CN115712950A - Automatic driving decision-making method for semi-trailer - Google Patents

Abstract

The invention discloses an automatic driving decision method for a semi-trailer vehicle, which comprises the following steps: acquiring the current motion state of the semi-trailer vehicle, and predicting all motion state spaces of the semi-trailer vehicle based on the dynamic characteristics of the semi-trailer vehicle and the current motion state; narrowing the range of the entire motion state space based on surrounding traffic environment information; calculating the running performance of the semi-trailer vehicle; a future movement state is preferred in the overall movement state space as a future travel target for the semi-trailer vehicle on the basis of the travel performance. The method is used for modeling the semi-trailer in motion prediction, motion stability analysis and driving performance analysis, so that the applicability of the strategy model to the semi-trailer is ensured, discontinuity of multi-target decision results is effectively avoided by reducing the target motion state space layer by layer, and meanwhile, the dependence on the hardware system performance is reduced by adopting a multi-level double-index optimization method.

Description

An automatic driving decision-making method for semi-trailer vehicles

technical field

The invention belongs to the technical field of automatic control of automobiles, in particular to an automatic driving decision-making method for a semi-trailer automobile.

Background technique

One of the important directions of the current development of the automobile industry is intelligence. The industry generally believes that commercial vehicle autonomous driving will be the first to land. With the continuous advancement of unmanned driving technology, the commercialization of assisted driving technology for single-body passenger vehicles has achieved success in basic scenarios. Autonomous driving technology still needs to explore and overcome some difficult problems. Among them, the comprehensive decision-making of the direction and speed of the trailer vehicle considering the driving stability is one of the difficult problems in the automatic driving of the trailer vehicle.

Regarding the decision-making and control of unmanned vehicles, the invention patent application number 201110007154.X proposes a device and method for local path planning of unmanned vehicles. The invention uses the method of mechanical field diagram to determine the future optional driving trajectory . This method considers traffic rules, static objects, dynamic objects, and abstract events in a unified way, which increases the complexity of the problem and makes it difficult to obtain ideal decision-making results.

The invention patent with application number 201410221906.6 describes a driverless car control system with social behavior interaction. This invention will use the hidden Markov model to judge the driving intention of other vehicles, and then determine the optimal trajectory of the vehicle on this basis. decision making. This method needs to establish a hidden Markov model for each adjacent traffic vehicle. Since the adjacent vehicles are difficult to track for a long time in practice, the accuracy of the established model is difficult to guarantee.

The invention patent with application number 201810007699.2 describes a driving task decision system and method for unmanned vehicles. This invention adds regularity evaluation to all tasks in the driving task set, and removes the driving tasks that do not conform to traffic rules. Driving tasks that are unsafe and inconsistent with the direction of the navigation guide to ensure the safety, legality and efficiency of vehicle driving. The proposed method does not have a targeted design for large multi-body commercial vehicles, and is only suitable for the application of small single-body vehicles, that is, passenger cars.

Contents of the invention

The purpose of the present invention is to provide a kind of automatic driving decision-making method for the semi-trailer vehicle, to solve the above-mentioned problems in the prior art.

In order to achieve the above object, the invention provides a kind of automatic driving decision-making method for semi-trailer vehicles, comprising:

Obtaining the current motion state of the semi-trailer, and predicting the entire motion state space of the semi-trailer based on the dynamic characteristics of the semi-trailer and the current motion state;

Narrowing down the scope of the entire motion state space based on the surrounding traffic environment information;

Calculate the driving performance of the semi-trailer;

Based on the driving performance, a future motion state is selected in the entire motion state space as the future driving target of the semi-trailer vehicle.

Optionally, the process of predicting the entire motion state space of the semi-trailer based on the dynamic characteristics of the semi-trailer and the current state of motion includes:

Based on the wheelbase and wheelbase of the semi-trailer, the range of yaw rate change is analyzed, and the range of acceleration change is analyzed based on the power specification of the semi-trailer;

Based on the current state of motion, analyze the yaw rate acceleration variation range of the semi-trailer vehicle at the next moment;

Uniformly discretize the variation range of the yaw rate and acceleration of the semi-trailer vehicle at the next moment, perform an orthogonal combination of each yaw rate and acceleration to obtain the future motion state, and obtain the entire motion state space based on the future motion state.

Optionally, the process of predicting the entire motion state space of the semi-trailer based on the dynamic characteristics of the semi-trailer and the current state of motion also includes:

Constructing a kinematics model based on the longitudinal acceleration and yaw rate of the semi-trailer, and obtaining the motion state of the semi-trailer based on the kinematics model;

Obtaining a future motion state based on the result of the motion state and the orthogonal combination, and performing a motion failure analysis of the semi-trailer vehicle based on the future motion state;

Based on the result of the motion failure analysis, future motion states that may cause motion failure in the entire motion state space are excluded.

Optionally, the process of performing the motion failure analysis of the semi-trailer vehicle based on the future motion state includes:

The lateral load transfer rate of the trailer and trailer is calculated based on the distance between the center of mass of the trailer and the trailer axle, the fifth wheel and the wheelbase, the distance to the roll axis, the sprung mass of the trailer, the total mass of the trailer, the lateral acceleration, and the body roll angle; based on The motion characteristics of the semi-trailer vehicle and the lateral load transfer rate are used to judge whether it is about to roll over, and obtain the rollover analysis result;

Based on the yaw rate of the trailer and the yaw rate of the trailer, the drift and folding problems of the semi-trailer are analyzed, and the analysis results of the drift and fold are obtained;

A motion failure analysis result is obtained based on the rollover analysis result and the flick fold analysis result.

Optionally, the process of narrowing the scope of the entire motion state space based on the surrounding traffic environment information includes:

Based on the position and speed of the vehicle, use environmental perception to obtain the position and line type of lane markings, speed limit signs, and status of traffic lights, and analyze the illegality of semi-trailer vehicles;

Based on the vehicle's motion state, environmental perception is used to obtain the road boundary and the motion state of other traffic participants, and the hazards of semi-trailer vehicles are analyzed.

Optionally, the process of obtaining the road boundary and other traffic participants' motion states by using environment perception includes:

The prediction results of several adjacent frames of the trailer and the trailer are connected through polygons to form a vehicle trajectory envelope, based on the vehicle trajectory envelope, the road boundary is obtained, and the vehicle exceeds the road boundary analysis based on the road boundary;

Based on the circumscribed circle of the semi-trailer vehicle, the circumscribed circle rough collision detection is carried out, and the motion state of other traffic participants is obtained to analyze the vehicle collision.

Optionally, the driving performance of the semi-trailer vehicle includes following performance and driving efficiency, the following performance includes position following and direction following, and the driving efficiency includes efficiency and comfort.

Optionally, the process of calculating the driving performance of the semi-trailer vehicle includes:

Acquiring the position followability based on the distance between the trailer and the center of mass position corresponding to the final state of the trailer's trajectory to the recommended path;

Obtaining the direction followability based on the angle between the trailer and the direction corresponding to the final state of the trailer's trajectory and the recommended path;

obtaining the efficacy based on the longitudinal velocity, the recommended vehicle speed, and the maximum longitudinal velocity and the minimum longitudinal velocity of the trailer at the end state of the trajectory;

The comfort is obtained based on the preview longitudinal acceleration and yaw rate of the trailer, the longitudinal acceleration and yaw rate of the trailer at the current moment, and the maximum value of the preview longitudinal acceleration and yaw rate of the trailer.

Optionally, the process of selecting a future motion state in the entire motion state space based on the driving performance as the future driving target of the semi-trailer vehicle includes:

Using hierarchical binomial optimization method to optimize the binary value of position followability, direction followability, efficacy and comfort respectively;

Combine the binary optimization results of position following and direction following into optimal path following, and combine the binary optimization results of efficacy and comfort into optimal driving efficiency;

Combining the optimal path following and the optimal driving efficiency binary optimization into optimal performance, and determining the future driving target of the semi-trailer based on the optimal performance

Technical effect of the present invention is:

1. The present invention models the semi-trailer in motion prediction, motion stability analysis, and driving performance analysis, which ensures the applicability of the strategy model to the semi-trailer.

2. The present invention adopts layer-by-layer reduction of the target motion state space, and finally optimizes the decision-making method, effectively avoids the discontinuity of multi-objective decision-making results, and ensures the stable driving of the vehicle.

3. The performance analysis of the present invention preferably adopts a multi-level double-index optimization method: first, each optimization only determines a weight value, which avoids the technical problem that the simultaneous optimization weight value of multiple indicators is difficult to determine; secondly, the optimization method of each level Similarly, the reuse rate of purchasing agent is improved, and the dependence of the operation of the method on the performance of the hardware system is reduced.

Description of drawings

The drawings constituting a part of the application are used to provide further understanding of the application, and the schematic embodiments and descriptions of the application are used to explain the application, and do not constitute an improper limitation to the application. In the attached picture:

Fig. 1 is the flow chart of the automatic driving decision-making method for semi-trailer vehicle in the embodiment of the present invention;

Fig. 2 is the semi-trailer vehicle kinematics model figure in the embodiment of the present invention;

Fig. 3 is the semi-trailer vehicle motion prediction effect figure in the embodiment of the present invention;

Fig. 4 is a schematic diagram of judging the legality of the expected driving area in the embodiment of the present invention;

5 is a schematic diagram of a plurality of trailer track envelopes formed by connecting polygons in an embodiment of the present invention;

FIG. 6 is a schematic diagram of rough collision detection of a circumscribed circle of a vehicle in an embodiment of the present invention.

Detailed ways

It should be noted that, in the case of no conflict, the embodiments in the present application and the features in the embodiments can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and embodiments.

It should be noted that the steps shown in the flowcharts of the accompanying drawings may be performed in a computer system, such as a set of computer-executable instructions, and that although a logical order is shown in the flowcharts, in some cases, The steps shown or described may be performed in an order different than here.

Embodiment one

As shown in Figure 2-6, a kind of automatic driving decision-making method for semi-trailer is provided in the present embodiment, comprises the following steps:

Step 1. Arrivable space analysis: According to the dynamic characteristics of the controlled vehicle and the current motion state, analyze all possible motion state spaces of the vehicle in the next decision cycle, which is the reachable space.

Step 2. Drivable space analysis: According to the surrounding traffic environment information, the motion state that causes the vehicle to enter the non-drivable area is excluded on the basis of the reachable space, and the motion state space is further reduced.

Step 3. Driving performance analysis: Considering the driving performance of the vehicle comprehensively, a certain motion state is optimally decided in the motion state space, which is used as the future driving target of the vehicle, and is output from the decision-making strategy.

In order to realize the decision-making of future driving goals, this model must obtain the macro-route and speed distribution scheme planned based on the destination point and arrival time before execution.

In order to realize the decision-making of future driving goals, in step 1, the maneuverability analysis is carried out firstly, and the method is as follows: firstly, the variation range analysis of the yaw rate is carried out according to the wheel base and the wheelbase of the vehicle, and the variation range analysis of the acceleration is carried out according to the power specification of the vehicle; Secondly, based on the current vehicle motion state, analyze the possible variation range of the vehicle’s yaw rate and acceleration in the next decision cycle; finally, uniformly discretize the above range of change, and combine each yaw rate and acceleration orthogonally as a future vehicle A motion state forms the initial motion state space.

In order to realize the decision-making of future driving goals, in step 1, based on the initial motion state space, it is necessary to exclude motion states that may cause vehicle motion failure (such as rollover, tail flick, folding), and the method is as follows:

挂车后轴中心点位置x_t,y_t，挂车航向角

合称为车辆运动状态

First, establish the kinematics model of the vehicle (semi-trailer vehicle) (Fig. 2), as shown in the following formula. The input of the model is the longitudinal acceleration a _h and the yaw rate ω _h of the trailer, and the output is the articulation of the trailer and the trailer Point position x _h , y _h , trailer speed v _h , trailer heading angle

The position of the center point of the rear axle of the trailer x _t , y _t , the heading angle of the trailer

Collectively referred to as vehicle motion state

如图3，并通过计算结果进行车辆运动失效分析，方法如下：Secondly, each combination of acceleration and yaw rate in the initial motion state space is brought into the above formula to calculate the future vehicle motion state

As shown in Figure 3, and the vehicle movement failure analysis is carried out through the calculation results, the method is as follows:

与

为拖车与挂车车身侧倾角。For the analysis of the vehicle rollover problem, calculate the lateral load transfer rate R ₁ and R ₂ of the trailer and the trailer according to the following formula. According to the motion characteristics of the semi-trailer vehicle, if R ₁ >0.6 or R ₂ >0.7, it is judged to be a rollover failure . Among them, d and e are the distances from the center of mass of the trailer to the axle of the trailer and the fifth wheel, respectively, and g is the acceleration due to gravity. T ₁ and T ₂ are the wheel bases of the trailer and trailer, respectively. h ₁ and h ₂ are the distances from the trailer and the trailer to the roll axis, respectively. m _1s is the sprung mass of the trailer, and m ₂ is the total mass of the trailer. a _ym1 and a _ym2 are the lateral acceleration of the trailer and the trailer respectively,

and

is the roll angle of the trailer and trailer body.

为铰接角绝对值的约束。对拖车与挂车之间的铰接角进行限制可间接避免在两结构单元在运动过程中发生机械碰撞。

为拖车的横摆角速度，

为挂车的横摆角速度，

为铰接角横摆角速度绝对值的约束。对拖车与挂车之间的铰接角速度之差进行限制可降低造成拖车后轴轮胎侧向力饱和或挂车车轴的侧向力饱和的概率，从而降低发生折叠与甩尾摆振失稳的可能。For the analysis of vehicle tail drift and folding, since tail drift can be regarded as the trailer sliding outward relative to the steering center, and folding can be regarded as the trailer sliding inward relative to the steering center, the two are unified by the following formula for analysis. in

is the constraint on the absolute value of the articulation angle. Limiting the articulation angle between the trailer and the trailer can indirectly avoid the mechanical collision between the two structural units during the movement.

is the yaw rate of the trailer,

is the yaw rate of the trailer,

is the constraint of the absolute value of the yaw rate at the articulation angle. Limiting the difference in articulation angular velocity between the trailer and the trailer can reduce the probability of saturation of the lateral force of the rear axle tire of the trailer or saturation of the lateral force of the trailer axle, thereby reducing the possibility of folding and tail-flicking instability.

Finally, the set of spatial positions x _h , y _h , x _t , and y _t corresponding to all motion states in the motion state space after excluding the above motion states that may cause failure is taken as the reachable space.

In order to realize the decision-making of future driving goals, in step 2, for each point in the reachable space, illegality and illegality need to be judged and eliminated at the same time to form a feasible space to ensure that the vehicle can drive safely and legally. The method is as follows:

For the analysis of vehicle illegality, the main illegal forms are crossing the line and speeding, which are determined by comparing the position and speed of the vehicle with the lane marking position and line type obtained from environmental perception, speed limit signs, and traffic light status. It is characterized in that: the following polynomial form is used to describe the lane markings, and the double rectangular bounding box is used to describe the own vehicle, so as to reduce the amount of calculation. The legitimacy judgment is shown in Figure 4.

y＝A ₃ x ³ +A ₂ x ² +A ₁ x+A ₀

For the analysis of vehicle risk problems, there are mainly two dangerous forms: exceeding the road boundary and colliding with other traffic participants, which are determined by comparing the vehicle's motion state with the road boundary obtained by environmental perception and the motion state of other traffic participants. In order to further ensure safety, the road boundary and other traffic participants should be multiplied by the safety factor before analysis. Its characteristics are: in order to further reduce the amount of calculation, the prediction results of several adjacent frames of the trailer and the trailer are connected through polygons to form a vehicle trajectory envelope, that is, the entire vehicle trajectory is composed of several trajectory envelopes, and the road boundary cross-border analysis is shown in the figure As shown in 5; the rough collision detection is performed through the circumscribed circle of the vehicle shape, and the circumscribed circle rough collision detection is first composed of the left and right front corner points of the vehicle in the tail frame and the left and right rear corner points of the vehicle in the first frame for the trajectory envelopes of the two vehicles that need to be calculated. Rectangle, because the rectangle is relatively narrow and long, it is necessary to surround the rectangle with multiple circles, as shown in Figure 6, and then calculate the distance from the center of the circle and the sum of the corresponding radius, if the former is greater than the latter, the trajectory envelope of this period is not A collision occurs, and vice versa.

In order to realize the decision-making of future driving goals, in step 3, the following indicators are used to evaluate all the positions in the feasible space, and the optimal value is selected as the target position-motion state. The first-level indicators are following performance and driving efficiency:

Following performance includes two sub-performances, which are position following and direction following:

Position followability reflects the degree of conformity between the evaluated position-motion state and the position of the pre-planned macro path:

In the formula, d _{h, ep} is the distance from the center of mass of the trailer vehicle corresponding to the end state of the trajectory to the recommended route, and d _{t, ep} is the distance from the center of mass of the trailer vehicle corresponding to the end state of the trajectory to the recommended route. min and max represent the maximum value and the minimum value, respectively.

Orientation followability reflects the degree of conformity of the evaluated position-motion state to the tangential direction of the pre-planned macro-path:

In the formula, θ _h is the angle between the direction of the trailer vehicle corresponding to the end state of the trajectory and the recommended route, and θ _t is the angle between the direction of the trailer vehicle corresponding to the end state of the trajectory and the recommended route. min and max represent the maximum value and the minimum value, respectively.

Driving efficiency includes two sub-performances, which are efficacy and comfort:

Efficacy reflects the speed at which the assessed position-motion state makes the driving task complete, and mainly considers the balance between the expected speed and the road speed limit:

where u _p is the longitudinal velocity of the trailer corresponding to the end state of the trajectory, u _ep is the recommended vehicle speed, u _p,max and up _p,min respectively represent the maximum and minimum longitudinal velocity of the trailer in the end state of the trajectory cluster.

Comfort reflects the physical and mental comfort of the occupants in the evaluated position-motion state, and mainly considers the ratio of the acceleration and yaw rate changes of the vehicle to the human capacity.

In the formula, J _{h, lon} is the longitudinal comfort evaluation index, a _{x, p} is the preview longitudinal acceleration of the trailer, a _x is the longitudinal acceleration of the trailer at the current moment, a _{x, pmax} is the maximum value of the trailer preview longitudinal acceleration; J _{h, lat} is the evaluation index of lateral comfort, r _hp is the preview yaw rate of the trailer, r _h is the yaw rate of the trailer at the current moment, r _{h, pmax} is the maximum value of the preview yaw rate of the trailer.

The layered binomial optimization method is used to optimize layer by layer: the first layer is to perform binary optimization of position followability, direction followability, efficacy, and comfort; the second layer is to optimize the binary value of position and speed followability The optimization is combined into the optimal path following, and the binary optimization of efficacy and comfort is combined into the optimal driving efficiency; the third layer, the binary optimization of path following and driving efficiency is combined into the optimal performance, and the final optimal Determine future driving goals. The characteristics of the binary optimization method are: only one weight is determined for each binary comparison, avoiding the simultaneous determination of multiple weights without method support; the same method is used for multi-level optimization, which not only ensures the stability of the method but also improves code reuse. rate, reducing the dependence on hardware computing and storage units. The binary optimization method is shown in the following formula, X ^* is the final output of this model, which is used to guide the work of the vehicle motion control unit to realize the automatic driving function:

x ^* = argmin(J) = argmin[λ·J ₁ +(1-λ)·J ₂ ]

This model uses reachable space analysis and drivable space analysis to reduce the target motion state space layer by layer, and finally optimizes the strategy of decision-making future motion state through driving performance analysis, realizing the motion decision-making function in the automatic driving system. By introducing the kinematics model of the semi-trailer, the motion prediction is more accurate, and the analysis of the typical motion failure forms of the semi-trailer, such as rollover, flicking and folding, is added to ensure the stability of the motion. In the performance analysis, the behavior of the trailer and trailer is considered at the same time, and the following performance and driving efficiency are comprehensively evaluated to ensure that the car can drive the car reasonably while taking into account the realization of the macro planning goals.

Embodiment two

As shown in Figure 1, the present embodiment provides an application example of an automatic driving decision-making method for a semi-trailer vehicle, including:

Step 0, preparation work: before the vehicle drives, complete the planning of the macro path and speed or event allocation based on the driving destination and time requirements, and input the wheelbase, wheelbase and power specifications of the controlled vehicle into the system; when the vehicle starts After driving, the perception results of traffic signals, road boundaries, and traffic participants are input to the system in real time, and the motion status of the vehicle is fed back to the system in real time.

Step 1. Analysis of original maneuverability: Based on the wheelbase, wheelbase and power specifications of the vehicle, analyze the suitable range of acceleration, deceleration and yaw rate of the vehicle.

Step 2, current maneuverability analysis: Based on the current vehicle motion state fed back in real time, analyze the possible range of acceleration, deceleration and yaw rate in a future control cycle, and separate them at equal intervals. Orthogonal combination every acceleration/deceleration and yaw rate to form the original motion state space.

Step 3: Establish a semi-trailer vehicle kinematics model, predict its future trajectory and attitude for each group of motion states in the original motion state space, and judge motion failure forms such as rollover, tail flick, and folding, and eliminate possible failures state of motion to form an accessible space.

Step 4: Carry out illegality and danger judgments for each group of motion states in the accessible space, where illegality judgments include violations of lane marking violations caused by location and speeding violations caused by speed, and hazards include crossing the physical boundary of the road The risk of collision with other traffic participants and the risk of collision with other traffic participants are eliminated, which may lead to illegal and dangerous motion states, thereby reducing the reachable space and forming a drivable space.

Step 5: Evaluate the driving performance of each group of motion states in the drivable space. The evaluation is divided into three levels: the first level is to evaluate the position of the trailer and the proximity of the trailer position to the macro path, and the distance between the trailer and the trailer’s heading and the macro path. The directional evaluation of the tangential consistency of the path, the comprehensive functional evaluation of the consistency of the speed with the expected speed and the speed limit of the law, and the comfort evaluation of the consistency of the acceleration and yaw rate with the tolerance of the driver and passengers; the second level is respectively Carry out the following performance evaluation of comprehensive position and direction, and the driving efficiency evaluation of comprehensive efficacy and comfort; the third level is the driving performance evaluation of comprehensive following performance and driving efficiency.

Step 6: By comparing the driving performance of each group of motion states in the drivable space, select the optimal motion state and its corresponding position and attitude, and output it as the target motion state to complete the work of this decision-making model.

The above is only a preferred embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any person familiar with the technical field can easily conceive of changes or changes within the technical scope disclosed in this application Replacement should be covered within the protection scope of this application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims (9)

1. An automatic driving decision method for a semi-trailer vehicle, comprising the steps of:

acquiring the current motion state of a semi-trailer vehicle, and predicting all motion state spaces of the semi-trailer vehicle based on the dynamic characteristics of the semi-trailer vehicle and the current motion state;

narrowing the range of the entire motion state space based on surrounding traffic environment information;

calculating the running performance of the semi-trailer vehicle;

a future movement state is preferred in the overall movement state space as a future travel target for the semi-trailer vehicle on the basis of the travel performance.

2. The automated driving decision method for a semi-trailer vehicle of claim 1, wherein predicting the overall motion state space of the semi-trailer vehicle based on the dynamics of the semi-trailer vehicle and the current motion state comprises:

analyzing a yaw angular speed change range based on the wheel base and the wheel base of the semi-trailer vehicle, and analyzing an acceleration change range based on the power specification of the semi-trailer vehicle;

analyzing the change range of the yaw velocity and the acceleration of the semi-trailer at the next moment based on the current motion state;

and carrying out uniform discretization on the change range of the yaw velocity and the acceleration of the semi-trailer at the next moment, carrying out orthogonal combination on each yaw velocity and each acceleration to obtain a future motion state, and obtaining all motion state spaces based on the future motion states.

3. The automated driving decision method for a semi-trailer vehicle of claim 2, wherein predicting the overall motion state space of the semi-trailer vehicle based on the dynamics of the semi-trailer vehicle and the current motion state further comprises:

constructing a kinematics model based on the longitudinal acceleration and the yaw angular velocity of the semi-trailer vehicle, and acquiring the motion state of the semi-trailer vehicle based on the kinematics model;

obtaining a future motion state based on the result of the motion state and orthogonal combination, and performing motion failure analysis on the semi-trailer based on the future motion state;

and eliminating future motion states in the overall motion state space which can cause motion failure based on the result of the motion failure analysis.

4. The automated driving decision method for a semi-trailer vehicle of claim 3, wherein the process of performing a kinematic failure analysis of the semi-trailer vehicle based on the future kinematic state comprises:

calculating the transverse load transfer rate of the trailer and the trailer based on the distance between the center of mass of the trailer and the trailer shaft and the fifth wheel, the wheel tread, the distance to a roll axis, the trailer spring load mass, the total mass of the trailer, the lateral acceleration and the vehicle body roll angle; judging whether the semi-trailer vehicle is about to turn over or not based on the motion characteristics of the semi-trailer vehicle and the transverse load transfer rate, and acquiring a rollover analysis result;

analyzing the drift and folding problems of the semi-trailer vehicle based on the yaw velocity of the trailer and the yaw velocity of the trailer to obtain a drift folding analysis result;

and obtaining a result of the movement failure analysis based on the rollover analysis result and the swing folding analysis result.

5. The automated driving decision method for a semi-trailer according to claim 1, wherein narrowing the entire motion state space based on ambient traffic environment information comprises:

based on the position and the speed of the vehicle, acquiring the position and the line type of a lane marking, the states of a speed limit sign and a traffic light by adopting environmental perception, and analyzing the illegal problem of the semi-trailer;

based on the motion state of the vehicle, the environment perception is adopted to obtain the motion states of road boundaries and other traffic participants, and the risk problem of semi-trailer vehicles is analyzed.

6. The automated driving decision method for semi-trailer vehicles according to claim 5, wherein the process of obtaining road boundaries and other traffic participant motion states using environmental awareness comprises:

connecting the prediction results of a plurality of adjacent frames of the trailer and the trailer through a polygon to form a vehicle track envelope, acquiring a road boundary based on the vehicle track envelope, and analyzing the vehicle exceeding the road boundary based on the road boundary;

and carrying out circumscribed circle rough collision detection based on the vehicle-shaped circumscribed circle of the semi-trailer vehicle, acquiring the motion states of other traffic participants, and analyzing vehicle collision.

7. The automated driving decision method for a semi-trailer vehicle of claim 1, wherein the driving performance of the semi-trailer vehicle comprises following performance and driving efficiency, the following performance comprises position following performance and direction following performance, and the driving efficiency comprises efficiency and comfort.

8. The automated driving decision method for semi-trailer vehicles of claim 7, wherein calculating the driving performance of the semi-trailer vehicle comprises:

acquiring the position following performance based on the distance from the center of mass position corresponding to the tail state of the trailer and the trailer motion track to the recommended path;

acquiring the direction following performance based on an included angle between the direction corresponding to the end state of the motion track of the trailer and the recommended path;

acquiring the effectiveness based on the longitudinal speed of the tail state of the motion trail of the trailer, the recommended vehicle speed, the maximum longitudinal speed and the minimum longitudinal speed;

the comfort is obtained based on the maximum values of the longitudinal acceleration and the yaw velocity of the preview of the trailer, the longitudinal acceleration and the yaw velocity of the trailer at the current moment, and the longitudinal acceleration and the yaw velocity of the preview of the trailer.

9. The automated driving decision method for a semi-trailer according to claim 1, wherein the process of preferring a future moving state in the total moving state space as a future driving target of the semi-trailer based on the driving performance comprises:

binary optimization of position following property, direction following property, efficacy and comfort is respectively carried out by adopting a layered two-item optimization method;

combining binary optimization results of position following performance and direction following performance into optimal path following performance, and combining binary optimization results of efficacy and comfort into optimal driving efficiency;

and optimizing and combining the optimal path following performance and the optimal running efficiency binary value into optimal performance, and determining a future running target of the semi-trailer based on the optimal performance.

Images