Method and system for cooperative adaptive obstacle avoidance control of main trailer vehicles on icy and snowy roads
By using a method of joint estimation and optimization of braking force distribution, the motion coordination problem of autonomous semi-trailer trains when avoiding obstacles on icy and snowy roads was solved, reducing the risk of semi-trailer sideslip and folding, and improving the feasibility and stability of obstacle avoidance control.
Patent Information
- Application Number
- CN202611429896.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-15
- Publication Date
- 2026-11-17
AI Technical Summary
When an autonomous semi-trailer truck performs emergency obstacle avoidance on icy or snowy roads with low adhesion, inaccurate estimation of road adhesion capacity, uncoordinated motion response between the tractor and the semi-trailer, and the inability of the fixed articulation angle constraint to reflect the actual stability boundary of the vehicle can cause control commands to exceed the available adhesion capacity of the tires, resulting in the semi-trailer skidding, swaying, or folding.
By jointly estimating the motion state of the tractor and semi-trailer and the road adhesion coefficient, a conservative adhesion coefficient is determined, the main-trailer plane coupled dynamic model is updated, the braking force distribution is optimized, and closed-loop correction is performed to ensure that the control commands are within the range of the tire's available adhesion capacity.
It reduces the impact of insufficient road surface adhesion recognition and tire model errors on the control results, reduces the risk of semi-trailer sideslip, shimmy and main trailer folding, and improves the feasibility and stability of obstacle avoidance control.
Smart Images

Figure CN123104638A_ABST
Abstract
Description
Technical Field
[0001] This application belongs to the field of intelligent commercial vehicle chassis control and autonomous driving technology, and in particular relates to a cooperative adaptive obstacle avoidance control method and system for main and trailer vehicles on icy and snowy roads. Background Technology
[0002] Semi-trailer truck trains, consisting of a tractor and a semi-trailer, are characterized by their large total mass, long length, and significant motion coupling between the tractor and semi-trailer. The tractor and semi-trailer are connected by an articulation device. During emergency turns, lane changes, or braking, the lateral and yaw movements of the semi-trailer act on the tractor through the articulation point. When the movements of the tractor and semi-trailer are not coordinated, the semi-trailer is prone to sideslip, swaying, or even folding.
[0003] On icy and snowy surfaces with low grip, the longitudinal and lateral forces that tires can provide are significantly reduced, and the road grip condition changes with snow accumulation, ice formation, snow melting, and the position of the wheels. When an autonomous semi-trailer truck performs an emergency lane change to avoid an obstacle, steering and braking simultaneously consume the limited tire grip capacity. If the controller uses a fixed or excessively high road grip coefficient, or only controls the trajectory of the tractor without considering the semi-trailer's response, the generated steering and braking force commands may exceed the actual available tire grip capacity.
[0004] For semi-trailer trucks on icy and snowy roads with low adhesion, if there is a lack of consistent data transmission and execution feedback between state estimation, dynamic safety constraints, master-trailer coordinated control, and wheel-end braking force distribution, the adhesion capability and execution capability adopted by the controller may be inconsistent with the actual working conditions. Summary of the Invention
[0005] In view of this, this application aims to propose a cooperative adaptive obstacle avoidance control method and system for main and trailer vehicles on icy and snowy roads, in order to solve the problem that when an autonomous semi-trailer truck train performs emergency obstacle avoidance on icy and snowy low-adhesion roads, the control commands exceed the available tire adhesion capacity and cause the semi-trailer to sideslip, oscillate or fold due to inaccurate estimation of road adhesion capacity, uncoordinated motion response between the tractor and the semi-trailer, and the inability of the fixed articulation angle constraint to reflect the actual stability boundary of the vehicle.
[0006] To achieve the above objectives, the technical solution of this application is implemented as follows: Firstly, this application provides a cooperative adaptive obstacle avoidance control method for main-trailer vehicles on icy and snowy roads, including: The motion states of the tractor and semi-trailer and the wheel end execution feedback are obtained. The motion states of the main vehicle and the road surface adhesion coefficient are jointly estimated, and the uncertainty is estimated based on the adhesion coefficient to determine the conservative adhesion coefficient. The motion state of the main trailer and the conservative adhesion coefficient are input into the plane-coupled dynamic model of the main trailer to update the tire slip angle, tire lateral force and adhesion boundary of each equivalent axle, and obtain the prediction model at the current vehicle running point. Based on the conservative adhesion coefficient, axle vertical load and longitudinal adhesion occupancy status, the remaining lateral adhesion capacity and dynamic articulation angle safety envelope are determined. The obstacle avoidance safety corridor, the dynamic articulation angle safety envelope, the tire adhesion boundary and the actuator capability boundary are used as prediction time domain constraints to jointly solve the target turning angle of the tractor front wheel, the target additional yaw moment of the tractor and the target additional yaw moment of the semi-trailer. Based on the vertical load, predicted lateral force, remaining adhesion, and braking capacity of each controllable wheel, the target additional yaw moment of the tractor and semi-trailer is optimized and distributed wheel-by-wheel braking force, and the target braking pressure of the electronically controlled braking system is corrected in a closed loop based on the real-time slip ratio of each wheel.
[0007] Secondly, based on the same inventive concept, this application also provides a tractor-trailer cooperative adaptive obstacle avoidance control system for icy and snowy roads, including: The vehicle information collection and joint estimation module is configured to acquire the motion state of the tractor and semi-trailer and the wheel end execution feedback, jointly estimate the motion state of the main and trailer vehicles and the road surface adhesion coefficient, and estimate the uncertainty based on the adhesion coefficient to determine the conservative adhesion coefficient. The prediction model update module is configured to input the motion state of the main trailer and the conservative adhesion coefficient into the main trailer planar coupled dynamic model, update the tire slip angle, tire lateral force and adhesion boundary of each equivalent axle, and obtain the prediction model at the current vehicle operating point. The dynamic safety constraint module is configured to determine the remaining lateral adhesion capability and dynamic articulation angle safety envelope based on the conservative adhesion coefficient, axle vertical load and longitudinal adhesion occupancy status, and to use the obstacle avoidance safety corridor, the dynamic articulation angle safety envelope, tire adhesion boundary and actuator capability boundary as prediction time domain constraints, and to jointly solve the target turning angle of the tractor front wheel, the target additional yaw moment of the tractor and the target additional yaw moment of the semi-trailer. The braking force distribution module is configured to optimize the distribution of the target additional yaw moment of the tractor and semi-trailer wheel by wheel-by-wheel braking force based on the vertical load, predicted lateral force, remaining adhesion and braking capacity of each controllable wheel, and to perform closed-loop correction of the target braking pressure of the electronic braking system based on the real-time slip ratio of each wheel.
[0008] Thirdly, based on the same inventive concept, this application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method described in the first aspect.
[0009] Fourthly, based on the same inventive concept, this application also provides a non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores computer instructions for causing the computer to perform the method as described in the first aspect.
[0010] Compared with existing technologies, the cooperative adaptive obstacle avoidance control method and system for main-trailer vehicles on icy and snowy roads described in this application has the following advantages: (1) This application jointly estimates the lateral motion state of the tractor, the motion state of the semi-trailer and the road adhesion coefficient, and constructs a conservative adhesion coefficient based on the uncertainty of the adhesion coefficient estimation. The conservative adhesion coefficient is used to update the tire force boundary and the main trailer cooperative prediction model, which helps to reduce the impact of insufficient road adhesion capability identification, sensor noise and tire model error on the control results, and reduces the situation where the controller outputs excessive steering or braking torque due to overestimation of the adhesion capability of the icy and snowy road surface.
[0011] (2) This application generates a dynamic articulation angle safety envelope online based on the conservative adhesion coefficient, longitudinal adhesion occupancy state, vehicle speed and main trailer vehicle structural parameters, and introduces the safety envelope, obstacle avoidance safety corridor, tire adhesion constraint and actuator capability constraint into the model predictive control process. By jointly solving the front wheel turning angle of the tractor, the additional yaw moment of the tractor and the additional yaw moment of the semi-trailer, it is beneficial to coordinate between obstacle avoidance trajectory tracking and main trailer stability, and reduce the risk of semi-trailer sideslip, swaying and main trailer folding.
[0012] (3) Based on the vertical load, predicted lateral force, remaining adhesion capacity and braking capacity of each controllable wheel, this application optimizes the distribution of the target additional yaw moment of the tractor and semi-trailer for wheel-by-wheel braking force, and performs closed-loop correction of the EBS target braking pressure based on the real-time slip ratio of each wheel; this is beneficial to reduce the distribution of excessive braking force to wheels that are already close to the adhesion limit, reduce the risk of wheel lock-up and lateral guidance capacity reduction during differential braking, and improve the rationality of the utilization of limited adhesion capacity between the tractor and trailer wheels.
[0013] (4) This application feeds back the actual braking force, actual slip ratio, brake saturation state and additional yaw moment actually generated by the main and trailer vehicles to the next control cycle to update the state estimation, prediction model and control input boundary. When the reference obstacle avoidance trajectory or the target additional yaw moment exceeds the current achievable capability of the vehicle, it outputs a request to reduce vehicle speed or replan to the upper planning module, which helps to reduce the prediction deviation caused by the inconsistency between the target control quantity and the actual execution quantity, and improves the executability and closed-loop adaptability of obstacle avoidance control on ice and snow low adhesion road surfaces. Attached Figure Description
[0014] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a schematic diagram of the overall process of the adaptive obstacle avoidance control method for tractor-trailer co-operation on icy and snowy roads described in the embodiments of this application; Figure 2 This is a schematic diagram of the planar coupled dynamics model of the tractor-semi-trailer, the articulation relationship between the main trailer and the trailer, the tire slip angle, and the vehicle geometric parameters as described in the embodiments of this application. Figure 3 This is a schematic diagram showing the relationship between sensor input, unscented Kalman filtering recursive process, and conservative adhesion coefficient output of the joint state and road surface adhesion estimation module described in this application embodiment; Figure 4 This is a schematic diagram illustrating the relationship between generating a dynamic hinge angle safety envelope based on a conservative adhesion coefficient and available lateral adhesion capability, and forming time-varying constraints for model predictive control, as described in an embodiment of this application. Figure 5 This is a schematic diagram showing the relationship between the state of the master-slave collaborative model predictive controller and the model input, objective function, time-varying constraints, online solution and control output of the controller described in the embodiments of this application. Figure 6 This is a schematic diagram illustrating the optimized distribution of braking force to each wheel of the tractor and semi-trailer, the conversion of target braking torque and braking pressure, and the EBS slip ratio closed-loop control relationship as described in the embodiments of this application. Figure 7 This is a schematic diagram of a tractor-trailer cooperative adaptive obstacle avoidance control system for icy and snowy roads, as described in an embodiment of this application. Figure 8 This is a schematic diagram of the hardware structure of the electronic device described in an embodiment of this application. Detailed Implementation
[0015] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with specific embodiments and the accompanying drawings.
[0016] It should be noted that, unless otherwise defined, the technical or scientific terms used in the embodiments of this application should have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms "first," "second," and similar terms used in the embodiments of this application do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed after the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are only used to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.
[0017] The embodiments of this application are described in detail below with reference to the accompanying drawings.
[0018] Example 1 This embodiment uses a 6×4 tractor-semi-trailer equipped with a steer-by-wire system and an electronic braking system (EBS) as the control object, wherein the semi-trailer mass... The weight of the tractor is 32,000 kg. The parameters are determined based on the vehicle's curb weight and load status. The vehicle is equipped with a tractor inertial measurement unit (IMU), a semi-trailer yaw rate sensor, an articulation angle sensor, a front wheel steering angle sensor, wheel speed sensors, a steer-by-wire actuator, an EBS (Electronic Steering System), and a vehicle controller. The planning module generates a reference trajectory, feasible corridor, road curvature, and desired vehicle speed based on obstacles and road boundaries.
[0019] In this embodiment, the semi-trailer mass This refers to the total mass of the semi-trailer body and its cargo. (Tractor vehicle mass) The equivalent mass of the load transmitted by the semi-trailer through the saddle is not counted twice; the influence of the saddle support load on the vertical load of each axle of the tractor is considered in the load calculation.
[0020] like Figure 1 As shown, the vehicle controller operates according to the sampling period. The process repeats steps S1 to S4, and feeds back the vehicle's motion status and the actual amount of data executed at the wheel ends to the next cycle.
[0021] This embodiment constructs a planar coupled dynamic model of the tractor-semi-trailer. For example... Figure 2 As shown, establish an inertial coordinate system. Tractor vehicle body coordinate system and semi-trailer body coordinate system The yaw angles of the tractor and the semi-trailer are respectively... and The yaw angular velocities are respectively and The hinge angle and the force at the hinge point satisfy the following: ; ; In the formula, For the hinge point, and These represent the forces acting at the articulation point in the coordinate systems of the tractor and the semi-trailer, respectively. It is a planar rotation matrix determined by the relative rotation angles of the two vehicle coordinate systems.
[0022] Figure 2 In , and These are the equivalent centers of the front axle of the tractor, the rear axle group of the tractor, and the axle group of the semi-trailer, respectively. , and The center of gravity of the tractor unit to , and The longitudinal distance; and The center of gravity of the semi-trailer is as follows: and The longitudinal distance; and These are the yaw moments of inertia of the tractor and the semi-trailer, respectively.
[0023] The equivalent lateral forces of the tractor's front axle, rear axle assembly, and semi-trailer axle assembly are denoted as follows: , and The corresponding sideslip angles are respectively , and The main mount motion state and the cooperative control inputs are as follows: ; In the formula, The lateral velocity of the tractor's center of gravity. For the steering angle of the front wheels of the tractor, and Additional yaw moments are applied to the tractor and semi-trailer, respectively. During model predictive control, a lateral error is added to the aforementioned motion states. and heading error This constitutes a six-dimensional prediction state. The tractor rear axle assembly and the semi-trailer axle assembly are treated as equivalent in the upper-level prediction model, and the braking force at each wheel end is calculated separately in step S4.
[0024] This embodiment uses a rigid planar articulated model on a smooth road, neglecting vehicle pitch, roll, and the planar flexibility of the articulation device. Vertical loads are used as known or estimated scheduling inputs. The vehicle coordinate axes are defined as longitudinally forward and laterally to the left, with yaw angle and yaw moment being positive counterclockwise. The positions of the articulation points in the tractor and semi-trailer coordinate systems are respectively... and The upper layer adopts an equivalent monorail model with the resultant lateral force of the axle acting at the axle center, retaining the steering projection of the longitudinal tire force and the yaw moment generated by the longitudinal braking force; the execution constraints at each wheel end are still handled separately in step S4. This model is used for forward driving conditions, and the execution control is switched at low speeds.
[0025] Step S1: Combined state and road surface adhesion estimation.
[0026] Specifically, in this embodiment, such as Figure 3 As shown, time synchronization, outlier removal, and filtering are performed on the inertial measurement unit of the tractor, the yaw rate sensor of the semi-trailer, the articulation angle sensor, the front wheel steering angle sensor, the wheel speed sensor, and the EBS feedback signal. The vehicle speed is estimated by combining wheel speed, effective rolling radius, and longitudinal acceleration from the inertial measurement unit, and wheel speed signals with significant slippage are downweighted. Longitudinal tire force is estimated based on drive torque, actual braking torque, and wheel rotation state. If the vehicle speed estimation fails or its reliability is insufficient, degraded control is implemented, and the wheel speed of a locked wheel is not directly equated to the vehicle speed.
[0027] To avoid an algebraic loop between the current control command and the current state estimate, the actual control quantity generated by the EBS at the previous sampling time is used as the known input for state prediction. This is used to determine the vehicle state from the... At all times towards the first The input vector for time step prediction is: ; ; In the formula, , The sum of the longitudinal forces of the main trailer defined according to the tire rolling direction; the longitudinal forces of each axle are determined by combining the known front and rear axle distribution information, and the resultant force in the vehicle coordinate system is calculated by the steering projection in step S2. The lateral acceleration of the tractor's center of gravity. The road surface adhesion coefficient is to be estimated.
[0028] State propagation employs the closed-loop dynamic structure of step S2, calculating tire force at each Sigma point using its adhesion coefficient component, without substituting the estimated adhesion coefficient with a conservative adhesion coefficient; the conservative adhesion coefficient is only used for subsequent predictive control and allocation. The road adhesion coefficient is modeled using a random walk between adjacent cycles, with process noise reflecting road surface changes. After compensating IMU measurements to the tractor's center of gravity and removing the gravity component, the measurement function is: ; The joint estimation system is represented as: ; In the formula, and These are the state transition function and the measurement function, respectively. This includes vehicle mass, yaw moment of inertia, geometric dimensions, tire model parameters, and cutoff time. The known scheduling quantities, such as the longitudinal force distribution information of the front and rear axles of the tractor, vertical load, and longitudinal moment of the foundation, have been obtained, but do not include the road adhesion coefficient to be estimated. and These are process noise and measurement noise, respectively. The currently known inputs used for measurement correction consist of the actual measured or reconstructed execution quantities, excluding control commands that have not yet been generated in the current cycle.
[0029] An unscented Kalman filter is used to recursively estimate the above five-dimensional states to obtain the posterior states. Covariance Matrix The state dimension in this embodiment Generate at each sampling time One Sigma point; Figure 3 In , and These represent the Sigma point, the predicted measurement, and the filter gain, respectively. and Let represent the process noise covariance matrix and the measurement noise covariance matrix, respectively. When the lateral excitation is insufficient or the measurement is not sensitive enough to the adhesion coefficient, the previous effective adhesion estimate is maintained and the adhesion estimate variance is increased. The conservative adhesion coefficient is calculated according to the following formula: ; In the formula, This is an estimated value for the road surface adhesion coefficient. For the corresponding estimated variance, The non-negative confidence reduction factor is... and These are the lower and upper limits of the adhesion coefficient, respectively. This indicates that the input is restricted to a range. The saturation function within. The adhesion coefficient used for control is reduced based on the estimated uncertainty and limited to a preset range.
[0030] Step S2: Update the master-attachment collaborative prediction model based on the conservative adhesion coefficient. Specifically, in this embodiment, step S2 updates the main-trailer plane coupling prediction model based on the posterior state and conservative adhesion coefficient output in step S1, as well as the current longitudinal speed, axle load, longitudinal tire force, and the actual control quantity at the previous sampling time.
[0031] Based on the continuity condition of the articulation point velocity, let the lateral velocity of the tractor's articulation point be: ; The semi-trailer's center of gravity velocity then satisfies: ; To avoid low-speed calculation errors, the following definition is provided: ; In the formula, This is the low-speed regularization speed parameter, in units of This is used to avoid the denominator in the calculation of the sideslip angle and slip ratio approaching zero.
[0032] The equivalent axle slip angles are: ; ; ; for The upper limit of available lateral force is calculated based on the conservative adhesion coefficient, vertical load, and longitudinal tire force: ; In the formula, These represent the front axle of the tractor unit, the rear axle group of the tractor unit, and the equivalent axle group of the semi-trailer, respectively. For the vertical load corresponding to the axle, The longitudinal resultant force is the force along the direction of tire rolling. This is the attachment safety reduction factor.
[0033] The lateral tire force is calculated based on the slip angle and the calibrated tire model, and is limited to the range mentioned above. Therefore, when the longitudinal braking force increases or the conservative coefficient of adhesion decreases, the lateral tire force allowed by the model decreases accordingly.
[0034] To simultaneously determine the longitudinal and lateral articulation forces, a closed-loop relationship is established using the plane dynamics of the main and trailer vehicles and the continuity condition of the acceleration at the articulation point. The sampling time scale is omitted in the following formulas; all variables are taken as values at the current solution time. Let... , and Representing vehicle units The longitudinal resultant force, lateral resultant force, and yaw moment about the center of mass, excluding the hinge force, are all acting on the object. In the equivalent monorail model: ; ; ; ; ; In the formula, , and These are the longitudinal resultant forces of the front axle, rear axle group of the tractor, and equivalent axle group of the semi-trailer along the tire rolling direction, respectively, with the driving direction being positive and the braking direction being negative; , and This corresponds to the lateral resultant force in the tire coordinate system. The yaw moment generated by the basic longitudinal tire force is calculated based on the wheel end position and the direction of the force: ; in, For the basic signed longitudinal tire force without differential braking increment; under pure braking conditions, ,therefore: ; Wheel end position and mapping coefficient The same definition is used as in step S4. This only indicates the additional yaw moment relative to the basic braking force, and is not included again in the full wheel-end longitudinal moment.
[0035] make: ; Substituting the planar dynamics of the two vehicle bodies into the acceleration continuity condition at the articulation point, we obtain: ; in: ; ; In the formula, The second-order coefficient matrix is determined by the vehicle mass, yaw moment of inertia, hinge point location, and hinge angle. Let be a known vector determined by external force, external torque, and yaw motion. When the mass and yaw moment of inertia are positive, If the equation is positive definite, the articulation forces in both directions can be uniquely determined. In practice, solving the above linear equations does not require explicit calculation of the matrix inverse. The articulation force of the semi-trailer is obtained from the following formula: ; Therefore, the four-state motion equations of the main suspension are closed: ; The longitudinal velocity is a scheduling quantity measured online, and its preview can be generated using the longitudinal relationship of the same closed model: ; Step S2 generates a longitudinal velocity preview based on the feasible control sequence of the previous cycle and the known or estimated longitudinal force of the shaft. This preview is fixed in this MPC optimization and corrected by the measured state in the next cycle. This includes longitudinal speed preview, conservative adhesion coefficient, axle vertical load, axle longitudinal force, foundation longitudinal moment, and vehicle structural parameters. The axle longitudinal force is treated as a known scheduling quantity within a single optimization. The change in resultant force caused by wheel-end redistribution is reviewed in step S4 and updated by subsequent feedback. This processing belongs to the rolling prediction approximation and does not list longitudinal speed as a separate decision variable for collaborative control.
[0036] By combining the reference trajectory, construct the predicted state and reference input: ; In the formula, For reference trajectory curvature, For reference vehicle speed. Lateral error. Taking the left side of the reference trajectory as positive, the heading error is defined as the heading angle of the tractor minus the tangent direction angle of the reference trajectory; the projection of the reference path is unique and... Within the interval, the error kinematics is: ; ; The reference vehicle speed is used for longitudinal planning and braking demand coordination, but does not directly replace the measured longitudinal speed. Combining the aforementioned error kinematics and closed-loop dynamics, linearization and discretization are performed using the current estimated state and the actual control quantity from the previous moment as the operating points, and the following definition is given: ; Obtain the discrete prediction model updated online: ; In the formula, , and These are the discrete coefficient matrices corresponding to the augmented state, control increment, and known reference input, respectively. The affine compensation terms generated for linearization are all updated at the current running point.
[0037] During augmented state initialization, the control quantity from the previous moment uses the actual front wheel steering angle and the actual additional yaw moment from the feedback reconstruction; the control quantity in the predictive recursion is the sequence to be optimized. Step S2 transfers the augmented state, the discrete prediction model, and the upper limit of available lateral forces for each equivalent axle to step S3.
[0038] Step S3: Dynamic safety constraint construction and master-slave collaborative model predictive control.
[0039] Specifically, in this embodiment, step S3 receives the augmented state, discrete prediction model, and upper limit of available lateral force for each equivalent axle output from step S2, and performs rolling optimization by combining the reference trajectory, road boundary, and obstacle information provided by the planning module. The cooperative control inputs are the front wheel steering angle of the tractor, the additional yaw moment of the tractor, and the additional yaw moment of the semi-trailer.
[0040] Based on the upper limit of the available lateral force for each equivalent axle, calculate the equivalent lateral acceleration control limit calculated based on the vehicle mass: ; Based on the above equivalent lateral acceleration control limits, longitudinal velocity, and the geometric parameters of the main and trailer vehicles, calculate the dynamic articulation angle safety limit: ; In the formula, For the mechanically permissible hinge angle, To be the minimum control limit, satisfying ; For safety reduction factor, The geometric length of the main trailer is used, and the angle is expressed in radians. This formula is a control limit mapping; its parameters and applicable range should be determined based on vehicle parameters and dynamic operating conditions, and it is not intended as a stability guarantee under arbitrary ice and snow conditions. When other conditions remain unchanged and the width limit is not reached, reduced adhesion, increased vehicle speed, or increased longitudinal adhesion occupancy tightens the limit. When the unlimited calculated value is lower than... When this happens, a slowdown or rescheduling should be triggered, not... To replace the insufficient adhesion capacity. The calculation relationship is as follows: Figure 4 As shown.
[0041] The planning module generates lateral safety corridors in the prediction time domain based on road boundaries and obstacle locations. Obstacle boundaries are expanded according to vehicle width, positioning error, obstacle detection error, and safety margin, and the semi-trailer sweep area is checked in conjunction with the predicted articulation angle to obtain the lateral error boundary. and .
[0042] The discrete prediction model established in step S2 is applied to the prediction time domain. Internal recursion. The predicted output includes at least the lateral error, heading error, articulation angle, and yaw rate of the tractor and semi-trailer: ; The objective function of the model predictive controller is: ; In the formula, , and These are the state tracking, control input, and control increment weight matrices, respectively. To predict the time domain, To control the time domain, and .
[0043] Reference output and Using the same component order, the reference values for lateral and heading errors are zero. The reference values for articulation angle and main trailer yaw rate are generated from the reference trajectory, reference speed, and identical main trailer geometry; these cannot all be set to zero under curve conditions. Weighting matrix. , and A positive semi-definite matrix is used, and the direction of the control increment is subject to a positive definite penalty. When the control time domain is shorter than the prediction time domain, the control increment exceeding the control time domain is set to zero.
[0044] The following constraints are applied in the prediction time domain: for Application status and tire adhesion constraints: ; ; ; ; for Apply control inputs and their rate of change constraints: , ; In the formula, and These are the lower and upper limits of the collaborative control input, respectively. and These represent the lower and upper limits of the control input variation, respectively. These constraints are defined in this embodiment. Take the lateral acceleration of the tractor's center of gravity, i.e. ; To control the equivalent limits used in the design, they are not interpreted as strictly analytical values of the independent limit lateral acceleration of each vehicle body.
[0045] Hard constraints are applied to the collision safety corridor, tire adhesion boundary, and dynamic articulation angle boundary; trajectory tracking accuracy, main trailer stability, and control smoothness are determined by the weight matrix in the objective function. , and Coordination is carried out. The upper and lower limits of the additional yaw moment are updated in real time based on the available braking force at the wheel end and the braking capacity fed back from the previous control cycle, in order to reduce the risk that the output command of step S3 will exceed the current estimated executable range, and will be corrected through the actual execution feedback of step S4.
[0046] Within each control cycle, the conservative adhesion coefficient, predicted longitudinal tire force, and vertical load used for this solution are used as known parameters to pre-calculate the upper limit of available lateral force and the dynamic safety boundary at each prediction time. For predicted conditions where the longitudinal tire force does not exceed the adhesion boundary, the tire adhesion constraint is transformed into upper and lower bound constraints for lateral force. ; The tire lateral force and lateral acceleration are approximated using a first-order approximation near the current operating point. The safety corridor and dynamic articulation angle boundaries remain fixed during this solution process. If the predicted longitudinal tire force exceeds the corresponding adhesion boundary, the predicted condition is deemed infeasible, triggering braking demand coordination or replanning. Based on the above processing, the objective function and constraints are reorganized into a quadratic programming problem and solved online, executing only the first set of control variables in the optimal control sequence. ; in, Send to the steering actuator by steer-by-wire and Send to step S4. In the next sampling period, re-estimate the vehicle state and update the prediction model and constraints. If no control sequence satisfies the hard constraints, request a reduction in the reference speed; if necessary, trigger emergency braking. The master-slave cooperative model predictive control relationship is as follows: Figure 5 As shown.
[0047] Step S4: Optimized distribution of braking force to the main trailer wheels and closed-loop control of EBS slip ratio.
[0048] Specifically, in this embodiment, step S4 receives the target additional yaw moment of the tractor and semi-trailer output in step S3, and combines the conservative adhesion coefficient, vertical load of each wheel, lateral force, wheel speed and braking capacity to solve the target braking force at each wheel end.
[0049] set up and These are the controllable moving wheels of the tractor and semi-trailer, respectively. Vehicle Unit The Middle The maximum available braking force for each wheel is: ; In the formula, Indicates a tractor unit. Indicates a semi-trailer; This is the braking adhesion safety reduction factor; This is the maximum braking torque at the wheel end; This is the effective rolling radius of the wheel.
[0050] The basic braking force is determined based on the vehicle's longitudinal deceleration requirements and the vertical load on each wheel. And an additional yaw moment is generated through differential braking of the left and right wheels: ; In the formula, and The longitudinal and lateral positions of the wheel relative to the center of mass of its vehicle unit. To correspond to the wheel turning angle, The mapping coefficient from wheel braking force to additional yaw moment.
[0051] Vectorize the braking forces of all controllable wheels Solving for the brake force distributor under the current known wheel-end conditions: ; In the formula, Mapping coefficients by vehicle unit Form a two-row matrix; Enter in the corresponding vehicle unit row All other elements are zero; ; The longitudinal braking demand vector of the main trailer; This serves as the basic braking force corresponding to this demand; This refers to the feasible target braking force vector obtained and actually issued in the previous control cycle. For the vehicle unit The The components corresponding to each wheel. For a diagonal normalized matrix, the first... Take the diagonal elements To prevent small positive numbers from being divided by zero, here To standardize wheel end numbering. , and It is a symmetric positive semidefinite weight matrix. The target braking force for each wheel is obtained under the following constraints: ; In the formula, and These represent the lower and upper limits of the wheel braking force variation in adjacent control cycles. When the target additional yaw moment exceeds the currently achievable range, the braking force combination closest to the target within the constraint range is output, and the achievable additional yaw moment range is fed back to step S3. After allocation, the longitudinal force and remaining lateral force boundaries are recalculated according to the target braking force of each wheel, and it is checked whether they still meet the constraints adopted in step S3. If not, the updated boundaries are used to resolve or a downgrade process is entered, and the allocation is no longer determined to be executable based solely on the upper and lower limits of the moment in the previous cycle.
[0052] The target braking force for each wheel is converted into target braking torque, base braking pressure, and target slip ratio: , , ; In the formula, For the calibration mapping of braking torque to braking pressure, To determine the calibration relationship for the target slip ratio based on the conservative adhesion coefficient, wheel load, and target braking force, For quasi-static feedforward conversion, wheel speed dynamics and conversion errors are corrected by wheel end slip ratio feedback. The target slip ratio is limited to a stable region before the current tire adhesion characteristics peak.
[0053] The actual braking slip ratio of the wheel is: ; In the formula, This is the projection of the wheel center velocity onto the direction of the wheel's rolling motion. The calculation takes into account the longitudinal velocity, lateral velocity, yaw rate, wheel end position, and turning angle of the vehicle body. Let be the wheel angular velocity. Define the slip ratio error: ; EBS corrects the wheel-end braking pressure based on slip ratio error: ; In the formula, This is the maximum permissible braking pressure at the wheel end. and These represent the proportional gain and integral gain of the closed-loop control corresponding to the wheel-end slip ratio. The adjustment direction of the slip ratio error feedback is as follows: reduce pressure when the actual slip ratio is too high, and increase or maintain pressure when it is too low. When the pressure reaches the upper or lower limit, a wheel-end fault occurs, or the control mode is switched, the integral term is limited, frozen, or reset to avoid integral saturation; a smooth switch is performed when the mode is restored. When the vehicle is at low speed, the control switches to braking torque or braking pressure tracking to avoid abnormal slip ratio calculation.
[0054] Based on the actual braking force fed back by EBS, the actual additional yaw moment of the tractor and semi-trailer is calculated according to the aforementioned yaw moment mapping. and The actual braking force, actual slip ratio, actual additional yaw moment, and the currently achievable yaw moment range are fed back to steps S1-S3 for state estimation, model updating, and constraint correction in the next sampling period. The wheel-end braking force distribution and EBS closed-loop relationship are as follows: Figure 6 As shown.
[0055] The above EBS discrete control formula uses the same sampling period as the upper-level control. This indicates that when EBS uses a shorter wheel-end control cycle, an independent wheel-end sampling sequence number should be used, and the sampling interval in the slip ratio error integral should be replaced with the actual wheel-end control cycle; the corresponding allocation target should be maintained between two adjacent upper-level command updates.
[0056] When the adhesion estimation reliability is insufficient, the conservative adhesion coefficient is reduced by increasing the estimation covariance. When the optimization problem is infeasible, a request is made to reduce the reference vehicle speed or trigger emergency braking, without relaxing the hard constraints on collision, tire adhesion, and mechanical articulation angle. For wheel ends that have lost braking output capability, their upper limit of distributable braking force is set to zero. For wheel ends with brake stagnation or inability to release pressure, their actual braking force is included in the prediction as an uncontrollable action, and redistribution is performed using normal wheel ends, followed by fault degradation. When the master-slave communication is abnormal, the timeout state is discontinued and deceleration or minimum risk control is entered. After obstacle avoidance, the additional yaw moment is smoothly reduced and normal lane tracking is restored. When a fault switch or sudden drop in adhesion causes a conflict between the braking force change constraint and the physical limit, the achievable physical limit is prioritized according to the fault degradation rules, and the old instructions that are no longer feasible are not executed.
[0057] It should be noted that, without changing the data transmission relationship in steps S1 to S4, the joint state estimator can also employ extended Kalman filtering, commensurate Kalman filtering, or other nonlinear state estimation methods; the tire model can employ a brush model, Fiala model, magic formula, piecewise linear model, or experimental calibration model; the main-trailer coordinated controller can employ linear time-varying model predictive control, nonlinear model predictive control, or other constrained optimization methods; braking force distribution can employ optimization models such as quadratic programming and solve them using algorithms such as the effective set method, or employ a control distribution method with saturation correction; EBS wheel-end control can employ proportional-integral control, gain scheduling control, or other slip ratio closed-loop control methods.
[0058] The conservative adhesion coefficient can be calculated separately for the whole vehicle, different axles, or different wheel ends, but it is used to synchronously update the tire force boundary, dynamic articulation angle safety envelope, and wheel end braking force constraint. The control system can adopt a centralized architecture or a distributed architecture in which the tractor controller and the semi-trailer controller communicate with each other.
[0059] For semi-trailers with other mass or load conditions, the mass, yaw moment of inertia, axle vertical load, and braking force distribution boundaries should be updated accordingly without changing the overall control structure between steps S1 to S4 of this application.
[0060] It should be noted that the above description describes some embodiments of this application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in a different order than that shown in the above embodiments and still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0061] Based on the same inventive concept, and corresponding to any of the above embodiments, the embodiments of this application also provide a tractor-trailer cooperative adaptive obstacle avoidance control system for tractors on icy and snowy roads.
[0062] like Figure 7 As shown, the icy and snowy road tractor-trailer cooperative adaptive obstacle avoidance control system includes: The vehicle information collection and joint estimation module 11 is configured to acquire the motion state of the tractor and semi-trailer and the wheel end execution feedback, jointly estimate the motion state of the main vehicle and the road surface adhesion coefficient, and estimate the uncertainty based on the adhesion coefficient to determine the conservative adhesion coefficient. The prediction model update module 12 is configured to input the motion state of the main trailer and the conservative adhesion coefficient into the plane coupled dynamic model of the main trailer, update the tire slip angle, tire lateral force and adhesion boundary of each equivalent axle, and obtain the prediction model at the current vehicle running point. The dynamic safety constraint module 13 is configured to determine the remaining lateral adhesion capability and dynamic articulation angle safety envelope based on the conservative adhesion coefficient, axle vertical load and longitudinal adhesion occupancy state. It uses the obstacle avoidance safety corridor, dynamic articulation angle safety envelope, tire adhesion boundary and actuator capability boundary as prediction time domain constraints, and jointly solves the target turning angle of the tractor front wheel, the target additional yaw moment of the tractor and the target additional yaw moment of the semi-trailer. The braking force distribution module 14 is configured to optimize the distribution of the target additional yaw moment of the tractor and semi-trailer wheel by wheel-by-wheel braking force based on the vertical load, predicted lateral force, remaining adhesion and braking capacity of each controllable wheel, and to perform closed-loop correction of the EBS target braking pressure based on the real-time slip ratio of each wheel.
[0063] For ease of description, the above system is described by dividing it into various modules based on their functions. Of course, in implementing the embodiments of this application, the functions of each module can be implemented in one or more software and / or hardware.
[0064] The system described in the above embodiments is used to implement the corresponding method in any of the foregoing embodiments and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.
[0065] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, embodiments of this application also provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the methods described in any of the above embodiments.
[0066] Figure 8 This embodiment illustrates a more specific hardware structure of an electronic device, which may include a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, memory 1020, input / output interface 1030, and communication interface 1040 are interconnected internally via the bus 1050.
[0067] The processor 1010 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.
[0068] The memory 1020 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program code is stored in the memory 1020 and is called and executed by the processor 1010.
[0069] The input / output interface 1030 is used to connect input / output modules to realize information input and output. The input / output modules can be configured as components in the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices may include keyboards, mice, touch screens, microphones, various sensors, etc., and output devices may include displays, speakers, vibrators, indicator lights, etc.
[0070] The communication interface 1040 is used to connect a communication module (not shown in the figure) to enable communication between this device and other devices. The communication module can communicate via wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).
[0071] Bus 1050 includes a pathway for transmitting information between various components of the device, such as processor 1010, memory 1020, input / output interface 1030, and communication interface 1040.
[0072] It should be noted that although the above-described device only shows the processor 1010, memory 1020, input / output interface 1030, communication interface 1040, and bus 1050, in specific implementations, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the embodiments of this specification, and not necessarily all the components shown in the figures.
[0073] The electronic devices described above are used to implement the corresponding methods in any of the foregoing embodiments and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.
[0074] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, this application also provides a non-transitory computer-readable storage medium that stores computer instructions for causing the computer to perform the methods described in any of the above embodiments.
[0075] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.
[0076] The computer instructions stored in the storage medium of the above embodiments are used to cause the computer to perform the methods described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.
[0077] Although this application has been described in conjunction with specific embodiments thereof, many substitutions, modifications, and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may be used with the embodiments discussed.
[0078] The embodiments of this application are intended to cover all such substitutions, modifications, and variations that fall within the broad scope of the claims of this application. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the embodiments of this application should be included within the protection scope of this application.
Claims
1. A cooperative adaptive obstacle avoidance control method for main-trailer vehicles on icy and snowy roads, characterized in that, include: The motion states of the tractor and semi-trailer and the wheel end execution feedback are obtained. The motion states of the main vehicle and the road surface adhesion coefficient are jointly estimated, and the uncertainty is estimated based on the adhesion coefficient to determine the conservative adhesion coefficient. The motion state of the main trailer and the conservative adhesion coefficient are input into the plane-coupled dynamic model of the main trailer to update the tire slip angle, tire lateral force and adhesion boundary of each equivalent axle, and obtain the prediction model at the current vehicle running point. Based on the conservative adhesion coefficient, axle vertical load and longitudinal adhesion occupancy status, the remaining lateral adhesion capacity and dynamic articulation angle safety envelope are determined. The obstacle avoidance safety corridor, the dynamic articulation angle safety envelope, the tire adhesion boundary and the actuator capability boundary are used as prediction time domain constraints to jointly solve the target turning angle of the tractor front wheel, the target additional yaw moment of the tractor and the target additional yaw moment of the semi-trailer. Based on the vertical load, predicted lateral force, remaining adhesion, and braking capacity of each controllable wheel, the target additional yaw moment of the tractor and semi-trailer is optimized and distributed wheel-by-wheel braking force, and the target braking pressure of the electronically controlled braking system is corrected in a closed loop based on the real-time slip ratio of each wheel.
2. The method according to claim 1, characterized in that, During joint estimation, the feedback signals from the tractor's inertial measurement unit, the semi-trailer's yaw rate sensor, the articulation angle sensor, the front wheel steering angle sensor, the wheel speed sensor, and the electric braking system are synchronized and filtered. The tractor's lateral speed, tractor's yaw rate, main-trailer articulation angle, semi-trailer's yaw rate, and road adhesion coefficient constitute the state to be estimated. The actual control quantity generated by the electric braking system at the previous sampling time is used as the known input for state prediction. When the lateral excitation or the observability of the adhesion coefficient is insufficient, the previous effective adhesion estimate is maintained and the variance of the adhesion estimate is increased.
3. The method according to claim 1, characterized in that, The conservative adhesion coefficient is obtained by subtracting the confidence reduction factor and the standard deviation of the adhesion estimate from the road adhesion coefficient estimate and then subjecting it to upper and lower limits. During the state propagation phase, the tire force is calculated using the estimated adhesion coefficient. The conservative adhesion coefficient is used for subsequent predictive control and braking force distribution, and is also used to update the tire lateral force boundary, dynamic articulation angle control limit, and wheel end available braking force boundary simultaneously.
4. The method according to claim 1, characterized in that, The main-trailer planar coupled dynamics model determines the motion relationship of the semi-trailer's center of gravity based on the continuous velocity condition at the articulation point. It solves the articulation force by combining the planar dynamics of the main-trailer vehicles with the continuous acceleration condition at the articulation point. Based on the conservative adhesion coefficient, the vertical load of each equivalent axle, and the longitudinal tire force, it determines the upper limit of the available lateral force of each equivalent axle. The tire model output is limited to the corresponding available lateral force range. The model is linearized and discretized at the current estimated state and the actual control quantity at the previous moment to form a linear time-varying prediction model.
5. The method according to claim 1, characterized in that, The generation of the dynamic hinge angle safety envelope includes: The equivalent lateral acceleration control limit is determined based on the upper limit of the available lateral force of each equivalent axle and the total vehicle mass. Then, the articulation angle control limit is determined based on the equivalent lateral acceleration control limit, longitudinal velocity and geometric characteristic length of the main and trailer vehicles. The amplitude is limited by the mechanically permissible articulation angle and the preset minimum control limit. When the unlimited hinge angle control limit is lower than the preset minimum control limit, a request to reduce the reference vehicle speed or replan is triggered.
6. The method according to claim 1, characterized in that, The model predictive controller uses lateral error, heading error, articulation angle, and yaw rate of the tractor and semi-trailer as predictive outputs. In the prediction time domain, the obstacle avoidance safety corridor, dynamic articulation angle boundary, and tire adhesion boundary are set as hard constraints, and the front wheel rotation angle, the additional yaw moment of the main and semi-trailer and their rate of change are applied as constraints. The objective function is formed by state tracking term, control input term and control increment term. The first set of control variables in the optimal control sequence is solved and executed in a rolling manner.
7. The method according to claim 1, characterized in that, When distributing braking force, the basic braking force is determined based on the longitudinal deceleration requirements and the vertical load of each wheel. The mapping between braking force and additional yaw moment is established based on the position of each wheel relative to the center of mass of its respective vehicle unit and the wheel rotation angle. Under the constraints of the adhesion boundary of each wheel, the braking capacity and the rate of change of braking force, the actual additional yaw moment of the tractor and semi-trailer is made to approach the corresponding target additional yaw moment.
8. The control method according to claim 7, characterized in that, After the braking force distribution is completed, the predicted longitudinal force of the vehicle unit is updated according to the target braking force and wheel rotation angle of each wheel, and the remaining lateral adhesion boundary is updated wheel by wheel in combination with the vertical load of each wheel, tire lateral force and conservative adhesion coefficient. The target braking force of each wheel is converted into target braking torque, basic braking pressure and target slip ratio. The electronic control braking system corrects the wheel-end braking pressure according to the deviation between the actual slip ratio and the target slip ratio, and switches to braking torque or braking pressure tracking at low speed.
9. The control method according to claim 1, characterized in that, The actual additional yaw moment of the tractor and semi-trailer is reconstructed based on the actual braking force fed back by the electronic braking system. The actual braking force, actual slip ratio, actual additional yaw moment and the current achievable additional yaw moment range are fed back to the next control cycle to update the state estimate, prediction model and control input boundary.
10. A tractor-trailer cooperative adaptive obstacle avoidance control system for tractors on icy and snowy roads, characterized in that, include: The vehicle information collection and joint estimation module is configured to acquire the motion state of the tractor and semi-trailer and the wheel end execution feedback, jointly estimate the motion state of the main and trailer vehicles and the road surface adhesion coefficient, and estimate the uncertainty based on the adhesion coefficient to determine the conservative adhesion coefficient. The prediction model update module is configured to input the motion state of the main trailer and the conservative adhesion coefficient into the main trailer planar coupled dynamic model, update the tire slip angle, tire lateral force and adhesion boundary of each equivalent axle, and obtain the prediction model at the current vehicle operating point. The dynamic safety constraint module is configured to determine the remaining lateral adhesion capability and dynamic articulation angle safety envelope based on the conservative adhesion coefficient, axle vertical load and longitudinal adhesion occupancy status, and to use the obstacle avoidance safety corridor, the dynamic articulation angle safety envelope, tire adhesion boundary and actuator capability boundary as prediction time domain constraints, and to jointly solve the target turning angle of the tractor front wheel, the target additional yaw moment of the tractor and the target additional yaw moment of the semi-trailer. The braking force distribution module is configured to optimize the distribution of the target additional yaw moment of the tractor and semi-trailer wheel by wheel-by-wheel braking force based on the vertical load, predicted lateral force, remaining adhesion and braking capacity of each controllable wheel, and to perform closed-loop correction of the target braking pressure of the electronic braking system based on the real-time slip ratio of each wheel.
