CN113515125A - Unmanned vehicle full-working-condition obstacle avoidance control method and performance evaluation method - Google Patents
Unmanned vehicle full-working-condition obstacle avoidance control method and performance evaluation method Download PDFInfo
- Publication number
- CN113515125A CN113515125A CN202110754791.7A CN202110754791A CN113515125A CN 113515125 A CN113515125 A CN 113515125A CN 202110754791 A CN202110754791 A CN 202110754791A CN 113515125 A CN113515125 A CN 113515125A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- repulsive force
- obstacle avoidance
- tracking
- control
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 58
- 238000011156 evaluation Methods 0.000 title claims abstract description 35
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 9
- 238000005457 optimization Methods 0.000 claims description 8
- 238000013459 approach Methods 0.000 claims description 5
- 239000011159 matrix material Substances 0.000 claims description 3
- 238000012887 quadratic function Methods 0.000 claims description 3
- 230000035484 reaction time Effects 0.000 claims description 3
- 230000004927 fusion Effects 0.000 claims 2
- 230000004888 barrier function Effects 0.000 claims 1
- 230000000694 effects Effects 0.000 description 12
- 238000013461 design Methods 0.000 description 5
- 238000011160 research Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 230000007423 decrease Effects 0.000 description 3
- 238000012937 correction Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000005096 rolling process Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000011217 control strategy Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
- G05D1/0221—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory involving a learning process
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
- G05D1/0223—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory involving speed control of the vehicle
Landscapes
- Engineering & Computer Science (AREA)
- Aviation & Aerospace Engineering (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
- Steering Control In Accordance With Driving Conditions (AREA)
Abstract
The invention relates to an unmanned vehicle all-condition obstacle avoidance control method and a performance evaluation method which are integrated with an obstacle avoidance path planning method based on an optimized artificial potential field and a trajectory tracking method based on model predictive control, and comprise the following steps: 1) under the common constraint of a plurality of potential fields such as an obstacle repulsive force field, a road boundary repulsive force field and the like, a real and effective obstacle avoidance reference path is planned; on the basis of a traditional artificial potential field, the distance between a vehicle and a target point is introduced to optimize an obstacle repulsive force field function, a road boundary repulsive force field is added, and the influence of the vehicle speed is considered; 2) sending the obstacle avoidance reference path to a track tracking control algorithm, outputting the control quantity of the turning angle and the vehicle speed of the front wheel of the vehicle, and controlling the running track of the vehicle; 3) and providing a weighted sum of root mean square values of the transverse tracking error, the yaw angle, the front wheel rotation angle and the front wheel rotation angle increment as a track tracking comprehensive evaluation index to obtain a change rule of the track tracking comprehensive evaluation index under the full-speed working condition. Compared with the prior art, the method has the advantages of high-efficiency obstacle avoidance of the unmanned automobile under the working condition of full speed, comprehensive evaluation of vehicle track tracking precision, driving safety, control performance of the controller and the like.
Description
Technical Field
The invention relates to an obstacle avoidance control method and a performance evaluation method for an unmanned vehicle, in particular to an obstacle avoidance control method for the full-speed working condition of the unmanned vehicle, which integrates an obstacle avoidance path planning method based on an optimized artificial potential field and a path tracking method based on model prediction control, realizes effective obstacle avoidance path planning and accurate track tracking control, and ensures the safe running of the unmanned vehicle under the full-speed working condition; the method defines a track tracking comprehensive evaluation index E, more comprehensively evaluates the track tracking precision, the driving safety and the control performance of a controller of the unmanned automobile in the track tracking process, and better provides a constructive suggestion for effective and safe obstacle avoidance of the automobile.
Background
Unmanned technology has developed rapidly since the 90 s of the 20 th century. The unmanned automobile senses the driving environment through sensors such as a radar and a camera, is controlled by an intelligent controller, avoids obstacles in the autonomous driving process, and guarantees the safety of vehicles and pedestrians. The vehicle path planning and decision-making module receives environment and vehicle state information in real time, sends control instructions to a steer-by-wire system, a brake-by-wire system, a drive-by-wire system and the like of the drive-by-wire chassis, controls the driving speed and direction in real time, and plays a vital role in guaranteeing the safety of vehicles. Therefore, in recent years, path planning and trajectory tracking control have gradually become a hot problem in the field of unmanned automobile research, and scholars at home and abroad have gained many achievements in the field.
(1) Relevant research for unmanned vehicle path planning
The method for planning the unmanned vehicle path mainly comprises a Dijkstra algorithm, an ant colony algorithm, a dynamic planning and the like, the algorithm principle is that a path with the shortest distance is planned according to a scene with known nodes, the path planning is greatly influenced by node arrangement, and the requirement of the unmanned vehicle for planning the path in a real scene is difficult to meet. Compared with the algorithm, an Artificial Potential Field (APF) method guides the vehicle to move by establishing a comprehensive Potential Field such as a virtual target gravitational Potential Field and an obstacle repulsive Potential Field, the type, the strength and the influence factors of the Artificial Potential Field can be flexibly adjusted, and the path planning requirement of the unmanned vehicle in an actual lane is met; most artificial potential field methods can be optimized by modern intelligent algorithms.
(2) Relevant research of unmanned automobile trajectory tracking
The unmanned vehicle trajectory tracking Control method mainly comprises a pure tracking method, a Stanley method, a backstepping method, a sliding mode variable structure Control method, a Model Predictive Control (MPC) method and the like. The pure tracking method and the Stanley method are based on geometric tracking, the corner of the front wheel of the vehicle is deduced by applying the Ackerman steering principle based on a vehicle kinematic model, and the real-time performance of system control is better; however, the constraint conditions of the control variables such as the front wheel turning angle, the yaw angle, and the vehicle turning radius when the vehicle is actually traveling are not taken into consideration, and the tracking path may not be realized for the actual vehicle. Control methods such as a backstepping method, a sliding mode variable structure control method and the like are widely applied, and the control principle is to adjust the vehicle speed and the yaw angular speed in real time, so that the vehicle pose error gradually tends to zero, and the aim of tracking the track is fulfilled. The track tracking control effect is greatly influenced by the unknown parameters of the control law, and the optimal parameter value combination needs to be obtained through repeated experiments. The model predictive control method introduces various vehicle motion constraint conditions, and the trajectory tracking process is closer to the actual situation, so that the method is widely applied. For example, Chinese patent "an unmanned vehicle obstacle avoidance method based on opportunistic constraint model predictive control" (patent number: CN 107357168A) discloses an unmanned vehicle obstacle avoidance method based on opportunistic constraint model predictive control, which comprises unmanned vehicle dynamics model modeling, collision-free condition design, collision-free opportunistic constraint design, cost function design, model predictive control optimization problem design and solution, and the like; the unmanned vehicle obstacle avoidance function is realized, and the unmanned vehicle obstacle avoidance system has the advantages of good environmental adaptability, consideration of the actual occupied area of the vehicle and the like.
(3) Relevant research of unmanned automobile obstacle avoidance performance evaluation method
At present, an obstacle avoidance performance evaluation method of an unmanned automobile is relatively single. For example, chinese patent "an unmanned vehicle trajectory tracking capability evaluation method" (patent No. CN 108108885 a) discloses an unmanned vehicle trajectory tracking capability evaluation method, which includes two parts, namely, a vehicle stability evaluation index and a trajectory tracking accuracy index, and by evaluating the tracking capability of the operation results of an unmanned vehicle control system using different control strategies under the conditions of a planned trajectory and an actual operation trajectory, guides the design and parameter adjustment of the unmanned vehicle control system, and evaluates the unmanned vehicle trajectory tracking accuracy and stability.
In summary, through document retrieval, investigation and analysis, the obstacle avoidance control method and the performance evaluation method for the unmanned vehicle have the following disadvantages:
(1) a path planning method based on an optimized artificial potential field and a path tracking method based on model predictive control cannot be deeply fused; the track planning, track tracking control method and control effect research under the full-speed working condition of the unmanned automobile are lacked, and the controller cannot be ensured to adapt to the full-speed working condition.
(2) The evaluation index system of the obstacle avoidance control performance of the unmanned automobile is single, and the control effect and stability of the trajectory tracking precision and the driving safety controller cannot be comprehensively reflected.
Disclosure of Invention
The invention mainly solves the technical problem of providing an unmanned vehicle all-working-condition obstacle avoidance control method and a performance evaluation method which are integrated with an obstacle avoidance path planning method based on an optimized artificial potential field and a path tracking method based on model predictive control aiming at the defects in the prior art.
In order to solve the technical problems, the invention adopts the following technical scheme:
an unmanned vehicle all-condition obstacle avoidance control method and a performance evaluation method which are integrated with an obstacle avoidance path planning method based on an optimized artificial potential field and a path tracking method based on model predictive control comprise the following steps:
1) under the common constraint of a plurality of potential fields such as an obstacle repulsive force field and a road boundary repulsive force field, a real and effective obstacle avoidance reference path is planned. On the basis of a traditional artificial potential field, the distance between a vehicle and a target point is introduced to optimize an obstacle repulsive force field function, a road boundary repulsive force field is added, and the influence of the vehicle speed is considered.
2) And sending the obstacle avoidance reference path to a track tracking control algorithm, outputting the control quantity of the front wheel rotation angle and the vehicle speed of the vehicle, and controlling the running track of the vehicle.
3) And taking the weighted sum of the root mean square values of the transverse tracking error, the yaw angle, the front wheel rotation angle and the front wheel rotation angle increment of the unmanned automobile as a track tracking comprehensive evaluation index to obtain the change rule of the unmanned automobile under the full-speed working condition, and comprehensively evaluating the track tracking precision, the driving safety and the control performance of the controller.
In the step 1), the gravitation force applied to the unmanned automobile is related to the distance between the target position and the automobile, and the direction of the gravitation force points to the target point. The greater the distance, the greater the attraction. Therefore, the gravitational field function is designed as follows:
in the formula: u shapeattIs a gravitational field function; eta is a target gravitational field gain coefficient; ρ (q, q)g) The distance of the vehicle from the target point.
And (3) carrying out derivation on the negative gradient of the gravitational field function to obtain the gravitational function:
in the formula: fattIs a function of gravity.
Introducing the distance between the vehicle and the target point on the basis of the traditional artificial potential field
Optimizing the obstacle repulsion field function (as shown in equation (3)) solves the problems of local optimality and unreachable target:
in the formula:
is the distance, U, between the vehicle and the target pointreqM is the gain factor of the repulsive force field of the obstacle, ρ (q, q), as a function of the repulsive force field0) Is the distance, rho, between the vehicle and the obstacle0Is an obstacleThe action distance of the repulsive force field, n is a constant.
As can be seen from equation (3), as the vehicle approaches the target point, the repulsive force applied to the vehicle decreases with the attractive force, and the attractive force and the repulsive force applied to the vehicle when the vehicle reaches the target point decrease to zero at the same time. Therefore, the problems of local optimization of the vehicle and unreachable target are solved. When the vehicle does not reach the target point, the repulsive force is:
in the formula: freq1And Freq2As a component of the repulsive force of the road boundary,
as shown in FIG. 2, Freq1In a direction from the obstacle to the vehicle, Freq2Is directed from the vehicle to the target point.
As shown in fig. 3, a road boundary repulsive force field is established to restrict a traveling region of the vehicle. The road boundary repulsive force field has a repulsive effect on vehicles on the road, so that the vehicles keep running on the center line of the lane, and the influence of the vehicle speed on the repulsive force field is properly considered. The road boundary repulsion function established is:
in the formula: frep,edgeIs road boundary repulsion; etaedgeIs road boundary repulsion gain factor; v is the vehicle speed; y is the ordinate of the vehicle in the coordinate system; d is the road width; w is the vehicle width.
If a vehicle runs on a driving lane, when the vehicle runs close to the boundary lines (i) and (iii) of the driving lane, the vehicle may collide with roadside guardrails, so that the boundary lines (i) and (iii) of the driving lane are required to repel the vehicle for ensuring the safety of the vehicleThe force increases dramatically. When in use
And
an exponential function is introduced in the repulsive force function. When the vehicle runs close to the boundary lines (i) and (iii) of the lane, the repulsive force borne by the vehicle exponentially increases, so that the vehicle is restrained from running close to the center line of the lane.
When the vehicle runs close to the lane boundary line (c), the vehicle may deviate from the running direction or the vehicle is about to change to a passing lane, so the repulsive force generated by the lane boundary line (c) should be set to be small. When in use
And
and then, introducing a quadratic function to dynamically adjust the magnitude of the repulsive force.
The higher the vehicle speed, the shorter the reaction time when the vehicle approaches the road boundary. Therefore, in order to secure driving safety, the vehicle speed is introduced into the road boundary repulsive force function. The higher the vehicle speed, the greater the repulsion force generated by the road boundary to the vehicle, so that the vehicle always runs close to the center line of the lane.
In the step 2), the controlled system, the model predictive controller and the state estimator form a complete model predictive control system, as shown in fig. 3. The model predictive controller is designed by combining constraint conditions, a predictive model and an objective function on the basis of three basic elements of a predictive model, rolling optimization and feedback correction of a predictive control theory. Combining with the constraint condition, the controller continuously solves the objective function, and the control variable sequence u is obtained by calculation*After (t), its first value will be applied to the controlled system, which performs the control. And the state estimator calculates the state quantity of the system and feeds the state quantity back to the controller, and the prediction model is continuously updated.
The state quantity deviation and control quantity deviation equation of the vehicle is set as follows:
in the formula:
is a state quantity deviation vector of the vehicle;
a control quantity deviation vector of the vehicle; x and y are respectively the horizontal and vertical coordinates of the vehicle in the inertial coordinate system;
is the vehicle yaw angle; deltafIs a front wheel corner; the index r indicates the reference value for the physical quantity.
Setting the prediction time domain to NpControl time domain as NcThe output vector at the time k is eta (k), the control increment vector at the time k is delta u (k), and the system output equation is obtained as follows:
Y＝Ψξ(k)+ΘΔU (7)
in the formula: Ψ and Θ are both coefficient matrices, Y ═ η (k +1), η (k +2),.., η (k + N)p)T，ΔU＝[Δu(k),Δu(k+1),...,Δu(k+Nc-1)]。
From equation (7), if the state quantity and the control time domain N at the current time are knowncThe control increment in the prediction module can predict the future prediction time domain NpThe internal output.
The problem of tracking the obstacle avoidance track of the unmanned vehicle is summarized as solving the following optimization problems:
in the formula: y ishcIs a hard constraint output; y isscIs the soft constrained output; y ishc,minAnd yhc,maxIs a hard constraint limit; y issc,minAnd ysc,maxIs a soft constraint limit; a is a coefficient matrix; epsilon is a relaxation factor; ρ, Q, R are both weighting coefficients. Soft constraints are set to ensure that a feasible solution can be obtained by solving in each control step, so that the range of the output quantity is properly amplified.
In the step 3), an obstacle avoidance reference path of the vehicle is drawn by using an optimized artificial potential field rule based on multiple constraints, then the planned obstacle avoidance reference path is tracked by adopting a trajectory tracking control method based on model predictive control under the working condition of full vehicle speed, and MATLAB and CarSim software are used for carrying out joint calculation under the full working condition. Taking the full-vehicle-speed working condition as an example for explanation, the parameters of the vehicle, such as the lateral error, the yaw angle, the front wheel rotation angle, and the like, and the maximum value and the root mean square value of the parameters are obtained along with the change of the vehicle speed. The maximum values of the parameters represent the track tracking performance of the vehicle when the vehicle passes through sharp bends at 30m, 47m and 81m, and the root mean square value represents the track tracking performance of the vehicle in the whole running process, so that the control effect is comprehensively reflected.
In order to more comprehensively evaluate the comprehensive performances of the vehicle such as tracking precision, driving safety, controller stability and the like in the process of obstacle avoidance track tracking and provide constructive suggestions for effective and safe obstacle avoidance of the vehicle, a track tracking comprehensive evaluation index E is defined as follows:
in the formula:
representing the track tracking precision of the unmanned automobile as the root mean square value of the transverse tracking error;
the root mean square value of the yaw angle represents the driving safety of the unmanned automobile;
the root mean square value of the corner of the front wheel represents the control performance of the unmanned vehicle trajectory tracking controller;
the control performance of the unmanned vehicle trajectory tracking controller is represented by the root mean square value of the front wheel steering angle increment.
Compared with the prior art, the invention has the following advantages:
(1) on the basis of a traditional manual potential field method, the obstacle repulsive force field and the road boundary repulsive force field are optimized, and a more real and effective obstacle avoidance path is planned;
(2) the invention provides a track tracking comprehensive evaluation method which takes the weighted sum of the root mean square values of the transverse tracking error, the yaw angle, the front wheel rotation angle and the front wheel rotation angle increment as a track tracking comprehensive evaluation index, obtains the change rule of the track tracking comprehensive evaluation index under the working condition of the full vehicle speed, comprehensively evaluates the track tracking effect of the vehicle, the driving safety and the control performance of a controller, and selects a proper critical vehicle speed for the effective and safe obstacle avoidance control of the unmanned vehicle.
Drawings
FIG. 1 is a flow chart of an unmanned vehicle full-condition obstacle avoidance control method and a performance evaluation method;
FIG. 2 is a stress situation of an unmanned vehicle in an artificial potential field;
FIG. 3 is a schematic diagram of a road boundary repulsive potential field for an unmanned vehicle obstacle avoidance path planning;
FIG. 4 is a model predictive control principle for unmanned vehicle trajectory tracking;
FIG. 5 illustrates the path planning and tracking effect of an unmanned vehicle;
FIG. 6 is a variation law of lateral error of the unmanned vehicle under three vehicle speed conditions;
FIG. 7 is a variation law of a lateral error of the unmanned vehicle under a full vehicle speed condition;
FIG. 8 is a change rule of a yaw angle of the unmanned vehicle under three vehicle speed conditions;
FIG. 9 is a diagram illustrating the variation law of the yaw angle of the unmanned vehicle under the full-speed condition;
FIG. 10 is a change rule of a front wheel corner of an unmanned vehicle under three vehicle speed conditions;
FIG. 11 is a variation law of a front wheel rotation angle of an unmanned vehicle under a full vehicle speed condition;
FIG. 12 is a graph showing a change in control amount vehicle speed when the target vehicle speed of the unmanned vehicle is 10 m/s;
FIG. 13 is a graph showing a change in control amount vehicle speed when the target vehicle speed of the unmanned vehicle is 20 m/s;
FIG. 14 is a graph showing a change in control amount vehicle speed when the target vehicle speed of the unmanned vehicle is 30 m/s;
FIG. 15 is a change rule of the front wheel steering angle increment of the unmanned vehicle under three vehicle speed conditions;
FIG. 16 is a variation law of the front wheel steering angle increment of the unmanned vehicle under the full vehicle speed condition;
FIG. 17 is a graph of the change in speed increment for an unmanned vehicle;
FIG. 18 is a diagram showing a change rule of a track following comprehensive evaluation index of an unmanned vehicle under a full vehicle speed condition;
Detailed Description
The following further describes preferred embodiments of the present invention in conjunction with the accompanying drawings so that the advantages and features of the present invention can be more readily understood by those skilled in the art, and the scope of the present invention is more clearly and clearly defined.
Fig. 1 is a flow chart of an unmanned vehicle all-condition obstacle avoidance control method and a performance evaluation method.
Step 1) planning a real and effective obstacle avoidance reference path under the common constraint of a plurality of potential fields such as an obstacle repulsive force field, a road boundary repulsive force field and the like; on the basis of a traditional artificial potential field, introducing the distance between a vehicle and a target point, optimizing a function of a repulsive force field of an obstacle, increasing a repulsive force potential field of a road boundary and considering the influence of the speed of the vehicle;
step 2) sending the obstacle avoidance reference path to a track tracking control algorithm, outputting the control quantity of the front wheel rotation angle and the vehicle speed of the vehicle, and controlling the vehicle running track;
and 3) providing a weighted sum of root mean square values of the transverse tracking error, the yaw angle, the front wheel rotating angle and the front wheel rotating angle increment as a track tracking comprehensive evaluation index, obtaining a change rule of the track tracking comprehensive evaluation index under the full-speed working condition, and comprehensively evaluating the track tracking precision of the vehicle, the driving safety and the control performance of the controller.
Fig. 2 shows the stress condition of the unmanned vehicle in the artificial potential field. The unmanned vehicle is simultaneously subjected to a repulsive field generated by the obstacle and a gravitational field generated by the target point.
Fig. 3 is a schematic diagram of a road boundary repulsive potential field of an unmanned vehicle obstacle avoidance path plan. The road boundary repulsion function established is:
in the formula: frep,edgeIs road boundary repulsion; etaedgeIs road boundary repulsion gain factor; v is the vehicle speed; y is the ordinate of the vehicle in the coordinate system; d is the road width; w is the vehicle width.
Assuming that a vehicle travels on a traveling lane, when the vehicle travels close to the lane boundary lines (i) and (iii), since the vehicle may collide against roadside guardrails, it is necessary that the repulsive force generated to the vehicle by the lane boundary lines (i) and (iii) sharply increases in order to secure the safety of the vehicle. When in use
And
an exponential function is introduced in the repulsive force function. When the vehicle runs close to the boundary lines (i) and (iii) of the lane, the repulsive force borne by the vehicle is exponentially increased, so that the vehicle is restrained from running close to the center line of the lane.
When the vehicle runs close to the lane boundary line (c), the running direction of the vehicle may deviate, or the vehicle may change to a passing lane, so the repulsive force generated by the lane boundary line (c) should be set to be small. When in use
And
and then, introducing a quadratic function to dynamically adjust the magnitude of the repulsive force.
The higher the vehicle speed, the shorter the reaction time will be when the vehicle approaches the road boundary. Therefore, in order to secure driving safety, the vehicle speed is introduced to the road boundary repulsive force function. The higher the vehicle speed, the greater the repulsion force generated by the road boundary to the vehicle, so that the vehicle always runs close to the center line of the lane.
Fig. 4 illustrates a model predictive control principle for unmanned vehicle trajectory tracking. The controlled system, the model predictive controller and the state estimator constitute a complete model predictive control system. The model predictive controller is designed by combining constraint conditions, a prediction model and an objective function on the basis of three basic elements of a prediction model, rolling optimization and feedback correction of a predictive control theory. And (4) combining the constraint conditions, continuously solving the objective function by the controller, calculating to obtain a control variable sequence, applying the first value to the controlled system, and executing control by the system. And the state estimator calculates the state quantity of the system and feeds the state quantity back to the controller, and the prediction model is continuously updated.
Fig. 5 shows the path planning and tracking effect of the unmanned vehicle. When the vehicle runs at the speed of 10m/s, 20m/s and 30m/s, an obstacle avoidance path can be successfully planned from a starting point, various obstacles on a road are avoided, and the planned obstacle avoidance reference path can be tracked. Even if a certain deviation exists when the vehicle tracks the path, the obstacle avoidance function is successfully realized. In the experiment, a planning result can be more conservative by adjusting an obstacle repulsion field gain coefficient m in the formula (3), and the path tracking effect can also be adjusted by adjusting parameters such as Q and R matrix values and calculation step length of a model prediction controller in the formula (8).
FIG. 6 shows the variation law of the lateral error of the unmanned vehicle under three vehicle speed conditions. In places close to obstacles, such as 30m, 47m and 81m, the curvature of the planned obstacle avoidance reference path suddenly increases, and the lateral error of vehicle tracking fluctuates greatly, and the reason is as follows: the normal running path of the vehicle needs to meet the requirement of continuous change of curvature, the turning angle of the vehicle is limited, the turning radius cannot be infinitely small, and the limitation is needed; and, further, to a front wheel steering angle constraint range.
FIG. 7 is a variation rule of the lateral error of the unmanned vehicle under the full speed condition. When the vehicle speed is lower than 22m/s, the maximum value of the lateral error is relatively low and is not higher than 0.2m, which shows that the vehicle has good track tracking performance when passing through sharp bends at positions of 30m, 47m and 81 m. However, when the vehicle speed is higher than 22m/s, the maximum value of the lateral error increases rapidly, the root mean square value of the lateral error rises sharply, which means that the lateral error of the vehicle starts to rise rapidly when the vehicle travels through a sharp curve, and the track following performance of the vehicle during the whole travel process also drops sharply. From the viewpoint of lateral error, the vehicle speed is preferably not more than 22m/s in order to achieve good tracking effect.
Fig. 8 is a change rule of the yaw angle of the unmanned vehicle under three vehicle speed conditions. In places close to obstacles, such as 30m, 47m and 81m, the curvature of the planned path is suddenly increased, the yaw angle tracked by the vehicle is greatly fluctuated, and the reason is analyzed: the normal running path of the vehicle needs to meet the continuous change of curvature, the turning angle of the vehicle is limited, the turning radius cannot be infinitely small, and the limitation is needed; and, further, to a front wheel steering angle constraint range.
Fig. 9 is a change law of the yaw angle of the unmanned vehicle under the full vehicle speed condition. When the vehicle speed exceeds 22m/s, the root mean square value of the yaw angle is increased sharply, which indicates that the vehicle direction is changed sharply many times in the track tracking process of the vehicle and is extremely dangerous under a high-speed working condition, so that the vehicle speed is prevented from exceeding the critical vehicle speed of 22m/s from the perspective of vehicle safety.
Fig. 10 is a change rule of the front wheel rotation angle of the unmanned vehicle under three vehicle speed conditions. The front wheel steering angle of the vehicle is limited between the constraint ranges of-0.54 rad and 0.332rad, but the front wheel steering angle slightly exceeds the constraint range at sharp bends of 30m, 60m, 90m and the like, which shows that the control effect of the trajectory tracking controller on the front wheel steering angle is not good on a straight road surface at the sharp bends with large curvature.
Fig. 11 is a change law of the rotation angle of the front wheel of the unmanned vehicle under the full vehicle speed condition. When the vehicle speed is lower than 25m/s, the maximum value of the front wheel rotation angle is limited in the constraint range. When the vehicle speed is more than 25m/s, the maximum value of the front wheel steering angle exceeds the constraint range of the front wheel steering angle, and the root mean square value of the front wheel steering angle gradually rises, which shows that the control effect of the controller on the front wheel steering angle is gradually weakened. Under high-speed working conditions, the speed control device is extremely dangerous, so that the controller is required to have strong control capability on the speed of the vehicle.
FIG. 12 is a graph showing the change in control amount vehicle speed when the target vehicle speed of the unmanned vehicle is 10 m/s. The vehicle speed is limited within the constraint range of [9.8m/s,10.2m/s ], the controller has good control capability on the vehicle speed, the vehicle speed can be effectively limited within the set constraint range under the working condition of the full vehicle speed, and guarantee is provided for the vehicle track tracking performance and the driving safety.
FIG. 13 shows the change in the control amount vehicle speed when the target vehicle speed of the unmanned vehicle is 20 m/s. The vehicle speed is limited within the restriction range [19.8m/s,20.2m/s ], the controller has good control capability on the vehicle speed, the vehicle speed can be effectively limited within the set restriction range under the working condition of the full vehicle speed, and the guarantee is provided for the track tracking performance and the driving safety of the vehicle.
FIG. 14 shows the change in the control amount vehicle speed when the target vehicle speed of the unmanned vehicle is 30 m/s. The vehicle speed is limited within a constraint range [29.8m/s,30.2m/s ], the controller has good control capability on the vehicle speed, the vehicle speed can be effectively limited within the set constraint range under the working condition of the full vehicle speed, and the guarantee is provided for the track tracking performance and the driving safety of the vehicle.
FIG. 15 is a change rule of the front wheel steering angle increment of the unmanned vehicle under three vehicle speed conditions.
Fig. 16 is a change rule of the front wheel steering angle increment of the unmanned vehicle under the full vehicle speed working condition. The nose wheel corner increments are limited to within a constrained range of [ -0.64rad,0.64rad ].
Fig. 17 shows a change law of the vehicle speed increment of the unmanned vehicle. The vehicle speed increase is limited within a constraint range of-0.05 m/s,0.05 m/s.
Fig. 18 is a change rule of a track following comprehensive evaluation index of the unmanned vehicle under a full vehicle speed condition. Along with the increase of the vehicle speed, the value of the track tracking comprehensive evaluation index slowly increases. When the vehicle speed exceeds the critical vehicle speed of 22m/s, the value of the comprehensive evaluation index of the track tracking rises sharply, which shows that the track tracking precision, the driving safety and the control performance of the controller of the vehicle all decline sharply; therefore, a reasonable critical vehicle speed should be set in the obstacle avoidance process.
Claims (5)
1. An unmanned vehicle all-condition obstacle avoidance control method and a performance evaluation method which are integrated with an obstacle avoidance path planning method based on an optimized artificial potential field and a trajectory tracking method based on model predictive control are characterized by comprising the following steps:
step 1) planning a real and effective obstacle avoidance reference path under the common constraint of a plurality of potential fields such as an obstacle repulsive force field, a road boundary repulsive force field and the like; on the basis of a traditional artificial potential field, introducing the distance between a vehicle and a target point, optimizing a function of a repulsive force field of an obstacle, increasing a repulsive force potential field of a road boundary and considering the influence of the speed of the vehicle;
step 2) sending the obstacle avoidance reference path to a track tracking control algorithm, outputting the control quantity of the front wheel rotation angle and the vehicle speed of the vehicle, and controlling the vehicle running track;
and 3) providing a weighted sum of root mean square values of the transverse tracking error, the yaw angle, the front wheel rotating angle and the front wheel rotating angle increment as a track tracking comprehensive evaluation index, obtaining a change rule of the track tracking comprehensive evaluation index under the full-speed working condition, and comprehensively evaluating the track tracking precision of the vehicle, the driving safety and the control performance of the controller.
2. The method for controlling obstacle avoidance of the unmanned vehicle based on the fusion optimization of the artificial potential field and the model predictive control according to claim 1, wherein in the step 1), a distance between the vehicle and a target point is introduced to optimize a function of a repulsive force field of an obstacle on the basis of a traditional artificial potential field:
introducing the distance between the vehicle and a target point on the basis of the traditional artificial potential field
Optimizing the obstacle repulsion field function (as shown in formula (1)) solves the problems of local optimization and unreachable target:
in the formula:
is the distance, U, between the vehicle and the target pointreqM is the gain factor of the repulsive force field of the obstacle, ρ (q, q), as a function of the repulsive force field0) Is the distance, rho, between the vehicle and the obstacle0Is the acting distance of the repulsive force field of the barrier, and n is a constant;
according to the formula (1), as the vehicle approaches the target point, the repulsive force applied to the vehicle is reduced along with the attractive force, and the attractive force and the repulsive force applied to the vehicle when the vehicle reaches the target point are reduced to zero at the same time; therefore, the problems that the vehicle is locally optimal and the target is inaccessible are solved; when the vehicle does not reach the target point, the repulsive force is:
in the formula: freq1And Freq2As a component of the repulsive force of the road boundary,
as shown in FIG. 2, Freq1In a direction from the obstacle to the vehicle, Freq2Is directed from the vehicle to the target point.
3. The method for controlling obstacle avoidance of the unmanned vehicle based on the combination of the optimized artificial potential field and the model predictive control according to claim 1, wherein in the step 1), a road boundary repulsive potential field is added, and the influence of the vehicle speed on the repulsive potential field is considered:
the road boundary repulsion function established is:
in the formula: frep,edgeIs road boundary repulsion; etaedgeIs road boundary repulsion gain factor; v is the vehicle speed; y is the ordinate of the vehicle in the coordinate system; d is the road width; w is the vehicle width;
assuming that a vehicle runs on a running lane, when the vehicle runs close to the boundary lines (i) and (iii) of the running lane, the vehicle runs close to roadside guardrails, so that the repulsive force generated by the boundary lines (i) and (iii) of the running lane to the vehicle is required to be increased sharply to ensure the safety of the vehicle; when in use
And
introducing an exponential function into the repulsive force function; when the vehicle runs close to the boundary lines of the lane, the repulsive force borne by the vehicle is exponentially increased, so that the vehicle is restrained to run close to the center line of the lane;
when the vehicle runs close to the lane boundary line II, the running direction of the vehicle may deviate, and the vehicle may be about to change to overtaking lane for running, so the repulsion generated by the lane boundary line II is set to be smaller; when in use
And
while introducing a quadratic function to dynamically adjust repulsionThe magnitude of the force;
if the vehicle speed is higher, the reaction time is shorter when the vehicle approaches the road boundary; therefore, in order to guarantee driving safety, the vehicle speed is introduced into a road boundary repulsion function; the higher the vehicle speed is, the larger the repulsion force generated by the road boundary to the vehicle is, and the vehicle is ensured to always run close to the center line of the lane.
4. The method for controlling obstacle avoidance of the unmanned vehicle based on the fusion of the optimized artificial potential field and the model predictive control according to claim 1, wherein in the step 2), a model predictive track tracking controller is established to control a vehicle driving track:
the problem of vehicle obstacle avoidance trajectory tracking is summarized as solving the following optimization problem:
in the formula: y ishcIs a hard constraint output; y isscIs the soft constrained output; y ishc,minAnd yhc,maxIs a hard constraint limit; y issc,minAnd ysc,maxIs a soft constraint limit; a is a coefficient matrix; epsilon is a relaxation factor; rho and Q, R are both weight coefficients; soft constraints are set to ensure that a feasible solution can be obtained by solving in each control step, and the output quantity range is properly amplified.
5. The method for controlling obstacle avoidance of the unmanned vehicle under the full operating conditions by fusing and optimizing the artificial potential field and the model predictive control according to claim 1, wherein in the step 3), the weighted sum of the root mean square values of the transverse tracking error, the yaw angle, the front wheel rotation angle and the front wheel rotation angle increment is used as a track tracking comprehensive evaluation index to obtain the change rule of the unmanned vehicle under the full vehicle speed operating conditions, comprehensively evaluate the vehicle track tracking accuracy, the driving safety and the control performance of a controller, and select a proper critical vehicle speed for the effective and safe obstacle avoidance control of the unmanned vehicle:
defining a track tracking comprehensive evaluation index E as follows:
in the formula:
representing the track tracking precision of the unmanned automobile as the root mean square value of the transverse tracking error;
the root mean square value of the yaw angle represents the driving safety of the unmanned automobile;
the root mean square value of the corner of the front wheel represents the control performance of the unmanned vehicle trajectory tracking controller;
the control performance of the unmanned vehicle trajectory tracking controller is represented by the root mean square value of the front wheel steering angle increment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110754791.7A CN113515125A (en) | 2021-07-05 | 2021-07-05 | Unmanned vehicle full-working-condition obstacle avoidance control method and performance evaluation method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110754791.7A CN113515125A (en) | 2021-07-05 | 2021-07-05 | Unmanned vehicle full-working-condition obstacle avoidance control method and performance evaluation method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113515125A true CN113515125A (en) | 2021-10-19 |
Family
ID=78066417
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110754791.7A Pending CN113515125A (en) | 2021-07-05 | 2021-07-05 | Unmanned vehicle full-working-condition obstacle avoidance control method and performance evaluation method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113515125A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114442637A (en) * | 2022-02-10 | 2022-05-06 | 北京理工大学 | Unmanned vehicle local dynamic obstacle avoidance path planning method |
| CN114995465A (en) * | 2022-08-02 | 2022-09-02 | 北京理工大学 | Multi-unmanned vehicle motion planning method and system considering vehicle motion capability |
| CN115145275A (en) * | 2022-06-24 | 2022-10-04 | 中国安全生产科学研究院 | Multi-robot formation obstacle avoidance control method based on improved artificial potential field method |
| WO2022228358A1 (en) * | 2021-04-25 | 2022-11-03 | 广州汽车集团股份有限公司 | Autonomous driving obstacle avoidance method and system, and storage medium |
| CN115489548A (en) * | 2022-09-20 | 2022-12-20 | 重庆大学 | Intelligent automobile park road path planning method |
| CN115489548B (en) * | 2022-09-20 | 2024-06-04 | 重庆大学 | Intelligent automobile park road path planning method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100235035A1 (en) * | 2006-03-22 | 2010-09-16 | Nissan Motor Co., Ltd. | Avoidance maneuver calculation device, avoidance control device, vehicle having each device, avoidance maneuver calculating method, and avoidance controlling method |
| CN102591332A (en) * | 2011-01-13 | 2012-07-18 | 同济大学 | Device and method for local path planning of pilotless automobile |
| CN110471408A (en) * | 2019-07-03 | 2019-11-19 | 天津大学 | Automatic driving vehicle paths planning method based on decision process |
| CN111845774A (en) * | 2020-07-20 | 2020-10-30 | 上海大学 | Automatic driving automobile dynamic trajectory planning and tracking method based on transverse and longitudinal coordination |
-
2021
- 2021-07-05 CN CN202110754791.7A patent/CN113515125A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100235035A1 (en) * | 2006-03-22 | 2010-09-16 | Nissan Motor Co., Ltd. | Avoidance maneuver calculation device, avoidance control device, vehicle having each device, avoidance maneuver calculating method, and avoidance controlling method |
| CN102591332A (en) * | 2011-01-13 | 2012-07-18 | 同济大学 | Device and method for local path planning of pilotless automobile |
| CN110471408A (en) * | 2019-07-03 | 2019-11-19 | 天津大学 | Automatic driving vehicle paths planning method based on decision process |
| CN111845774A (en) * | 2020-07-20 | 2020-10-30 | 上海大学 | Automatic driving automobile dynamic trajectory planning and tracking method based on transverse and longitudinal coordination |
Non-Patent Citations (1)
| Title |
|---|
| 黄颖涛: "自动驾驶车辆轨迹跟踪模型预测控制方法研究" * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2022228358A1 (en) * | 2021-04-25 | 2022-11-03 | 广州汽车集团股份有限公司 | Autonomous driving obstacle avoidance method and system, and storage medium |
| CN114442637A (en) * | 2022-02-10 | 2022-05-06 | 北京理工大学 | Unmanned vehicle local dynamic obstacle avoidance path planning method |
| CN114442637B (en) * | 2022-02-10 | 2023-11-10 | 北京理工大学 | Unmanned vehicle local dynamic obstacle avoidance path planning method |
| CN115145275A (en) * | 2022-06-24 | 2022-10-04 | 中国安全生产科学研究院 | Multi-robot formation obstacle avoidance control method based on improved artificial potential field method |
| CN115145275B (en) * | 2022-06-24 | 2024-04-30 | 中国安全生产科学研究院 | Multi-robot formation obstacle avoidance control method based on improved artificial potential field method |
| CN114995465A (en) * | 2022-08-02 | 2022-09-02 | 北京理工大学 | Multi-unmanned vehicle motion planning method and system considering vehicle motion capability |
| CN114995465B (en) * | 2022-08-02 | 2022-11-15 | 北京理工大学 | Multi-unmanned vehicle motion planning method and system considering vehicle motion capability |
| CN115489548A (en) * | 2022-09-20 | 2022-12-20 | 重庆大学 | Intelligent automobile park road path planning method |
| CN115489548B (en) * | 2022-09-20 | 2024-06-04 | 重庆大学 | Intelligent automobile park road path planning method |
| CN115683145B (en) * | 2022-11-03 | 2024-06-11 | 北京踏歌智行科技有限公司 | Automatic driving safety obstacle avoidance method based on track prediction |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111413966B (en) | Progressive model prediction unmanned planning tracking cooperative control method | |
| CN110377039B (en) | Vehicle obstacle avoidance track planning and tracking control method | |
| Xu et al. | Preview path tracking control with delay compensation for autonomous vehicles | |
| CN113515125A (en) | Unmanned vehicle full-working-condition obstacle avoidance control method and performance evaluation method | |
| CN110780594B (en) | Path tracking method and system of intelligent vehicle | |
| Qian et al. | A hierarchical model predictive control framework for on-road formation control of autonomous vehicles | |
| US20200278686A1 (en) | Iterative Feedback Motion Planning | |
| CN108227491B (en) | Intelligent vehicle track tracking control method based on sliding mode neural network | |
| Kloeser et al. | NMPC for racing using a singularity-free path-parametric model with obstacle avoidance | |
| JP7014453B2 (en) | Maintenance control method | |
| CN112965476B (en) | High-speed unmanned vehicle trajectory planning system and method based on multi-window model | |
| Liu et al. | Simulation performance evaluation of pure pursuit, stanley, lqr, mpc controller for autonomous vehicles | |
| CN110262229B (en) | Vehicle self-adaptive path tracking method based on MPC | |
| CN111750866B (en) | Intelligent automobile transverse and longitudinal coupling path planning method based on regional virtual force field | |
| CN112947469A (en) | Automobile track-changing track planning and dynamic track tracking control method | |
| Xue et al. | Active collision avoidance system design based on model predictive control with varying sampling time | |
| CN111538328A (en) | Priority hierarchical prediction control method for obstacle avoidance trajectory planning and tracking control of autonomous driving vehicle | |
| Zakaria et al. | Dynamic curvature steering control for autonomous vehicle: performance analysis | |
| Chen et al. | An adaptive path tracking controller based on reinforcement learning with urban driving application | |
| CN113408062A (en) | Automatic driving full-working-condition road surface self-adaptive MPC (MPC) trajectory tracking control and evaluation method | |
| Xu et al. | Interval trajectory tracking for AGV based on MPC | |
| Obayashi et al. | Real-time autonomous car motion planning using NMPC with approximated problem considering traffic environment | |
| Timings et al. | Efficient minimum manoeuvre time optimisation of an oversteering vehicle at constant forward speed | |
| Kim et al. | A heuristic obstacle avoidance algorithm using vanishing point and obstacle angle | |
| Tao et al. | Path following of autonomous vehicles with an optimized brain emotional learning–based intelligent controller |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20211019 |