CN105109483A - Driving method and driving system - Google Patents

Driving method and driving system Download PDF

Info

Publication number
CN105109483A
CN105109483A CN201510532385.0A CN201510532385A CN105109483A CN 105109483 A CN105109483 A CN 105109483A CN 201510532385 A CN201510532385 A CN 201510532385A CN 105109483 A CN105109483 A CN 105109483A
Authority
CN
China
Prior art keywords
action
decision
obstacle
vehicle
making
Prior art date
Application number
CN201510532385.0A
Other languages
Chinese (zh)
Inventor
方啸
高红博
王继贞
徐达学
尹飞飞
Original Assignee
奇瑞汽车股份有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 奇瑞汽车股份有限公司 filed Critical 奇瑞汽车股份有限公司
Priority to CN201510532385.0A priority Critical patent/CN105109483A/en
Publication of CN105109483A publication Critical patent/CN105109483A/en

Links

Classifications

    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W30/00Purposes of road vehicle drive control systems not related to the control of a particular sub-unit, e.g. of systems using conjoint control of vehicle sub-units, or advanced driver assistance systems for ensuring comfort, stability and safety or drive control systems for propelling or retarding the vehicle
    • B60W30/08Active safety systems predicting or avoiding probable or impending collision or attempting to minimise its consequences
    • B60W30/09Taking automatic action to avoid collision, e.g. braking and steering
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W40/00Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models
    • B60W40/02Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models related to ambient conditions
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W2554/00Input parameters relating to objects

Abstract

The invention discloses a driving method and a driving system, belonging to the technical field of vehicle safety. The driving system comprises an environmental awareness module and a collision avoidance control module. The environmental awareness module is used for monitoring running environments of vehicles in the running process of the vehicles, predicting whether emergencies caused by obstacles would occur in the running environments of the vehicles and obtaining the positions of the obstacles when the emergencies caused by the obstacles would occur in the running environments of the vehicles; the collision avoidance control module is used for determining target decision actions by adopting a machine learning algorithm according to the positions of the obstacles, and controlling the vehicles to run according to the target decision actions. By applying the driving method and the driving system, the problems that the driving system in the related technology has limitations, the stability is lower and the flexibility is poorer are solved, and the beneficial effects of expanding the application range of the driving system and improving the stability and flexibility of the driving system are achieved. The driving method and the driving system are used for collision avoidance driving of vehicles.

Description

Drive manner and system
Technical field
The present invention relates to technical field of vehicle safety, particularly a kind of drive manner and system.
Background technology
Along with the fast development of Eltec, the vehicles such as vehicle have become the requisite vehicle in life.And popularizing along with vehicle, the vehicle on road is more and more intensive, and traffic safety is also more and more important.
Usually, in the process that chaufeur travels at steering vehicle, some may occur and such as insert the accidents such as car, barrier obstruction, now, chaufeur can M/C bearing circle, throttle, brakes etc. are to avoid vehicle and obstacle to collide, but due to when there is accident, chaufeur is usually in tension, chaufeur is to bearing circle, and the accuracy of the operation such as throttle, brake is lower.For this reason, correlation technique provides a kind of control loop, this control loop comprises: environment sensing module and collision avoidance control module, chaufeur is stored according to the ambient condition amount of collision avoidance experience setting in steering vehicle process and the corresponding relation of decision-making action in collision avoidance control module, wherein, ambient condition amount can be the position of obstacle, environment sensing module obtains vehicle current ambient condition amount when can there is the accident caused by obstacle in the running environment of vehicle, collision avoidance control module can according to the current ambient condition amount of vehicle from ambient condition amount with determine the objective decision action that the ambient condition amount current with vehicle is corresponding the corresponding relation of decision-making action, then travel according to this objective decision action control vehicle, thus avoid vehicle and obstacle to collide.
Realizing in process of the present invention, contriver finds that correlation technique at least exists following problem:
Control loop in correlation technique is for travelling according to controlling vehicle with the collision avoidance experience of chaufeur, by the restriction of driver experience, control loop has certain limitation, and due to be with the collision avoidance experience of chaufeur be according to control vehicle travel, therefore, the stability of control loop is lower, and alerting ability is poor.
Summary of the invention
Have limitation to solve control loop in correlation technique, stability is lower, the problem that alerting ability is poor, the invention provides a kind of drive manner and system.Described technical scheme is as follows:
First aspect, provides a kind of control loop, and described control loop comprises: environment sensing module and collision avoidance control module,
Described environment sensing module is used for the running environment of monitoring described vehicle in vehicle travel process; Predict the accident that whether can occur in the running environment of described vehicle to be caused by obstacle; When can there is the accident caused by obstacle in the running environment of described vehicle, obtain the position of described obstacle;
Described collision avoidance control module is used for the position according to described obstacle, adopts the action of machine learning algorithm determination objective decision; According to described objective decision action control, vehicle travels.
Alternatively, described collision avoidance control module is used for:
From experience thesaurus, determine that the target corresponding with the position of described obstacle can perform an action, the position that have recorded the obstacle determined by machine learning algorithm in advance in described experience thesaurus and the corresponding relation that can perform an action, the position of each obstacle corresponding at least one can perform an action, each performing an action comprises decision-making action and strengthens signal one to one with described decision-making action, and described enhancing signal is used to indicate and the immediately return of described enhancing signal decision-making action one to one when performing;
Calculate described target can perform an action in the infinite return cumulative sum in future of each decision-making action;
Using decision-making action maximum for infinite for described future return cumulative sum as described objective decision action.
Alternatively, described collision avoidance control module is for obtaining the decision-making action of chaufeur;
Described environment sensing module for obtain perform described chaufeur decision-making action after the primary importance of described obstacle;
Described collision avoidance control module is used for from experience thesaurus, determine that the target corresponding with the primary importance of described obstacle can perform an action, the position that have recorded the obstacle determined by machine learning algorithm in advance in described experience thesaurus and the corresponding relation that can perform an action, the position of each obstacle corresponding at least one can perform an action, each performing an action comprises decision-making action and strengthens signal one to one with described decision-making action, and described enhancing signal is used to indicate and the immediately return of described enhancing signal decision-making action one to one when performing;
Described collision avoidance control module is for the infinite return cumulative sum in future of each decision-making action in calculating described target and can performing an action; Using decision-making action maximum for infinite for described future return cumulative sum as described objective decision action.
Alternatively, described collision avoidance control module is used for:
Calculate the enhancing signal of following n state corresponding to the decision-making action of described chaufeur, obtain n and strengthen signal, described n be greater than or equal to 1 integer;
Judge whether to exist in described n enhancing signal to meet pre-conditioned enhancing signal;
When existence meets pre-conditioned enhancing signal in described n enhancing signal, trigger the primary importance of described obstacle after described environment sensing module obtains the decision-making action performing described chaufeur;
Wherein, when meeting decision-making action corresponding to pre-conditioned enhancing signal described in performing, the obstacle in the running environment of described vehicle and described vehicle can collide.
Alternatively, described environment sensing module is used for:
Judge whether there is obstacle in the running environment of described vehicle;
When there is obstacle in the running environment of described vehicle, judge whether described obstacle is in preset range;
When described obstacle is in described preset range, determine the accident that can occur in the running environment of described vehicle to be caused by described obstacle.
Second aspect, provides a kind of drive manner, and described method comprises:
The running environment of described vehicle is monitored in vehicle travel process;
Predict the accident that whether can occur in the running environment of described vehicle to be caused by obstacle;
The accident caused by obstacle if can occur in the running environment of described vehicle, then obtain the position of described obstacle;
According to the position of described obstacle, adopt the action of machine learning algorithm determination objective decision;
According to described objective decision action control, vehicle travels.
Alternatively, the described position according to described obstacle, adopts the action of machine learning algorithm determination objective decision, comprising:
From experience thesaurus, determine that the target corresponding with the position of described obstacle can perform an action, the position that have recorded the obstacle determined by machine learning algorithm in advance in described experience thesaurus and the corresponding relation that can perform an action, the position of each obstacle corresponding at least one can perform an action, each performing an action comprises decision-making action and strengthens signal one to one with described decision-making action, and described enhancing signal is used to indicate and the immediately return of described enhancing signal decision-making action one to one when performing;
Calculate described target can perform an action in the infinite return cumulative sum in future of each decision-making action;
Using decision-making action maximum for infinite for described future return cumulative sum as described objective decision action.
Alternatively, the described position according to described obstacle, adopts the action of machine learning algorithm determination objective decision, comprising:
Obtain the decision-making action of chaufeur;
The primary importance of described obstacle after obtaining the decision-making action performing described chaufeur;
From experience thesaurus, determine that the target corresponding with the primary importance of described obstacle can perform an action, the position that have recorded the obstacle determined by machine learning algorithm in advance in described experience thesaurus and the corresponding relation that can perform an action, the position of each obstacle corresponding at least one can perform an action, each performing an action comprises decision-making action and strengthens signal one to one with described decision-making action, and described enhancing signal is used to indicate and the immediately return of described enhancing signal decision-making action one to one when performing;
Calculate described target can perform an action in the infinite return cumulative sum in future of each decision-making action;
Using decision-making action maximum for infinite for described future return cumulative sum as described objective decision action.
Alternatively, the primary importance of described obstacle after the decision-making action of the described chaufeur of described acquisition execution, comprising:
Calculate the enhancing signal of following n state corresponding to the decision-making action of described chaufeur, obtain n and strengthen signal, described n be greater than or equal to 1 integer;
Judge whether to exist in described n enhancing signal to meet pre-conditioned enhancing signal;
If described n strengthens in signal to exist and meets pre-conditioned enhancing signal, then the primary importance of described obstacle after obtaining the decision-making action performing described chaufeur;
Wherein, when meeting decision-making action corresponding to pre-conditioned enhancing signal described in performing, the obstacle in the running environment of described vehicle and described vehicle can collide.
Alternatively, whether can there is the accident caused by obstacle in the running environment of the described vehicle of described prediction, comprise:
Judge whether there is obstacle in the running environment of described vehicle;
If there is obstacle in the running environment of described vehicle, then judge whether described obstacle is in preset range;
If described obstacle is in described preset range, then determine the accident that can occur in the running environment of described vehicle to be caused by described obstacle.
The beneficial effect that technical scheme provided by the invention is brought is:
The drive manner that the embodiment of the present invention provides and system, control loop comprises: environment sensing module and collision avoidance control module, and environment sensing module is used for the running environment of monitor vehicle in vehicle travel process; The accident caused by obstacle whether can be there is in the running environment of prediction vehicle; When can there is the accident caused by obstacle in the running environment of vehicle, obtain the position of obstacle; Collision avoidance control module is used for the position according to obstacle, adopts the action of machine learning algorithm determination objective decision; Travel according to objective decision action control vehicle.Because the present invention adopts the action of machine learning algorithm determination objective decision, the determination of objective decision action is not by the restriction of the experience of chaufeur, the control loop solved in correlation technique has limitation, stability is lower, the problem that alerting ability is poor, reach the field of application expanding control loop, improve the stability of control loop and the beneficial effect of alerting ability.
Should be understood that, it is only exemplary that above general description and details hereinafter describe, and can not limit the present invention.
Accompanying drawing explanation
In order to be illustrated more clearly in the technical scheme in the embodiment of the present invention, below the accompanying drawing used required in describing embodiment is briefly described, apparently, accompanying drawing in the following describes is only some embodiments of the present invention, for those of ordinary skill in the art, under the prerequisite not paying creative work, other accompanying drawing can also be obtained according to these accompanying drawings.
Fig. 1 is the structural representation of a kind of implementation environment involved by drive manner that each embodiment of the present invention provides;
Fig. 2 is the block diagram of the control loop that one embodiment of the invention provides;
Fig. 3 is the method flow diagram of a kind of drive manner that one embodiment of the invention provides;
Fig. 4 is the method flow diagram of a kind of drive manner that another embodiment of the present invention provides;
Fig. 5 be embodiment illustrated in fig. 4 provide a kind ofly predict the method flow diagram whether accident caused by obstacle can occur in the running environment of vehicle;
Fig. 6 be embodiment illustrated in fig. 4 provide a kind ofly determine the schematic diagram whether obstacle is positioned at preset range;
Fig. 7 is the method flow diagram that a kind of position according to obstacle provided embodiment illustrated in fig. 4 adopts the action of machine learning algorithm determination objective decision;
Fig. 8 is the position of a kind of basis provided embodiment illustrated in fig. 4 by machine learning algorithm determination obstacle and the schematic diagram of the corresponding relation that can perform an action;
Fig. 9 is that the another kind provided embodiment illustrated in fig. 4 adopts the method flow diagram of machine learning algorithm determination objective decision action according to the position of obstacle;
Figure 10 is a kind of method flow diagram obtaining the primary importance of the obstruction of the decision-making action performing chaufeur provided embodiment illustrated in fig. 4.
Accompanying drawing to be herein merged in specification sheets and to form the part of this specification sheets, shows embodiment according to the invention, and is used from specification sheets one and explains principle of the present invention.
Embodiment
In order to make the object, technical solutions and advantages of the present invention clearly, below in conjunction with accompanying drawing, the present invention is described in further detail, and obviously, described embodiment is only a part of embodiment of the present invention, instead of whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art, not making other embodiments all obtained under creative work prerequisite, belong to the scope of protection of the invention.
Please refer to Fig. 1, it illustrates the structural representation of a kind of implementation environment involved by drive manner that each embodiment of the present invention provides, see Fig. 1, road S comprises three tracks, be respectively track S1, track S2 and track S3, vehicle C1 travels on the S1 of track, and vehicle C2 travels on the S2 of track, vehicle C3 travels on the S3 of track, and vehicle C1, vehicle C2 are identical with the travel direction of vehicle C3.In the process that vehicle C1 and vehicle C2 travels, vehicle C3 travels from the lane change ahead of vehicle C2 to track S2 suddenly (plugging in car), now, vehicle C3 can be called obstacle for vehicle C2, when this obstacle occurs, the bearing circle of the rotation vehicle C2 of the chaufeur conditioned reflex of vehicle C2 collides to avoid vehicle C2 and vehicle C3, the chaufeur of vehicle C2 rotates the angle and direction difference of the bearing circle of vehicle C2, vehicle C2 can be made according to different route, illustratively, vehicle C2 can according to the path L1 shown in Fig. 1, arbitrary route in path L2 and path L3, known see Fig. 1, the hand of rotation of the bearing circle that this 3 paths is corresponding is all rotate to the left side of the chaufeur of vehicle C2, and the anglec of rotation corresponding to path L2 is less than the anglec of rotation corresponding to path L1, the anglec of rotation that path L1 is corresponding is less than the anglec of rotation corresponding to path L3.
Under normal circumstances, suddenly the plug in car of vehicle C3 can cause the chaufeur of vehicle C2 be in biotonus and accurately cannot hold the anglec of rotation of bearing circle, if the anglec of rotation of the bearing circle of vehicle C2 is too small, vehicle C2 may travel according to path L2, cause vehicle C2 and vehicle C3 to occur to swipe and even collide, and then cause traffic accident; If the anglec of rotation of the bearing circle of vehicle C2 is excessive, vehicle C2 may travel according to path L3, causes vehicle C2 and vehicle C1 to occur to swipe and even collide, and then causes traffic accident.
In this implementation environment, when vehicle C3 plugs in car, the desired ride path of vehicle C2 is path L1, the control loop that the embodiment of the present invention provides can be arranged on vehicle C2, make when vehicle C3 plugs in car, vehicle C2 travels according to path L1, avoids vehicle C2 and vehicle C3, vehicle C1 collides, thus the generation avoided traffic accident.
Please refer to Fig. 2, it illustrates the block diagram of the control loop 200 that one embodiment of the invention provides, this control loop 200 may be used for vehicular drive, when this control loop 200 can exist the accident caused by obstacle in the running environment of vehicle, avoids vehicle and obstacle to collide.See Fig. 2, this control loop 200 can include but not limited to: environment sensing module 210 and collision avoidance control module 220.
Environment sensing module 210 is for the running environment of monitor vehicle in vehicle travel process; The accident caused by obstacle whether can be there is in the running environment of prediction vehicle; When can there is the accident caused by obstacle in the running environment of vehicle, obtain the position of obstacle.
Collision avoidance control module 220, for the position according to obstacle, adopts the action of machine learning algorithm determination objective decision; Travel according to objective decision action control vehicle.
Alternatively, collision avoidance control module 220 for:
From experience thesaurus, determine that the target corresponding with the position of obstacle can perform an action, the position that have recorded the obstacle determined by machine learning algorithm in advance in experience thesaurus and the corresponding relation that can perform an action, the position of each obstacle corresponding at least one can perform an action, each performing an action comprises decision-making action and strengthens signal one to one with decision-making action, strengthens signal and is used to indicate and strengthens the immediately return of signal decision-making action one to one when performing;
The infinite return cumulative sum in future of each decision-making action during calculating target can perform an action;
Using decision-making action maximum for infinite for future return cumulative sum as objective decision action.
Alternatively, collision avoidance control module 220 is for obtaining the decision-making action of chaufeur;
Environment sensing module 210 is for obtaining the primary importance of the obstruction of the decision-making action performing chaufeur;
Collision avoidance control module 220 is for determining that from experience thesaurus the target corresponding with the primary importance of obstacle can perform an action, the position that have recorded the obstacle determined by machine learning algorithm in advance in experience thesaurus and the corresponding relation that can perform an action, the position of each obstacle corresponding at least one can perform an action, each performing an action comprises decision-making action and strengthens signal one to one with decision-making action, strengthens signal and is used to indicate and strengthens the immediately return of signal decision-making action one to one when performing;
Collision avoidance control module 220 is for the infinite return cumulative sum in future of each decision-making action in calculating target and can performing an action; Using decision-making action maximum for infinite for future return cumulative sum as objective decision action.
Alternatively, collision avoidance control module 220 for:
Calculate the enhancing signal of following n state corresponding to the decision-making action of chaufeur, obtain n enhancing signal, n be greater than or equal to 1 integer;
Judge whether to exist in n enhancing signal to meet pre-conditioned enhancing signal;
When existence meets pre-conditioned enhancing signal in n enhancing signal, trigger the primary importance that environment sensing module 210 obtains the obstruction of the decision-making action performing chaufeur;
Wherein, perform meet decision-making action corresponding to pre-conditioned enhancing signal time, the obstacle in the running environment of vehicle and vehicle can collide.
Alternatively, environment sensing module 210 for:
Judge whether there is obstacle in the running environment of vehicle;
When there is obstacle in the running environment of vehicle, whether disturbance in judgement thing is in preset range;
When obstacle is in preset range, determine the accident that can occur in the running environment of vehicle to be caused by obstacle.
In sum, the control loop that the embodiment of the present invention provides comprises: environment sensing module and collision avoidance control module, and environment sensing module is used for the running environment of monitor vehicle in vehicle travel process; The accident caused by obstacle whether can be there is in the running environment of prediction vehicle; When can there is the accident caused by obstacle in the running environment of vehicle, obtain the position of obstacle; Collision avoidance control module is used for the position according to obstacle, adopts the action of machine learning algorithm determination objective decision; Travel according to objective decision action control vehicle.Because the present invention adopts the action of machine learning algorithm determination objective decision, the determination of objective decision action is not by the restriction of the experience of chaufeur, the control loop solved in correlation technique has limitation, stability is lower, the problem that alerting ability is poor, reach the field of application expanding control loop, improve the stability of control loop and the beneficial effect of alerting ability.
The control loop that the embodiment of the present invention provides can be applied to method hereafter, and in the embodiment of the present invention, drive manner can vide infra the description in each embodiment.
Please refer to Fig. 3, it illustrates the method flow diagram of the drive manner that one embodiment of the invention provides, this drive manner can perform by control loop as shown in Figure 2, and see Fig. 3, the method flow process can comprise following several step:
In step 301, the running environment of monitor vehicle in vehicle travel process.
In step 302, the accident that whether can occur in the running environment of vehicle to be caused by obstacle is predicted.
In step 303, the accident caused by obstacle if can occur in the running environment of vehicle, then obtain the position of obstacle.
In step 304, according to the position of obstacle, adopt the action of machine learning algorithm determination objective decision.
In step 305, travel according to objective decision action control vehicle.
In sum, the drive manner that the embodiment of the present invention provides, by the running environment of monitor vehicle in vehicle travel process; The accident caused by obstacle whether can be there is in the running environment of prediction vehicle; When can there is the accident caused by obstacle in the running environment of vehicle, obtain the position of obstacle; According to the position of obstacle, adopt the action of machine learning algorithm determination objective decision; Travel according to objective decision action control vehicle.Because the present invention adopts the action of machine learning algorithm determination objective decision, the determination of objective decision action is not by the restriction of the experience of chaufeur, the drive manner solved in correlation technique has limitation, stability is lower, the problem that alerting ability is poor, reach the field of application expanding drive manner, improve the stability of drive manner and the beneficial effect of alerting ability.
Alternatively, step 304 can comprise:
From experience thesaurus, determine that the target corresponding with the position of obstacle can perform an action, the position that have recorded the obstacle determined by machine learning algorithm in advance in experience thesaurus and the corresponding relation that can perform an action, the position of each obstacle corresponding at least one can perform an action, each performing an action comprises decision-making action and strengthens signal one to one with decision-making action, strengthens signal and is used to indicate and strengthens the immediately return of signal decision-making action one to one when performing;
The infinite return cumulative sum in future of each decision-making action during calculating target can perform an action;
Using decision-making action maximum for infinite for future return cumulative sum as objective decision action.
Alternatively, step 304 can comprise:
Obtain the decision-making action of chaufeur;
Obtain the primary importance of the obstruction of the decision-making action performing chaufeur;
From experience thesaurus, determine that the target corresponding with the primary importance of obstacle can perform an action, the position that have recorded the obstacle determined by machine learning algorithm in advance in experience thesaurus and the corresponding relation that can perform an action, the position of each obstacle corresponding at least one can perform an action, each performing an action comprises decision-making action and strengthens signal one to one with decision-making action, strengthens signal and is used to indicate and strengthens the immediately return of signal decision-making action one to one when performing;
The infinite return cumulative sum in future of each decision-making action during calculating target can perform an action;
Using decision-making action maximum for infinite for future return cumulative sum as objective decision action.
Further, obtain the primary importance of the obstruction of the decision-making action performing chaufeur, comprising:
Calculate the enhancing signal of following n state corresponding to the decision-making action of chaufeur, obtain n enhancing signal, n be greater than or equal to 1 integer;
Judge whether to exist in n enhancing signal to meet pre-conditioned enhancing signal;
If n strengthens in signal to exist and meet pre-conditioned enhancing signal, then obtain the primary importance of the obstruction of the decision-making action performing chaufeur;
Wherein, perform meet decision-making action corresponding to pre-conditioned enhancing signal time, the obstacle in the running environment of vehicle and vehicle can collide.
Alternatively, step 302 can comprise:
Judge whether there is obstacle in the running environment of vehicle;
If there is obstacle in the running environment of vehicle, then whether disturbance in judgement thing is in preset range;
If obstacle is in preset range, then determine the accident that can occur in the running environment of vehicle to be caused by obstacle.
Above-mentioned all alternatives, can adopt and combine arbitrarily formation optional embodiment of the present invention, this is no longer going to repeat them.
In sum, the drive manner that the embodiment of the present invention provides, by the running environment of monitor vehicle in vehicle travel process; The accident caused by obstacle whether can be there is in the running environment of prediction vehicle; When can there is the accident caused by obstacle in the running environment of vehicle, obtain the position of obstacle; According to the position of obstacle, adopt the action of machine learning algorithm determination objective decision; Travel according to objective decision action control vehicle.Because the present invention adopts the action of machine learning algorithm determination objective decision, the determination of objective decision action is not by the restriction of the experience of chaufeur, the drive manner solved in correlation technique has limitation, stability is lower, the problem that alerting ability is poor, reach the field of application expanding drive manner, improve the stability of drive manner and the beneficial effect of alerting ability.
Please refer to Fig. 4, it illustrates the method flow diagram of the drive manner that another embodiment of the present invention provides, this drive manner can perform by control loop as shown in Figure 2, and see Fig. 4, the method flow process can comprise following several step:
In step 401, the running environment of monitor vehicle in vehicle travel process.
Wherein, the process of the running environment of this monitor vehicle can the environment sensing module as shown in Figure 2 in control loop perform, and wherein, environment sensing module can comprise pick up camera, radar etc., and the embodiment of the present invention does not limit this.The running environment of vehicle can comprise the road conditions etc. of vehicle place travel, and the embodiment of the present invention does not limit this.
In step 402, the accident that whether can occur in the running environment of vehicle to be caused by obstacle is predicted.
Wherein, the process whether accident caused by obstacle can occur in the running environment of prediction vehicle can the environment sensing module as shown in Figure 2 in control loop perform, in embodiments of the present invention, environment sensing module, just can according to the accident that whether can occur to be caused by obstacle in the running environment of the running environment prediction vehicle of vehicle in the process of the running environment of monitor vehicle.
Illustratively, please refer to Fig. 5, shown in it be embodiment illustrated in fig. 4 provide a kind ofly predict the method flow diagram whether accident caused by obstacle can occur in the running environment of vehicle.See Fig. 5, in embodiments of the present invention, environment sensing module predicts that the accident that whether can occur to be caused by obstacle in the running environment of vehicle can comprise following several step:
In sub-step 4021, judge whether there is obstacle in the running environment of vehicle.
Environment sensing module can judge whether there is obstacle in the running environment of vehicle according to the monitoring data of pick up camera, radar etc., and wherein, in embodiments of the present invention, that hinders vehicle forward can be called obstacle.Illustratively, as shown in Figure 1, when vehicle C2 normally travels, the plug in car of vehicle C3 hinders vehicle C2 and moves ahead, and therefore, vehicle C3 belongs to obstacle for vehicle C2, when vehicle C2 travels according to path L3, vehicle C1 can hinder vehicle C2 to move ahead, and therefore, vehicle C1 belongs to obstacle for vehicle C2.
In embodiments of the present invention, illustratively, whether there is vehicle C3 or vehicle C1 in the image that the environment sensing module of the control loop on vehicle C2 can be caught according to pick up camera, judge whether there is obstacle in the running environment of vehicle C2.
In sub-step 4022, if there is obstacle in the running environment of vehicle, then whether disturbance in judgement thing is in preset range.
If in step 4021, there is obstacle in the running environment of environment sensing module determination vehicle, then whether environment sensing module disturbance in judgement thing is in preset range, and wherein, preset range can be arranged according to actual conditions, and the embodiment of the present invention does not limit this.Illustratively, preset range can be in the scope of vehicle periphery 5 meters on track, vehicle place, namely, preset range can be the center of circle with vehicle, 5 meters of scopes being the circle of radius and determining, in embodiments of the present invention, preferably, preset range can be the center of circle with vehicle, the scope that 5 meters of semicircles being the circle of radius is positioned at vehicle front are determined, the embodiment of the present invention does not limit this.
Illustratively, as shown in Figure 6, suppose that vehicle C3 is the obstacle in vehicle C2 running environment, the scope that preset range can be determined for the dotted line shade in Fig. 6, as shown in Figure 6, vehicle C3 is in preset range.
In sub-step 4023, if obstacle is in preset range, then determine the accident that can occur in the running environment of vehicle to be caused by obstacle.
If in step 4022, environment sensing module determination obstacle is in preset range, then the accident caused by obstacle can occur in the running environment of environment sensing module determination vehicle.Illustratively, for Fig. 6, because vehicle C3 (obstacle) is in preset range, therefore, the accident caused by obstacle can be there is in the running environment of environment sensing module determination vehicle C2.
In step 403, the accident caused by obstacle if can occur in the running environment of vehicle, then obtain the position of obstacle.
If in step 402, the accident caused by obstacle can occur in the running environment of environment sensing module determination vehicle, then environment sensing module obtains the position of obstacle.
Wherein, the position of obstacle can comprise: the distance of obstacle distance vehicle, and the orientation etc. at obstacle place, the embodiment of the present invention does not limit this.
Illustratively, environment sensing module obtains the position of vehicle C3 (obstacle), and this position can be X (t).It should be noted that, this position also can be understood as the ambient condition amount of vehicle C2, and the embodiment of the present invention does not limit this.
In step 404, according to the position of obstacle, adopt the action of machine learning algorithm determination objective decision.
Wherein, according to the position of obstacle, the action of machine learning algorithm determination objective decision is adopted the collision avoidance control module as shown in Figure 2 in control loop to perform.Environment sensing module can send the position of obstacle to collision avoidance control module, make collision avoidance control module according to the position of obstacle, adopts the action of machine learning algorithm determination objective decision.
In embodiments of the present invention, collision avoidance control module according to any one method shown in lower Fig. 7 or lower Fig. 9 according to the position of obstacle, can adopt the action of machine learning algorithm determination objective decision.
Illustratively, please refer to Fig. 7, shown in it is a kind of position according to obstacle provided embodiment illustrated in fig. 4, and adopt the method flow diagram of machine learning algorithm determination objective decision action, see Fig. 7, the method flow process can comprise following several step:
In sub-step 4041a, from experience thesaurus, determine that the target corresponding with the position of obstacle can perform an action.
Wherein, the position that have recorded the obstacle determined by machine learning algorithm in advance in experience thesaurus and the corresponding relation that can perform an action, the position of each obstacle corresponding at least one can perform an action, each performing an action comprises decision-making action and strengthens signal one to one with decision-making action, strengthens signal and is used to indicate and strengthens the immediately return of signal decision-making action one to one when performing.
Suppose that the position of obstacle represents with X, decision-making action u represents, strengthen signal r to represent, illustratively, the position of the obstacle recorded in experience thesaurus can be as shown in table 1 below with the corresponding relation that can perform an action, wherein, decision-making action can be the dynamics value of step on the accelerator, the dynamics value touched on the brake, the anglec of rotation etc. of bearing circle, the embodiment of the present invention does not limit this.
Table 1
Ginseng is shown in Table 1, and have recorded the position of n obstacle in this table 1, and the position of each obstacle is corresponding multiple to perform an action, and each performing an action comprises decision-making action and strengthen signal one to one with this decision-making action.Illustratively, corresponding the performing an action in position X (t) of obstacle comprises decision-making action u1 (t), u2 (t) and u3 (t) totally 3 decision-making actions, the enhancing signal of decision-making action u1 (t) correspondence is r1 (t), the enhancing signal of decision-making action u2 (t) correspondence is r2 (t), the enhancing signal of decision-making action u3 (t) correspondence is r3 (t), corresponding the performing an action of position X (t+1) of obstacle can be understood with reference to corresponding the performing an action in position X (t) of obstacle, the embodiment of the present invention does not repeat them here.
In embodiments of the present invention, the position of vehicle C3 (obstacle) is X (t), then collision avoidance control module can determine from table 1 that the target that X (t) is corresponding can perform an action, and the target that this X (t) is corresponding can perform an action and comprise decision-making action u1 (t), u2 (t) and u3 (t) totally 3 decision-making actions.
In sub-step 4042a, the infinite return cumulative sum in future of each decision-making action during calculating target can perform an action.
After collision avoidance control module determines that the target of X (t) correspondence can perform an action, the infinite return cumulative sum in future of each decision-making action in can be able to performing an action according to the enhancing calculated signals target of each decision-making action.Wherein, the infinite return cumulative sum in future of each decision-making action during collision avoidance control module can be able to perform an action according to infinite return in future cumulative sum computing formula calculating target.
Wherein, following infinite return cumulative sum computing formula is:
R(t)=r(t+1)+αr(t+2)+α 2r(t+3)+…
Wherein, R (t) represents following infinite return cumulative sum, r (t+1) represents the enhancing signal in t+1 moment, r (t+2) represents the enhancing signal in t+2 moment, r (t+3) represents the enhancing signal in t+3 moment, and α represents commutation factor, known see above-mentioned formula, what have the greatest impact to future returns cumulative sum is the enhancing signal in t+1 moment, and the enhancing signal in t+2 moment, t+3 moment is decayed with exponential form on the impact of future returns cumulative sum.
Illustratively, collision avoidance control module calculates the infinite return cumulative sum in future of decision-making action u1 (t), u2 (t) and u3 (t) respectively.
It should be noted that, when the position of obstacle is X (t), collision avoidance control module performs the position that any one decision-making action corresponding to X (t) all can change obstacle, the position of obstacle is made to become X (t+1), the control module of collision avoidance simultaneously can obtain an enhancing signal corresponding to X (t+1), when the position of obstacle is X (t+1), collision avoidance control module performs the position that any one decision-making action corresponding to X (t+1) all can change obstacle, the position of obstacle is made to become X (t+2), the control module of collision avoidance simultaneously can obtain an enhancing signal corresponding to X (t+2), the like.Illustratively, suppose that collision avoidance control module performs decision-making action u1 (t) corresponding to X (t), the position of obstacle is made to become X (t+1), the enhancing signal that collision avoidance control module obtains X (t+1) corresponding is r1 (t+1), the decision-making action that collision avoidance control module performs X (t+1) corresponding is u1 (t+1), the position of obstacle is made to become X (t+2), the enhancing signal that collision avoidance control module obtains X (t+2) corresponding is r1 (t+2), the like, the infinite return cumulative sum in future that then can obtain decision-making action u1 (t) corresponding according to infinite return in above-mentioned future cumulative sum computing formula is R1 (t)=r1 (t+1)+α r1 (t+2)+α 2r1 (t+3)+
In sub-step 4043a, using decision-making action maximum for infinite for future return cumulative sum as objective decision action.
Collision avoidance control module calculate target can perform an action in future of each decision-making action after infinite return cumulative sum, the decision-making action that future returns cumulative sum is maximum can be determined in infinite return in the future cumulative sum of all decision-making actions from target can perform an action, and using decision-making action maximum for infinite for future return cumulative sum as objective decision action.Illustratively, suppose that the future of decision-making action u1 (t), infinite return cumulative sum was R1 (t), the infinite return cumulative sum in future of decision-making action u2 (t) is R2 (t), the infinite return cumulative sum in future of decision-making action u3 (t) is R3 (t), and R1 (t) > R2 (t) > R3 (t), then collision avoidance control module using decision-making action u1 (t) as objective decision action.
It should be noted that, known see table 1, for position X (t) of the obstacle of any time t, multiple different decision-making action all can be had can to select for control loop.The position X (t+1) of the obstacle of next moment t+1 and corresponding enhancing signal r (t+1) also can be different to select different decision-making actions to mean.Although the standard of control loop trade-off decision action is dependent on the return strengthening signal and bring, this does not represent that control loop will select bring at subsequent time the decision-making action of maximal rewards in t.For dynamic changing process, the standard of optimizing decision Action Selection will be dependent on Bellman's principle of optimal, that is, after considering this decision-making action, remaining (future) is if the enhancing signal of the state likely existed, alternative action and feedback all optimums.
Also it should be noted that, before sub-step 4041a, control loop can first by the machine learning algorithm position obtaining the obstacle shown in table 1 and the corresponding relation that can perform an action.Wherein, the process of study can be carried out in computing machine, can at computing machine (such as, Matlab software at computing machine) innerly carry out simulated experiment, create vehicle running environment model, and design multiple plug in car situation, enable the control loop autonomous learning collision avoidance strategy of vehicle.Control loop after study can store learning experience (corresponding relation as shown in table 1), is installed on vehicle by control loop afterwards and uses.
Illustratively, please refer to Fig. 8, shown in it is a kind of position by machine learning algorithm determination obstacle provided embodiment illustrated in fig. 4 and the schematic diagram of the corresponding relation that can perform an action.See Fig. 8, environment sensing module can obtain position X (t) of obstacle, and position X (t) of obstacle is sent to collision avoidance control module, collision avoidance control module can to make a policy action u (t) according to position X (t) of obstacle, this decision-making action u (t) can change the position of vehicle, and then make the position of obstacle become X (t+1), simultaneously, environment sensing module can feed back to collision avoidance control module one and strengthen signal r (t), this enhancing signal r (t) represents the return immediately after performing decision-making action u (t), usually, strengthen signal can numerically exist, different numerical value is in order to evaluate " good " of the decision-making action made, " bad ", and the numerical value strengthening signal shows that more greatly corresponding decision-making action is better, the numerical value strengthening signal is less shows that corresponding decision-making action is poorer.Equally, for new position X (t+1), collision avoidance control module can be the decision-making action u (t+1) made new advances, and strengthens signal r (t+1) from obtaining one.The like go down, i.e. collision avoidance control module can be mutual with environment sensing module in each moment, by " good ", " bad " of the enhancing signal of environment sensing module feedback, on-line control decision strategy, to obtain maximum return in follow-up decision action, whole decision process is made to be tending towards optimum, finally, according to the quality strengthening signal determination decision-making action, the corresponding relation shown in table 1 can be obtained.
Again illustratively, please refer to Fig. 9, shown in it be the another kind provided embodiment illustrated in fig. 4 according to the position of obstacle, adopt the method flow diagram of machine learning algorithm determination objective decision action, see Fig. 9, the method flow process can comprise following several step:
In sub-step 4041b, obtain the decision-making action of chaufeur.
In embodiments of the present invention, when can there is the accident caused by obstacle in the running environment of vehicle, chaufeur can people be the action that makes a policy, such as, and chaufeur manual operation bearing circle, throttle, brake etc.Collision avoidance control module can obtain the decision-making action of chaufeur, and illustratively, collision avoidance control module can by reading throttle, and brake, the service data of bearing circle, obtains the decision-making action of chaufeur.
In sub-step 4042b, obtain the primary importance of the obstruction of the decision-making action performing chaufeur.
The decision-making action that chaufeur is made can change the position of vehicle, and then cause the position of obstacle to change, illustratively, the position performing the obstruction of the decision-making action of chaufeur can become primary importance, therefore, collision avoidance control module can obtain the primary importance of the obstruction of the decision-making action performing chaufeur, and suppose that the primary importance of obstacle is X (t+1), then collision avoidance control module can obtain the primary importance X (t+1) of obstacle.
Illustratively, please refer to Figure 10, shown in it is the method flow diagram that the collision avoidance control module provided embodiment illustrated in fig. 4 obtains the primary importance of the obstruction of the decision-making action performing chaufeur, see Figure 10, in embodiments of the present invention, collision avoidance control module obtain perform chaufeur decision-making action obstruction primary importance can comprise following several step:
In sub-step 4042b1, calculate the enhancing signal of following n state corresponding to the decision-making action of chaufeur, obtain n enhancing signal, n be greater than or equal to 1 integer.
Wherein, a following n state that is to say n the position in the future of obstacle.Suppose that the decision-making action of chaufeur is for decision-making action u2 (t) shown in table 1, then this decision-making action u2 (t) can change the position of obstacle, the position of obstacle is made to change into X (t+1) from X (t), the position X (t+1) of this obstacle can for first state in following n state of decision-making action u2 (t) correspondence of the person of sailing, the enhancing signal of position X (t+1) correspondence of then collision avoidance control module dyscalculia thing, known see table 1, the enhancing signal that the position X (t+1) of obstacle is corresponding comprises r1 (t+1), r2 (t+1), r3 (t+1) and r4 (t+1) is totally 4 enhancing signals, and the corresponding decision-making action of each enhancing signal, therefore, collision avoidance control module calculates the enhancing signal of following n-1 state corresponding to each decision-making action.Illustratively, collision avoidance control module calculates enhancing signal r1 (t+1) corresponding to decision-making action u1 (t+1), and calculate the position X (t+2) performing decision-making action u1 (t+1) obstruction, suppose that enhancing signal corresponding to decision-making action that this X (t+2) is corresponding is r1 (t+2), the like, collision avoidance control module can obtain n and strengthen signal.
In sub-step 4042b2, judge whether to exist in n enhancing signal to meet pre-conditioned enhancing signal.
Wherein, perform meet decision-making action corresponding to pre-conditioned enhancing signal time, the obstacle in the running environment of vehicle and vehicle can collide.Illustratively, pre-conditioned can be strengthen signal be less than or equal to-1, and also, collision avoidance control module judges n strengthens in signal whether there is the enhancing signal being less than or equal to-1.
In sub-step 4042b3, if n strengthens in signal to exist and meet pre-conditioned enhancing signal, then obtain the primary importance of the obstruction of the decision-making action performing chaufeur.
If in step 4042b2, collision avoidance control module is determined to exist in n enhancing signal to meet pre-conditioned enhancing signal, illustrate that the decision-making action of chaufeur exists error, therefore, the primary importance of the obstruction of the decision-making action performing chaufeur can be obtained by environment sensing module, and then revised by the decision-making action of primary importance to chaufeur of collision avoidance control module according to the obstruction of the decision-making action of execution chaufeur.
In sub-step 4043b, from experience thesaurus, determine that the target corresponding with the primary importance of obstacle can perform an action.
Wherein, the position that have recorded the obstacle determined by machine learning algorithm in advance in experience thesaurus and the corresponding relation that can perform an action, the position of each obstacle corresponding at least one can perform an action, each performing an action comprises decision-making action and strengthens signal one to one with decision-making action, strengthens signal and is used to indicate and strengthens the immediately return of signal decision-making action one to one when performing.
In sub-step 4044b, the infinite return cumulative sum in future of each decision-making action during calculating target can perform an action.
In sub-step 4045b, using decision-making action maximum for infinite for future return cumulative sum as objective decision action.
The implementation procedure of above-mentioned steps 4043b to step 4045b with embodiment illustrated in fig. 7 in step 4041a to step 4043a identical or similar, its implementation procedure can with reference to the step 4041a in embodiment illustrated in fig. 7 to step 4043a, and the embodiment of the present invention does not repeat them here.
In step 405, travel according to objective decision action control vehicle.
After the action of collision avoidance control module determination objective decision, can travel according to objective decision action control vehicle.Illustratively, collision avoidance control module controls vehicle according to objective decision action u1 (t) and travels.
It should be noted that, the sequencing of the drive manner step that the embodiment of the present invention provides can suitably adjust, step also according to circumstances can carry out corresponding increase and decrease, illustratively, if in sub-step 4042b2, collision avoidance control module is determined not exist in n enhancing signal to meet pre-conditioned enhancing signal, illustrate that the decision-making action of chaufeur does not exist error, control loop also can realize the effect of collision avoidance without the need to the decision-making action revising chaufeur, now, step 4043b to step 4045b also can not perform.Anyly be familiar with those skilled in the art in the technical scope that the present invention discloses, the method changed can be expected easily, all should be encompassed within protection scope of the present invention, therefore repeat no more.
In sum, the drive manner that the embodiment of the present invention provides, by the running environment of monitor vehicle in vehicle travel process; The accident caused by obstacle whether can be there is in the running environment of prediction vehicle; When can there is the accident caused by obstacle in the running environment of vehicle, obtain the position of obstacle; According to the position of obstacle, adopt the action of machine learning algorithm determination objective decision; Travel according to objective decision action control vehicle.Because the present invention adopts the action of machine learning algorithm determination objective decision, the determination of objective decision action is not by the restriction of the experience of chaufeur, the drive manner solved in correlation technique has limitation, stability is lower, the problem that alerting ability is poor, reach the field of application expanding drive manner, improve the stability of drive manner and the beneficial effect of alerting ability.
In the last few years, the safety issue of vehicle received increasing concern.Show according to investigations, the first six national communication death tolls of world's vehicle population in 2011 is respectively: the U.S.: 32310 people, China: 62000 people, Japan: 4612 people, Germany: 4009 people, Italy: 3800 people, Russia: 27900 people.Kuomintang-Communist generation traffic accident 204196 in 2012, dead 59997 people, injured 224327 people, direct property loss 117489.6 ten thousand yuan.As can be seen from data, annual toll on traffic is more than 100,000 people in the world, and wherein the annual toll on traffic of China is more than 60,000 people, and direct economic loss reaches more than 10 hundred million yuan more than a year.Therefore, traffic accident brings huge loss to personal safety and national economy property.The drive manner that the embodiment of the present invention provides, can avoid vehicle and obstacle to collide, can reduce the generation of traffic accident, and then the economic loss that minimizing traffic accident brings.
It should be noted that: the control loop that above-described embodiment provides is when steering vehicle, only be illustrated with the division of above-mentioned each functional module, in practical application, can distribute as required and by above-mentioned functions and be completed by different functional modules, inner structure by equipment is divided into different functional modules, to complete all or part of function described above.In addition, the drive manner that above-described embodiment provides and system embodiment belong to same design, and its implementation procedure refers to embodiment of the method, repeats no more here.
One of ordinary skill in the art will appreciate that all or part of step realizing above-described embodiment can have been come by hardware, the hardware that also can carry out instruction relevant by program completes, described program can be stored in a kind of computer-readable recording medium, the above-mentioned storage medium mentioned can be read-only memory (ROM), disk or CD etc.
The foregoing is only preferred embodiment of the present invention, not in order to limit the present invention, within the spirit and principles in the present invention all, any amendment done, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.

Claims (10)

1. a control loop, is characterized in that, described control loop comprises: environment sensing module and collision avoidance control module,
Described environment sensing module is used for the running environment of monitoring described vehicle in vehicle travel process; Predict the accident that whether can occur in the running environment of described vehicle to be caused by obstacle; When can there is the accident caused by obstacle in the running environment of described vehicle, obtain the position of described obstacle;
Described collision avoidance control module is used for the position according to described obstacle, adopts the action of machine learning algorithm determination objective decision; According to described objective decision action control, vehicle travels.
2. control loop according to claim 1, is characterized in that,
Described collision avoidance control module is used for:
From experience thesaurus, determine that the target corresponding with the position of described obstacle can perform an action, the position that have recorded the obstacle determined by machine learning algorithm in advance in described experience thesaurus and the corresponding relation that can perform an action, the position of each obstacle corresponding at least one can perform an action, each performing an action comprises decision-making action and strengthens signal one to one with described decision-making action, and described enhancing signal is used to indicate and the immediately return of described enhancing signal decision-making action one to one when performing;
Calculate described target can perform an action in the infinite return cumulative sum in future of each decision-making action;
Using decision-making action maximum for infinite for described future return cumulative sum as described objective decision action.
3. control loop according to claim 1, is characterized in that,
Described collision avoidance control module is for obtaining the decision-making action of chaufeur;
Described environment sensing module for obtain perform described chaufeur decision-making action after the primary importance of described obstacle;
Described collision avoidance control module is used for from experience thesaurus, determine that the target corresponding with the primary importance of described obstacle can perform an action, the position that have recorded the obstacle determined by machine learning algorithm in advance in described experience thesaurus and the corresponding relation that can perform an action, the position of each obstacle corresponding at least one can perform an action, each performing an action comprises decision-making action and strengthens signal one to one with described decision-making action, and described enhancing signal is used to indicate and the immediately return of described enhancing signal decision-making action one to one when performing;
Described collision avoidance control module is for the infinite return cumulative sum in future of each decision-making action in calculating described target and can performing an action; Using decision-making action maximum for infinite for described future return cumulative sum as described objective decision action.
4. control loop according to claim 3, is characterized in that,
Described collision avoidance control module is used for:
Calculate the enhancing signal of following n state corresponding to the decision-making action of described chaufeur, obtain n and strengthen signal, described n be greater than or equal to 1 integer;
Judge whether to exist in described n enhancing signal to meet pre-conditioned enhancing signal;
When existence meets pre-conditioned enhancing signal in described n enhancing signal, trigger the primary importance of described obstacle after described environment sensing module obtains the decision-making action performing described chaufeur;
Wherein, when meeting decision-making action corresponding to pre-conditioned enhancing signal described in performing, the obstacle in the running environment of described vehicle and described vehicle can collide.
5., according to the arbitrary described control loop of Claims 1-4, it is characterized in that,
Described environment sensing module is used for:
Judge whether there is obstacle in the running environment of described vehicle;
When there is obstacle in the running environment of described vehicle, judge whether described obstacle is in preset range;
When described obstacle is in described preset range, determine the accident that can occur in the running environment of described vehicle to be caused by described obstacle.
6. a drive manner, is characterized in that, described method comprises:
The running environment of described vehicle is monitored in vehicle travel process;
Predict the accident that whether can occur in the running environment of described vehicle to be caused by obstacle;
The accident caused by obstacle if can occur in the running environment of described vehicle, then obtain the position of described obstacle;
According to the position of described obstacle, adopt the action of machine learning algorithm determination objective decision;
According to described objective decision action control, vehicle travels.
7. method according to claim 6, is characterized in that,
The described position according to described obstacle, adopts the action of machine learning algorithm determination objective decision, comprising:
From experience thesaurus, determine that the target corresponding with the position of described obstacle can perform an action, the position that have recorded the obstacle determined by machine learning algorithm in advance in described experience thesaurus and the corresponding relation that can perform an action, the position of each obstacle corresponding at least one can perform an action, each performing an action comprises decision-making action and strengthens signal one to one with described decision-making action, and described enhancing signal is used to indicate and the immediately return of described enhancing signal decision-making action one to one when performing;
Calculate described target can perform an action in the infinite return cumulative sum in future of each decision-making action;
Using decision-making action maximum for infinite for described future return cumulative sum as described objective decision action.
8. method according to claim 6, is characterized in that,
The described position according to described obstacle, adopts the action of machine learning algorithm determination objective decision, comprising:
Obtain the decision-making action of chaufeur;
The primary importance of described obstacle after obtaining the decision-making action performing described chaufeur;
From experience thesaurus, determine that the target corresponding with the primary importance of described obstacle can perform an action, the position that have recorded the obstacle determined by machine learning algorithm in advance in described experience thesaurus and the corresponding relation that can perform an action, the position of each obstacle corresponding at least one can perform an action, each performing an action comprises decision-making action and strengthens signal one to one with described decision-making action, and described enhancing signal is used to indicate and the immediately return of described enhancing signal decision-making action one to one when performing;
Calculate described target can perform an action in the infinite return cumulative sum in future of each decision-making action;
Using decision-making action maximum for infinite for described future return cumulative sum as described objective decision action.
9. method according to claim 8, is characterized in that, the primary importance of described obstacle after the decision-making action of the described chaufeur of described acquisition execution, comprising:
Calculate the enhancing signal of following n state corresponding to the decision-making action of described chaufeur, obtain n and strengthen signal, described n be greater than or equal to 1 integer;
Judge whether to exist in described n enhancing signal to meet pre-conditioned enhancing signal;
If described n strengthens in signal to exist and meets pre-conditioned enhancing signal, then the primary importance of described obstacle after obtaining the decision-making action performing described chaufeur;
Wherein, when meeting decision-making action corresponding to pre-conditioned enhancing signal described in performing, the obstacle in the running environment of described vehicle and described vehicle can collide.
10., according to the arbitrary described method of claim 6 to 9, it is characterized in that whether the accident caused by obstacle can occur in the running environment of the described vehicle of described prediction, comprising:
Judge whether there is obstacle in the running environment of described vehicle;
If there is obstacle in the running environment of described vehicle, then judge whether described obstacle is in preset range;
If described obstacle is in described preset range, then determine the accident that can occur in the running environment of described vehicle to be caused by described obstacle.
CN201510532385.0A 2015-08-24 2015-08-24 Driving method and driving system CN105109483A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510532385.0A CN105109483A (en) 2015-08-24 2015-08-24 Driving method and driving system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510532385.0A CN105109483A (en) 2015-08-24 2015-08-24 Driving method and driving system

Publications (1)

Publication Number Publication Date
CN105109483A true CN105109483A (en) 2015-12-02

Family

ID=54657679

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510532385.0A CN105109483A (en) 2015-08-24 2015-08-24 Driving method and driving system

Country Status (1)

Country Link
CN (1) CN105109483A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109109863A (en) * 2018-07-28 2019-01-01 华为技术有限公司 Smart machine and its control method, device

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101395647A (en) * 2006-02-28 2009-03-25 丰田自动车株式会社 Object course prediction method, device, program, and automatic driving system
US20090093960A1 (en) * 2007-10-04 2009-04-09 Jeffrey Scott Puhalla Method and system for obstacle avoidance for a vehicle
CN102431553A (en) * 2011-10-18 2012-05-02 奇瑞汽车股份有限公司 Active safety system and method of vehicle
CN103171554A (en) * 2011-12-26 2013-06-26 现代自动车株式会社 System and method for controlling inter-vehicle distance using side and rear sensor
CN103303309A (en) * 2013-06-26 2013-09-18 奇瑞汽车股份有限公司 Automobile rear-end collision early-warning method and system

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101395647A (en) * 2006-02-28 2009-03-25 丰田自动车株式会社 Object course prediction method, device, program, and automatic driving system
US20090093960A1 (en) * 2007-10-04 2009-04-09 Jeffrey Scott Puhalla Method and system for obstacle avoidance for a vehicle
CN102431553A (en) * 2011-10-18 2012-05-02 奇瑞汽车股份有限公司 Active safety system and method of vehicle
CN103171554A (en) * 2011-12-26 2013-06-26 现代自动车株式会社 System and method for controlling inter-vehicle distance using side and rear sensor
CN103303309A (en) * 2013-06-26 2013-09-18 奇瑞汽车股份有限公司 Automobile rear-end collision early-warning method and system

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
方啸: "GrHDP算法的研究及其在虚拟交互中的应用", 《中国博士学位论文全文数据库》 *
方啸: "基于自适应动态规划算法的小车自主导航控制策略设计", 《燕山大学学报》 *
方啸: "移动机器人自主寻路避障启发式动态规划算法", 《农业机械学报》 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109109863A (en) * 2018-07-28 2019-01-01 华为技术有限公司 Smart machine and its control method, device

Similar Documents

Publication Publication Date Title
Luo et al. A dynamic automated lane change maneuver based on vehicle-to-vehicle communication
US9934688B2 (en) Vehicle trajectory determination
US9405293B2 (en) Vehicle trajectory optimization for autonomous vehicles
EP3092599B1 (en) Systems and methods for mimicking a leading vehicle
CN103373351B (en) System and method for vehicle lateral control
CN105015545B (en) A kind of autonomous lane change decision-making technique of pilotless automobile
DE102015114464A9 (en) Uniform motion planner for an autonomous vehicle while avoiding a moving obstacle
CN103158705B (en) Method and system for controlling a host vehicle
JP2017102907A (en) Method of selecting optimized rail track
Guldner et al. Analysis of automatic steering control for highway vehicles with look-down lateral reference systems
CN106218637B (en) A kind of automatic Pilot method
Makarem et al. Model predictive coordination of autonomous vehicles crossing intersections
US10239529B2 (en) Autonomous vehicle operation based on interactive model predictive control
CN102815299B (en) By in fixed for track/lane sensing of lane markings identification that keeps
CN103754221B (en) Vehicle adaptive cruise control system
JPWO2017010264A1 (en) Vehicle control device, vehicle control method, and vehicle control program
JP6308032B2 (en) System and method for generating driving maneuvers
CN107264531B (en) The autonomous lane-change of intelligent vehicle is overtaken other vehicles motion planning method in a kind of semi-structure environment
CN108692734B (en) Path planning method and device
Kurt et al. Hybrid-state driver/vehicle modelling, estimation and prediction
US9096267B2 (en) Efficient data flow algorithms for autonomous lane changing, passing and overtaking behaviors
CN106940933B (en) A kind of intelligent vehicle decision lane-change method based on intelligent transportation system
CN103448723B (en) Use the Lane Keeping System of rear pick up camera
US10074279B1 (en) Inference-aware motion planning
US8170739B2 (en) Path generation algorithm for automated lane centering and lane changing control system

Legal Events

Date Code Title Description
PB01 Publication
C06 Publication
SE01 Entry into force of request for substantive examination
C10 Entry into substantive examination
RJ01 Rejection of invention patent application after publication

Application publication date: 20151202

RJ01 Rejection of invention patent application after publication