the Transformer's prediction of traffic interaction trends. It should be noted that the planned acceleration, planned speed, tractor-side and trailer-side power, and the recursive results of the dual-side SOC in step S22 are not fixed parameters, but rather vary with the acceleration correction amount. It changes with the changes. For any given acceleration correction sequence The corresponding planned speed sequence, planned acceleration sequence, tractor-trailer power distribution results, and SOC change results can all be calculated in step S22, and then substituted into the prediction prior terms, road speed constraint terms, energy terms, and ride comfort terms in step S23. Subsequent step S25 obtains the optimal acceleration correction sequence by minimizing this objective function. By combining the forward slope, speed limit, curvature, battery status, and regenerative braking capability, the predicted acceleration obtained in step S21 is modified to a limited extent, and the road advance reference acceleration and reference speed are further obtained, so as to reduce high-power drive, rapid deceleration and mechanical braking intervention, and improve the energy saving and smoothness of the road advance type ACC control of pure electric trailer train.

[0129] Step S24: To ensure that the road prediction energy-saving planning results obtained in the subsequent step S25 based on the objective function in step S23 can be actually executed by the pure electric trailer train, it is necessary to set constraints on the planning speed, planning acceleration, acceleration change rate, battery power of the tractor side and trailer side, SOC of the tractor side and trailer side, and acceleration correction amount in the prediction time domain.

[0130] First, the planned speed should meet the road's permissible speed constraints: ; in, Taking into account road speed limits, curvature safety speeds, and traffic flow reference speeds, To correct a given acceleration sequence At that time, the velocity recursion relationship obtained in step S22 is... Each prediction step has a candidate planning speed, and the corresponding optimal road aiming reference speed is obtained after minimizing the objective function in step S25. .

[0131] The planned acceleration should satisfy the combined constraints of vehicle longitudinal dynamics, power battery power, and comfort. Step S21 has already constructed a dynamic acceleration boundary based on road gradient, vehicle speed, and power battery charging / discharging power. This step no longer uses a single global constant boundary, but instead uses the dynamic boundary calculation method from step S21 to constrain the planned acceleration. The difference is that the dynamic boundary in step S21 was used to correct the initial predicted acceleration output by the Transformer, while the dynamic boundary in this step is used to constrain the planned acceleration obtained from the road's pre-aiming energy-saving planning.

[0132] Specifically, in the first One prediction step, based on the planned location obtained in step S22. Match the corresponding road gradient and use the planned speed Substituting the predicted velocity in step S21 into the dynamic acceleration boundary calculation formula in step S21, we obtain the upper and lower boundaries of dynamic acceleration in the planning stage: , .

[0133] in, and These represent the combined effects of the current planned speed, road gradient, maximum available discharge power of the tractor-trailer, maximum available charging power, and vehicle running resistance, respectively, on the [number]th [day / time]. The maximum allowable acceleration and deceleration boundaries for each prediction step. Therefore, the planned acceleration should satisfy:

[0134] in, To provide a given acceleration correction amount At that time, the predicted prior acceleration obtained in step S21 is corrected in step S22 to form the first... The candidate planned accelerations for each prediction step are determined by the above constraints, which prevent the optimized planned accelerations from exceeding the actual dynamic boundary of the vehicle under the current gradient, speed, and combined power conditions of the tractor and trailer.

[0135] Meanwhile, to improve the smoothness of velocity planning, the rate of change of planned acceleration should satisfy:

[0136] in, This is the upper limit of the rate of change of acceleration, used to limit abrupt changes in acceleration between adjacent prediction steps.

[0137] The battery power of both the tractor and trailer sides should meet their respective charge and discharge capacity constraints:

[0138]

[0139] The above constraints correspond to the discharge power on the tractor side, respectively. Trailer-side discharge power tractor side recovery charging power and trailer-side recovery charging power The upper and lower limits. When the trailer is not equipped with an energy storage system or an electric drive axle, the trailer-side power constraint is not included in the solution.

[0140] Tractor and trailer side Minimum power requirements must be met:

[0141] Among them, the tractor side and trailer side All must not be lower than the corresponding minimum permissible state of charge. and When the trailer is not equipped with an energy storage system, only the tractor side is retained. constraint.

[0142] Considering that the prediction results output in step S21 already include information on the motion of the vehicle ahead, road traffic conditions, and vehicle dynamics, this step limits the acceleration correction to prevent the road preview energy-saving optimization from deviating excessively from the prediction model results.

[0143] in, This represents the maximum correction amount for single-step predicted acceleration.

[0144] Furthermore, to ensure priority is given to coasting and regenerative braking in downhill or speed-reduced scenarios, the equivalent power of mechanical braking should meet the following requirements:

[0145] in, For the first The upper limit of mechanical braking intervention allowed in each prediction step. This upper limit can be dynamically adjusted based on the operating phase probability output in step S21. When the vehicle is in the... The probability that a prediction step is in the coasting phase or the regenerative braking phase is relatively high, i.e. or When it is large, reduce This allows the vehicle to prioritize coasting or regenerative braking to adjust its speed; when the safe distance is insufficient, the speed limit ahead is significantly reduced, or road constraints require the vehicle to decelerate quickly, the speed limit is increased. Mechanical braking is allowed to supplement braking force.

[0146] When there is a significant downhill slope or a reduced speed limit ahead at the pre-aiming point, the vehicle gradually reduces its drive output before entering the area through the aforementioned speed constraints, charging power constraints, and mechanical braking constraints, and decelerates as much as possible through coasting and regenerative braking. When there is an uphill slope at the pre-aiming point, the vehicle reasonably maintains its kinetic energy while not exceeding the road's permissible speed and safe following distance through speed constraints and discharge power constraints, avoiding short-term high-power drive demand after entering the slope.

[0147] Step S25: Solve for the road preview reference velocity and reference acceleration; use the acceleration correction sequence in the prediction time domain as the optimization variable:

[0148] Solve the following optimization problem:

[0149] And satisfy the velocity, acceleration, rate of change of acceleration, battery power, etc. in step S24. Predictive correction amount and mechanical braking intervention constraints.

[0150] The optimal acceleration correction sequence is obtained after solving:

[0151] The road aiming reference acceleration, i.e., the energy-saving ACC reference acceleration sequence, is as follows:

[0152] The velocity correction is obtained recursively from the optimal acceleration correction sequence. Let:

[0153] Then we have:

[0154] The road aiming reference speed, i.e. the energy-saving ACC reference speed sequence, is further obtained as follows:

[0155] in, The predicted velocity obtained by recursion in step S21, This is the velocity correction amount corresponding to the optimal acceleration correction amount. Therefore, the predicted velocity sequence obtained in step S21... The reference speed output in step S25, which is used as the base sequence for generating the road aiming reference speed, is not derived from the current speed alone, but is modified based on the predicted prior speed.

[0156] The rolling optimization method uses only the result of the first optimization step as the road aiming reference input for the current control cycle, i.e.:

[0157]

[0158] in, and These are the road aiming reference acceleration and road aiming reference speed output in the current control cycle of step S2, respectively, which are used to fuse with the basic car-following acceleration and mission reference speed in the subsequent step S4; To predict the optimal road aiming reference acceleration for the first prediction step in the time domain, The reference speed for the next predicted time step is obtained by recursion from the optimal acceleration correction sequence.

[0159] When the next sampling time arrives, the road preview information, surrounding traffic conditions, battery status, and transportation task status are reacquired, and steps S21 to S25 are re-executed.

[0160] Furthermore, in this embodiment, step S3 includes the following steps: Step S31: Identify the primary following target and potential cutting-in targets. For vehicles in the current lane and in front of the current vehicle, select the vehicle with the closest longitudinal distance as the primary following target, denoted as... The method for determining it is as follows:

[0161] in, This is the set of candidate vehicles ahead of this vehicle in its current lane.

[0162] For vehicles in adjacent lanes, determine whether there is a tendency for them to cut into the current vehicle's lane based on their lateral distance, lateral speed, and lane center position. Let the lateral position of the current vehicle's lane centerline be... Then the first Cross-lane approach time of adjacent vehicles for:

[0163] When the lateral velocity direction of an adjacent vehicle is pointing towards the lane of this vehicle, and If the value is less than a set threshold, the vehicle will be included in the potential entry target set. .

[0164] Step S32: For the first in the potential entry target set... For each vehicle, construct a risk feature vector for entry. :

[0165] in, This is for turn signal, lane line crossing, or trajectory deviation indicators. The offline-calibrated logic function is used to calculate the first... Risk of entering the market for individual vehicles :

[0166] in, The intercept parameter of the risk logic function is used to input the risk. The cutting-in risk feature weight vector describes the impact of longitudinal distance, relative speed, lateral distance, lateral speed, lateral approach time, and turning / trajectory deflection markers on cutting-in risk. It is obtained through offline calibration using historical mixed traffic samples. The maximum cutting-in risk among all potential cutting-in targets is:

[0167] If there are no potential entry points at present, then let .

[0168] Furthermore, the intensity of mixed traffic disturbance is calculated based on the number of vehicles in the surrounding area, average relative speed, and acceleration fluctuations within the sensing range. :

[0169] in, To determine the number of vehicles in the surrounding area within the sensing range. This is a normalized parameter for the number of vehicles. and These are the normalized parameters for velocity and acceleration, respectively. , and These are the weighting coefficients for the vehicle quantity disturbance term, relative speed disturbance term, and acceleration disturbance term, respectively. They are used to characterize the influence of the surrounding vehicle density, speed difference, and acceleration fluctuation on the intensity of mixed traffic disturbance. They are calibrated offline based on typical mixed traffic conditions and stored in the vehicle controller.

[0170] Step S33: Calculate the coordination and motion stability of the primary car-following target. A cooperative stability evaluation metric is constructed based on its communication reliability, speed fluctuation, and acceleration fluctuation. :

[0171] in, To improve the reliability of communication between the primary and secondary targets. The standard deviation of the target velocity within a short time window. The standard deviation of the principal target acceleration within a short time window. and These are the normalized parameters. , and The weighting coefficients for the communication reliability term, speed stability term, and acceleration stability term are respectively obtained through offline calibration using typical vehicle interaction data such as reliable communication, weak communication, non-cooperative driving, and manual driving, and satisfy the following:

[0172] The larger the value, the more stable the vehicle in front is and the more suitable it is to adopt an aggressive following strategy; The smaller the value, the stronger the fluctuation in the movement of the vehicle in front or the poorer the coordination, and the more conservative the following should be.

[0173] Step S34: Determine the conservative-aggressive strategy fusion coefficient. The conservative-aggressive dual-strategy fusion coefficient is determined based on the main following target coordination stability, potential entry risks, and the intensity of mixed traffic disturbances. :

[0174] in, This indicates that the variable is restricted to... Within the range; This is the basic bias term for the strategy fusion coefficient. , and The weighting coefficients are respectively the cooperative stability of the main car-following target, the potential cutting-in risk, and the intensity of mixed traffic disturbance. They are calibrated offline based on simulation data and real vehicle data collected under typical operating conditions such as cutting-in vehicles, free-flow car-following, congested car-following, and non-cooperative leading vehicles, and stored in the vehicle controller.

[0175] Step S35: Determine the desired headway and following safety distance. Let the desired headway under the aggressive strategy be... The expected time interval under the conservative strategy is And satisfy ACC expected time interval at the current moment. for:

[0176] in, and It can be determined offline based on the vehicle's design speed range, braking performance, regulatory safety requirements, and typical car-following data. When surrounding vehicles are stable, communication with the vehicle ahead is reliable, and the risk of cutting in is low... To improve traffic efficiency, the system uses a smaller expected time interval; however, this may be necessary when there are non-cooperative vehicles, cutting-in risks, or significant traffic disturbances. To reduce this, the system uses a larger expected time interval to improve security.

[0177] Based on this, construct the safe distance corresponding to the primary following target:

[0178] in, For minimum stationary safe distance, The main focus is on the target speed. Assuming the vehicle in front has the maximum deceleration, The risk distance correction factor is obtained from the offline calibration of the vehicle test conditions. This is the effective braking deceleration of the vehicle.

[0179] Furthermore, in this embodiment, the specific process of solving the transportation task constraints and the desired ACC acceleration in step S4 is as follows: Step S41: Based on the remaining driving range and the required arrival time of the task Calculate the average speed required to meet the transportation task at the current moment. :

[0180] in, For the current time, To prevent extremely small positive numbers with a denominator of zero, the task reference speed is obtained by considering the current road speed limit. :

[0181] Calculate the mission urgency based on the expected arrival time deviation. :

[0182] in, This is the time deviation normalization parameter. The larger the value, the higher the risk of being late; The smaller the value, the easier it is to meet the task time requirements.

[0183] Step S42: Calculation Condition difference and energy demand level; based on current road forecasting information and historical power consumption per unit mileage, estimate the power required to complete the remaining route. calculate:

[0184] in, This represents the number of discrete points on the remaining route. For the first Predicted power consumption per unit mileage for the remaining road segments. This refers to the length of the road segment. The distance required to complete the remaining transportation task. for:

[0185] in, For power battery capacity, To complete the remaining transportation tasks ; definition State difference for:

[0186] At that time, it indicated that the current power level was insufficient to reliably complete the remaining tasks, requiring increased energy-saving control or triggering a power replenishment plan correction. According to Energy demand level is calculated using state difference. :

[0187] in, for State difference normalization parameter.

[0188] Step S43: Adjust the ACC control weights according to the transportation task status, and define the primary following risk as:

[0189] in, The conservative-active strategy fusion coefficients obtained in step S34. Because... A larger value indicates that the system is more inclined to actively follow the market, and the current risk of following the market is lower. The larger the value, the higher the risk of following the lead.

[0190] Based on task urgency Energy demand level and the risk of following the main trend Correcting the safety, energy efficiency, and task time weights in the ACC acceleration solution:

[0191]

[0192]

[0193] in, , and These are respectively: safety weight, task time weight, and energy-saving weight; , and For the corresponding basic weights; , and These are adjustment coefficients. The above basic weights and adjustment coefficients are determined by typical transportation tasks and different... The offline simulation and real vehicle data calibration under different conditions, road gradients, and car-following risks are obtained and stored in the vehicle controller. When the vehicle is running, the parameter combination under the corresponding condition category is called according to the urgency of the task, the energy demand, and the main car-following risk.

[0194] Step S44: Generate the base following acceleration based on the distance and speed errors between the main following target and the vehicle. Distance error and speed error They are respectively:

[0195]

[0196] in, The longitudinal distance between the primary target vehicle and the vehicle itself. Basic car-following acceleration. for:

[0197] in, , and To control the gain in the case of car-following, Accelerate towards the primary target. For communication reliability. The above control gain was obtained through simulation calibration of typical car-following, rapid deceleration, cutting in, and congestion conditions.

[0198] Based on the risk of main following Determine the fusion coefficient between road anti-aiming acceleration (energy-saving ACC reference acceleration) and car-following acceleration. :

[0199] When the risk of main-carrying is low When the speed is relatively high, the system places greater emphasis on the road-aiming energy-saving reference acceleration obtained in step S2; when the main following risk is high, The system is smaller and places greater emphasis on the safety control of the primary target.

[0200] Initial ACC acceleration after fusion for:

[0201] in, The road aiming reference acceleration output in step S2, This is a correction factor for task speed. It was obtained from offline calibration of the transportation task conditions.

[0202] Step S45: Calculate the desired acceleration of the ACC at the current moment. To optimize the variables, construct the ACC acceleration objective function for the current control cycle:

[0203]

[0204]

[0205]

[0206]

[0207] The first item is used to ensure safe following, the second item is used to meet the speed requirements of the transportation task, the third item is used to inherit the energy-saving results of road pre-aiming, the fourth item is used to track and fuse the initial acceleration, and the fifth item is used to limit acceleration abrupt changes. and To integrate acceleration tracking and ride comfort weights, their values ​​are obtained through offline calibration of typical ACC comfort and execution response conditions.

[0208] The desired acceleration for ACC should satisfy the following constraints:

[0209]

[0210]

[0211] The predicted distance is:

[0212] The desired acceleration of the ACC at the current moment is obtained by solving the following single-step constrained optimization problem:

[0213] And it satisfies all the above security constraints and execution constraints. The result is... It also reflects the results of road pre-planning for energy conservation, mixed traffic safety requirements, and transportation task constraints.

[0214] Furthermore, in this embodiment, the coordinated execution process of drive, regenerative braking, and mechanical braking in step S5 is as follows: Step S51: Based on the vehicle's longitudinal dynamics, determine the current required target wheel-end longitudinal force. for:

[0215] in, This represents the road slope angle at the current location. When... At that time, the vehicle needs to be driven; when At that time, the vehicle needs to brake.

[0216] Step S52: Based on the target wheel end longitudinal force obtained in step S51 Determine whether the vehicle is currently in driving or braking mode. When ≥0, it means the vehicle requires electric motor drive; when When the value is less than 0, it indicates that the vehicle needs to decelerate and brake.

[0217] When the vehicle requires electric motor drive, the target driving force is allocated according to the electric drive system capabilities of the tractor and trailer:

[0218] in, This is the driving force distribution coefficient for the tractor unit at the current moment. If the trailer is not equipped with an electric drive axle, then... =1, =0; if the trailer is equipped with an electric drive axle, then According to the tractor and trailer The motor efficiency and electric drive axle torque capacity are determined. The target drive torque for the tractor and trailer sides is:

[0219] After considering the constraints of maximum driving torque and maximum discharge power on the tractor and trailer sides, the actual driving torque is obtained respectively. and At this time, the vehicle is in driving mode, and both regenerative braking force and mechanical braking force are zero.

[0220] When the vehicle needs to decelerate and brake, the total braking force required is:

[0221] Under braking conditions, the regenerative braking systems on the tractor and trailer sides are prioritized to meet braking requirements. The maximum available regenerative braking forces on the tractor and trailer sides are as follows:

[0222]

[0223] Define the regenerative braking distribution coefficient of the tractor :

[0224] in, and These are the current articulation angle and articulation angular velocity, respectively. When the trailer... When the charging power is low and the trailer allows for a higher charging power, the charging power can be reduced. Increase the trailer-side recovery ratio; when the articulation angle or articulation velocity is large, increase... Limit the proportion of regenerative braking on the trailer side to avoid amplifying the risk of train yaw or folding due to braking disturbances on the trailer side.

[0225] The actual regenerative braking forces on the tractor and trailer sides are as follows:

[0226]

[0227] The corresponding regenerative braking torque is:

[0228] When the regenerative braking capacity of the tractor and trailer is insufficient to meet the total braking force requirement, the remaining portion is supplemented by mechanical braking:

[0229] If the trailer is not equipped with an electric drive axle or the energy storage system cannot receive recovered energy, then... =0, the system coordinates the regenerative braking of the tractor and the mechanical braking of the trailer to avoid the rapid increase of the articulation angle due to the lag in the braking response of the trailer or the excessive braking on one side of the tractor.

[0230] The total mechanical braking torque is:

[0231] If the mechanical braking torque is distributed according to each brake shaft, then the mechanical braking torque of the l-th brake shaft is:

[0232] in, For the first The braking force distribution coefficient of each brake axle. Further, based on the brake pressure-braking torque relationship, the mechanical braking torque is converted into braking pressure:

[0233] in, For the first The target braking pressure of each brake shaft For the first The conversion coefficient from braking pressure to braking torque on each brake axle. The target braking pressure should satisfy 0 ≤ ≤ Ultimately, , , , and The signals are sent to the tractor electric drive system, trailer electric drive axle, regenerative braking control unit, and mechanical braking actuator respectively, to realize the underlying execution of the desired acceleration of the pure electric trailer-trailer.

[0234] The present invention also provides an electronic device, comprising: one or more processors and a memory; wherein the memory is used to store one or more programs, and when the one or more programs are executed by the one or more processors, the one or more processors implement the above-described multi-constraint cooperative ACC control method for a pure electric trailer train.

[0235] The present invention also provides a computer-readable medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the above-described multi-constraint cooperative ACC control method for a pure electric trailer train.

[0236] Those skilled in the art will understand that all or part of the functions of the various methods / modules in the above embodiments can be implemented by hardware or by computer programs. When all or part of the functions in the above embodiments are implemented by computer programs, the program can be stored in a computer-readable storage medium, which may include: read-only memory, random access memory, disk, optical disk, hard disk, etc., and the program is executed by a computer to achieve the above functions. For example, the program can be stored in the memory of a device, and when the program in the memory is executed by the processor, all or part of the above functions can be achieved.

[0237] In addition, when all or part of the functions in the above embodiments are implemented by computer programs, the programs can also be stored in storage media such as servers, other computers, disks, optical discs, flash drives, or portable hard drives. They can be downloaded or copied to the memory of the local device, or the system of the local device can be updated. When the program in the memory is executed by the processor, all or part of the functions in the above embodiments can be implemented.

[0238] The above describes specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A multi-constraint cooperative ACC control method for pure electric trailer-trailer trains, characterized in that, Includes the following steps: Step S1: Construct a unified multi-source information input for the ACC control of pure electric trailer truck trains, including the status of the vehicle and surrounding vehicles, road preview information, energy status of the tractor and trailer, and transportation task constraint information; Step S2: Based on the data obtained in Step S1, traffic flow status, and vehicle combination mass, predict the future longitudinal state. Then, introduce acceleration correction to establish a phase-aware longitudinal resistance and tractor-trailer dual energy storage power model. At the same time, construct the road pre-aiming energy-saving planning objective function and plan the energy-saving ACC reference speed and reference acceleration in advance in the prediction time domain through rolling optimization. Step S3: Construct a conservative-active dual-strategy ACC control mechanism for mixed traffic environments; Identify the primary car-following target and potential cutting-in targets, calculate the cutting-in risk, the cooperative stability of the primary car-following target, and the intensity of mixed traffic disturbances, and determine the fusion coefficient of the conservative-proactive dual strategy accordingly, thereby determining the expected ACC distance at the current moment and calculating the safe distance corresponding to the primary car-following target. Step S4: Generate the basic car-following acceleration and fuse it with the energy-saving ACC reference acceleration; ACC expected acceleration at the current moment To optimize the variables, an ACC acceleration objective function is constructed, consisting of a term ensuring car-following safety, a term meeting the speed requirements of the transportation task, a term inheriting the energy-saving results of road pre-aiming, a term for tracking and fusing the initial acceleration, and a term limiting acceleration mutations. The expected acceleration of the ACC at the current moment is obtained by solving a single-step constraint optimization problem. Step S5: Drive, regenerative braking and mechanical braking are executed in coordination according to the ACC desired acceleration.

2. The multi-constraint cooperative ACC control method for a pure electric trailer train according to claim 1, characterized in that, In step S1, the vehicle status includes the vehicle's longitudinal position, speed, longitudinal acceleration, yaw angle, yaw rate, articulation angle between the tractor and semi-trailer, equivalent mass of the vehicle, and state of charge of the power battery along the road centerline; the surrounding vehicle status includes the longitudinal position, lateral position, longitudinal speed, longitudinal acceleration, lateral speed, and communication reliability of the vehicles in front of the vehicle and vehicles in adjacent lanes; the road preview information includes the location of preview points and the road slope, road curvature, road speed limit, traffic flow reference speed, curve safety speed, and road permissible speed at each preview point; the tractor-side energy status includes the tractor's power battery state of charge, maximum permissible discharge power, maximum permissible charging power, maximum drive torque, and maximum regenerative braking torque; the trailer-side energy status includes the trailer's energy storage system state of charge, maximum permissible discharge power, maximum permissible charging power, maximum drive torque of the trailer's electric drive axle, and maximum regenerative braking torque; the transportation task constraint information includes the remaining driving mileage, the task-required arrival time, the estimated arrival time, the type of cargo, the target unit mileage energy consumption, the set of available charging stations or battery swapping stations, and the refueling time.

3. The multi-constraint cooperative ACC control method for a pure electric trailer train according to claim 2, characterized in that, In step S2, a phase-aware macro-micro fusion longitudinal state prediction model based on Transformer is established, which predicts the current state at the current time. forward The micro-tracking state sequence within each sampling time point, constructed based on road preview information, is the first... The macroscopic road-traffic feature vector corresponding to each prediction step, along with the energy states of the tractor and trailer sides, the electric drive axle capacity, and the vehicle combination mass, are used as prediction inputs to output the future... The initial longitudinal state sequence for the prediction step includes the initial predicted velocity, initial predicted acceleration, and the pure electric vehicle's operating phase probability. Based on the pure electric vehicle's operating phase probability, the vehicle's position in the prediction step is determined in advance. The prediction step is in the driving, coasting, regenerative braking or mechanical braking condition; then, the upper and lower boundaries of the prediction acceleration are determined according to the current comprehensive available discharge power and comprehensive available charging power of the tractor and trailer, the initial prediction acceleration is corrected by boundary correction, and smoothing correction is performed in combination with the acceleration change rate constraint to obtain the corrected acceleration, i.e., the prediction acceleration; then, the velocity and longitudinal position sequence in the prediction time domain are generated according to the corrected acceleration sequence.

4. The multi-constraint cooperative ACC control method for a pure electric trailer train according to claim 3, characterized in that, The road pre-aiming energy-saving planning objective function mentioned in step S2 includes prediction prior terms. Road speed constraints Energy Item Smoothness item The optimal acceleration correction sequence is obtained by minimizing the objective function. Furthermore, when solving the problem, constraints are set for the planned speed, planned acceleration, rate of change of acceleration, battery power on the tractor side and trailer side, SOC on the tractor side and trailer side, and acceleration correction amount in the prediction time domain. In this way, combined with the slope ahead, speed limit, curvature, battery status and regenerative braking capability, the predicted acceleration is modified to a limited extent to obtain the road pre-aiming reference acceleration, i.e. the energy-saving ACC reference acceleration. Then, the speed correction amount is obtained by recursively deriving the optimal acceleration correction sequence, thereby obtaining the road aiming reference speed, i.e. the energy-saving ACC reference speed; Predicting priors ; in, and These are the predicted velocity and predicted acceleration, respectively. and These are the planned speed and planned acceleration after road energy-saving corrections. ; Acceleration for optimization in road energy conservation planning; This refers to the acceleration correction amount to be optimized in the road energy-saving planning. , It is the velocity correction amount, calculated based on the acceleration correction amount. To predict the step size; Road speed constraint ;in, For the first The permissible speed at each pre-aiming point; Energy Item ; Among them, the The discharge power of the tractor and trailer sides at each pre-aiming point are respectively , The regenerative braking power of the tractor and trailer sides is respectively , The equivalent power of mechanical braking is tractor and trailer sides They are respectively All of these are calculated based on the phase-sensing longitudinal resistance and the dual energy storage power model of the tractor-trailer; , , and They are respectively used to characterize discharge power penalty, recovered energy utilization, mechanical braking penalty, and bilateral. Equilibrium constraints and Determined based on the operating phase probability of pure electric vehicles and Offline calibration; Smoothness term ;in, The predicted acceleration of the previous prediction step. The sampling period.

5. The multi-constraint cooperative ACC control method for a pure electric trailer train according to claim 4, characterized in that, In step S3, based on the positional relationship between the vehicle and surrounding vehicles, each surrounding vehicle is divided into candidate following targets or potential cutting targets. For vehicles in the current lane and in front of the vehicle, the vehicle with the closest longitudinal distance is selected as the primary following target. For the potential cutting target set, the first... For each vehicle, construct a risk feature vector for entry. The offline calibrated logic function is used to calculate the first... Risk of entering the market for individual vehicles : ； in, The intercept parameter of the risk logic function is used to input the risk. This is the entry risk feature weight vector, used to describe the degree of influence of longitudinal distance, relative speed, lateral distance, lateral speed, lateral approach time, and steering / trajectory deflection indicators on entry risk. It is the transpose symbol; The intensity of mixed traffic disturbance is calculated based on the number of surrounding vehicles, average relative speed, and acceleration fluctuations within the sensing range. And identify the maximum entry risk among all potential entry targets. ; A cooperative stability evaluation metric is constructed based on the communication reliability, velocity fluctuations, and acceleration fluctuations of the primary and secondary targets. ; Current ACC expected time interval for: ； in, The conservative-aggressive dual-strategy fusion coefficient is given by the expected time interval under the aggressive strategy. The expected time interval under the conservative strategy is ; The safe distance corresponding to the main following target for: ； in, For minimum stationary safe distance, This represents the current speed of the vehicle. The main focus is on the target speed. Assuming the vehicle in front has the maximum deceleration, This is a correction factor for the distance to the entry risk. This is the effective braking deceleration of the vehicle.

6. The multi-constraint cooperative ACC control method for a pure electric trailer train according to claim 5, characterized in that, In step S4, the distance error between the main target and the vehicle is used as a basis. and speed error Generate basic car-following acceleration ; ； in, , and To control the gain in the case of car-following, Accelerate towards the primary target. For communication reliability; Then, based on the risk of main follow-up. Determine the fusion coefficient between the energy-saving ACC reference acceleration and the base car-following acceleration. ; Initial ACC acceleration after fusion for: ； in, The road aiming reference acceleration output in step S2 is the energy-saving ACC reference acceleration. Adjustment factor for task speed; For task reference speed.

7. The multi-constraint cooperative ACC control method for a pure electric trailer train according to claim 6, characterized in that, The safe car-following term in the ACC acceleration objective function in step S4 ; Meet the speed requirements for transportation tasks Inheriting the results of road pre-planning energy conservation ; Tracking the initial acceleration term Limiting acceleration mutation terms in, The longitudinal distance between the main target vehicle and the vehicle itself. The acceleration of this vehicle at the current moment. , and These are respectively: safety weight, task time weight, and energy-saving weight; and To integrate acceleration tracking and smoothness weights.

8. The multi-constraint cooperative ACC control method for a pure electric trailer train according to claim 7, characterized in that, In step S5, the required target wheel-end longitudinal force is first calculated based on the vehicle's longitudinal dynamics and the desired ACC acceleration. Then determine whether the vehicle is currently in a driving or braking condition; When the vehicle requires motor drive, the target driving force is allocated according to the electric drive system capacity of the tractor and trailer to obtain the target driving torque on the tractor and trailer sides. At the same time, the actual driving torque is obtained by combining the maximum driving torque and maximum discharge power on the tractor and trailer sides. When the vehicle needs to decelerate and brake, the actual regenerative braking force and corresponding regenerative braking torque of the tractor and trailer are calculated by combining the current maximum available regenerative braking force on the tractor and trailer sides, the required total braking force, and the regenerative braking distribution coefficient of the tractor. When the regenerative braking capacity of the tractor and trailer is insufficient to meet the total braking force requirement, the remaining part is supplemented by mechanical braking.

9. A multi-constraint cooperative ACC control method for a pure electric trailer train according to claim 8, characterized in that, The regenerative braking power on the tractor side, the regenerative braking power on the trailer side, and the equivalent power of mechanical braking are respectively: ； ； ； in, and The regenerative braking energy recovery efficiency is respectively for the tractor side and the trailer side; The actual regenerative braking forces at the pre-aiming points on the tractor side and trailer side are respectively Mechanical braking force is The regenerative braking distribution coefficient of the tractor is It is based on the tractor and trailer The difference, hinge angle, and hinge angular velocity are dynamically adjusted; At each pre-aiming point, the maximum available regenerative braking force on the tractor and trailer sides is respectively... The calculations were based on the maximum regenerative braking torque on the tractor side and trailer side, the maximum allowable charging power on the tractor side and trailer side, the transmission ratio on the tractor side and trailer side, the transmission efficiency, the regenerative braking energy recovery efficiency, the wheel rolling radius, and the planned speed after road pre-aiming energy saving correction; The total braking force required at each pre-aiming point is ; Discharge power on the tractor side and trailer side ; in, These refer to the driving forces on the tractor side and trailer side when the vehicle is in driving mode. and These refer to the drive efficiency of the electric drive systems on the tractor side and the trailer side, respectively.

10. A multi-constraint cooperative ACC control method for a pure electric trailer train according to claim 9, characterized in that, In step S4, based on the remaining mileage and the required arrival time, the average speed needed to meet the transportation task at the current moment is calculated, and the task reference speed is obtained by combining this with the current road speed limit. ; Calculate the urgency of the task based on the expected arrival time deviation. Then, based on the current road forecasting information and historical power consumption per unit mileage, the remaining power required for the route is estimated and calculated, thereby determining the amount of electricity needed to complete the remaining transportation task. At the same time, determine whether the current battery status is sufficient to complete the remaining tasks, and based on... Energy demand level is calculated using state difference. Then, the risk of main follow-the-leader is determined based on the fusion coefficient of the conservative and aggressive dual strategies. And based on the urgency of the task Energy demand level and the risk of following the main trend The safety weight, energy-saving weight, and task time weight in the ACC acceleration solution are corrected.

Images