CN109760681B - Channel changing control method and device - Google Patents
Channel changing control method and device Download PDFInfo
- Publication number
- CN109760681B CN109760681B CN201711064941.1A CN201711064941A CN109760681B CN 109760681 B CN109760681 B CN 109760681B CN 201711064941 A CN201711064941 A CN 201711064941A CN 109760681 B CN109760681 B CN 109760681B
- Authority
- CN
- China
- Prior art keywords
- vehicle
- lane
- action
- data
- lane change
- 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.)
- Active
Links
Images
Abstract
The embodiment of the invention discloses a lane change control method and a lane change control device, wherein the method comprises the following steps: acquiring current environment data of the surrounding environment of the controlled vehicle at the current moment; obtaining the score of the lane changing action at the current moment according to a pre-obtained action evaluation model and current environment data, wherein the lane changing action comprises acceleration and a steering angle; determining the lane changing action corresponding to the highest score as the selected lane changing action at the current moment; and controlling the controlled vehicle to act according to the selected lane changing action. The embodiment of the invention comprehensively considers the acceleration and the steering angle in the lane changing process, avoids the stability problem caused by forcibly decoupling the longitudinal and transverse plans, and increases the robustness and the comfort of passengers.
Description
Technical Field
The invention relates to the technical field of automatic driving, in particular to a lane changing control method and a lane changing control device.
Background
The automatic lane changing is one of key technologies for realizing automatic driving of vehicles, but is limited by nonlinear characteristics of automobile dynamics and actual complex driving environment, and how to ensure safe, quick and stable lane changing of the vehicles in the automatic driving process is the key and difficult point of the research of the current automatic driving technology.
In the automatic lane changing control process, a local route is calculated by identifying the adjacent lane line of the controlled vehicle and the position and the speed of the vehicle on the adjacent lane line, so that the controlled vehicle moves to the center line of the lane where the adjacent vehicle arrives through the local route. The existing lane change control technology has the following defects: the coupled longitudinal and transverse control plans are decoupled forcibly, and the longitudinal acceleration and the transverse steering are planned separately, so that the robustness of the automatic driving automobile is not high and the comfort of passengers is not good in the lane changing process.
Disclosure of Invention
In view of the above, the invention provides a lane change control method and device, which can solve the problems of low robustness and poor comfort in the lane change process of the existing automatic driving automobile.
The lane change control method provided by the embodiment of the invention comprises the following steps:
acquiring current environment data of the surrounding environment of the controlled vehicle at the current moment;
obtaining the score of the lane changing action at the current moment according to a pre-obtained action evaluation model and the current environment data, wherein the lane changing action comprises acceleration and a steering angle;
determining the lane changing action corresponding to the highest score as the selected lane changing action at the current moment;
and controlling the controlled vehicle to act according to the selected lane changing action.
Optionally, the method for establishing the action evaluation model specifically includes:
acquiring a lane change action training set, wherein the lane change action training set comprises a plurality of groups of lane change data, and each group of lane change data comprises each lane change action executed by a driver controlling a first vehicle in a lane change process and vehicle speed data of the first vehicle and environmental data of the surrounding environment thereof when the lane change action is executed;
training an objective function based on a convolutional neural network pair according to the lane change action training set
Training is carried out, and after the training is converged, the action evaluation model q (s, a) is obtained;
wherein, a is the current lane changing action, s is the current environment data, a 'is the lane changing action at the next moment, s' is the environment data at the next moment, γ is the learning rate, and R(s) is the instant report at the current moment.
Alternatively to this, the first and second parts may,
in the lane changing process, the instant return is positively correlated with the first comfort data and the second comfort data, and the instant return is also positively correlated with the first safety data and/or the second safety data;
wherein the first comfort data is inversely related to a degree of change of a lateral acceleration of the first vehicle over a preset time period; the second comfort data is inversely related to a degree of change of the longitudinal acceleration of the first vehicle over the preset time period; the first safety data are positively correlated with a first distance and the vehicle speed of a second vehicle, the second vehicle is in front of the first vehicle and is positioned in a lane to be turned, the first distance is the longitudinal distance between the second vehicle and the first vehicle, and the first safety data are also negatively correlated with the vehicle speed, the maximum braking deceleration and the reaction time delay of the first vehicle; the second safety data are positively correlated with a second distance and the vehicle speed of the first vehicle, the second distance is the longitudinal distance between a third vehicle and the first vehicle, the third vehicle is behind the first vehicle and is positioned in the lane to be turned, and the second safety data are also negatively correlated with the vehicle speed, the maximum braking deceleration and the reaction time delay of the third vehicle;
at the end of the lane change, the instant return is 100.
Optionally, in the lane changing process, the instant report specifically includes:
wherein, [ f ]0,f1]Is a lateral acceleration spectrum, [ f ]2,f3]Is a spectrum of longitudinal acceleration, axFor the lateral acceleration of the first vehicle within the predetermined time period, ayLongitudinal acceleration, y, of the first vehicle in the preset time period1Is the first distance, y2Is said second distance, v1Speed, v, of said first vehicle2Speed, v, of said second vehicle3Is the speed of the third vehicle,
Optionally, the obtaining, according to a pre-obtained action evaluation model and the current environment data, a score of the lane change action at the current time specifically includes:
inputting the current environment into the action evaluation model q (s, a) to obtain a scoring function q (a) of the lane-changing action at the current moment;
the determining the lane change action corresponding to the highest score as the selected lane change action at the current time specifically includes:
according to the formula
Determining the selected lane action aIs selected by。
The lane change control device provided by the embodiment of the invention comprises: the device comprises an acquisition unit, a scoring unit, a determination unit and a control unit;
the acquisition unit is used for acquiring current environment data of the surrounding environment of the controlled vehicle at the current moment;
the scoring unit is used for obtaining the score of the lane changing action at the current moment according to a pre-obtained action evaluation model and the current environment data, wherein the lane changing action comprises acceleration and a steering angle;
the determining unit is used for determining the lane changing action corresponding to the highest score as the selected lane changing action at the current moment;
and the control unit is used for controlling the controlled vehicle to act according to the selected lane changing action.
Optionally, the apparatus further includes: a model training unit; the model training unit is specifically configured to:
acquiring a lane change action training set, wherein the lane change action training set comprises a plurality of groups of lane change data, and each group of lane change data comprises each lane change action executed by a driver controlling a first vehicle in a lane change process and vehicle speed data of the first vehicle and environmental data of the surrounding environment thereof when the lane change action is executed;
training an objective function based on a convolutional neural network pair according to the lane change action training set
Training is carried out, and after the training is converged, the action evaluation model q (s, a) is obtained;
wherein, a is the current lane changing action, s is the current environment data, a 'is the lane changing action at the next moment, s' is the environment data at the next moment, γ is the learning rate, and R(s) is the instant report at the current moment.
Alternatively to this, the first and second parts may,
in the lane changing process, the instant return is positively correlated with the first comfort data and the second comfort data, and the instant return is also positively correlated with the first safety data and/or the second safety data;
wherein the first comfort data is inversely related to a degree of change of a lateral acceleration of the first vehicle over a preset time period; the second comfort data is inversely related to a degree of change of the longitudinal acceleration of the first vehicle over the preset time period; the first safety data are positively correlated with a first distance and the vehicle speed of a second vehicle, the second vehicle is in front of the first vehicle and is positioned in a lane to be turned, the first distance is the longitudinal distance between the second vehicle and the first vehicle, and the first safety data are also negatively correlated with the vehicle speed, the maximum braking deceleration and the reaction time delay of the first vehicle; the second safety data are positively correlated with a second distance and the vehicle speed of the first vehicle, the second distance is the longitudinal distance between a third vehicle and the first vehicle, the third vehicle is behind the first vehicle and is positioned in the lane to be turned, and the second safety data are also negatively correlated with the vehicle speed, the maximum braking deceleration and the reaction time delay of the third vehicle;
at the end of the lane change, the instant return is 100.
Optionally, in the lane changing process, the instant report specifically includes:
wherein, [ f ]0,f1]Is a lateral acceleration spectrum, [ f ]2,f3]Is a spectrum of longitudinal acceleration, axFor the lateral acceleration of the first vehicle within the predetermined time period, ayLongitudinal acceleration, y, of the first vehicle in the preset time period1Is the first distance, y2Is said second distance, v1Speed, v, of said first vehicle2Speed, v, of said second vehicle3Is the speed of the third vehicle,
Optionally, the scoring unit is specifically configured to:
inputting the current environment into the action evaluation model q (s, a) to obtain a scoring function q (a) of the lane-changing action at the current moment;
the determining the lane change action corresponding to the highest score as the selected lane change action at the current time specifically includes:
according to the formula
Determining the selected lane action aIs selected by。
An embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed, the following steps are implemented:
acquiring current environment data of the surrounding environment of the controlled vehicle at the current moment;
obtaining the score of the lane changing action at the current moment according to a pre-obtained action evaluation model and the current environment data, wherein the lane changing action comprises acceleration and a steering angle;
determining the lane changing action corresponding to the highest score as the selected lane changing action at the current moment;
and controlling the controlled vehicle to act according to the selected lane changing action.
The embodiment of the present invention further provides a vehicle control unit, including: a memory and a processor;
the memory for storing a computer program that when executed by the processor is capable of performing the steps of:
acquiring current environment data of the surrounding environment of the controlled vehicle at the current moment;
obtaining the score of the lane changing action at the current moment according to a pre-obtained action evaluation model and the current environment data, wherein the lane changing action comprises acceleration and a steering angle;
determining the lane changing action corresponding to the highest score as the selected lane changing action at the current moment;
and controlling the controlled vehicle to act according to the selected lane changing action.
Compared with the prior art, the invention has at least the following advantages:
in the embodiment of the invention, the environmental data around the controlled vehicle is firstly collected, the combined score of the acceleration and the steering angle at the current moment is obtained according to the environmental data and the action evaluation model obtained by pre-training, namely the similarity between the acceleration and the steering angle and the real driving action in the current state is obtained, and the lane changing action corresponding to the highest score is adopted to control the action of the controlled vehicle, so that the lane changing process of the controlled vehicle is close to the real lane changing action of a driver in the current state, and the comfort level of passengers is ensured. The embodiment of the invention comprehensively considers the acceleration and the steering angle in the lane changing process, avoids the stability problem caused by forcibly decoupling the longitudinal and transverse plans, and increases the robustness and the comfort of passengers.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments described in the present application, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a schematic diagram of a vehicle lane changing to the left;
fig. 2 is a schematic flow chart of a lane change control method according to an embodiment of the present invention;
fig. 3 is a schematic flow chart of another lane change control method according to an embodiment of the present invention;
FIG. 4 is a diagram illustrating a training process of a convolutional neural network according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of a lane change control device according to an embodiment of the present invention.
Detailed Description
In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Before describing the embodiments of the present invention, a specific procedure for changing lanes of a vehicle will be described. The process of changing lanes to the left is shown in fig. 1, where a vehicle a will change lanes to the left, with a vehicle C in front of the adjacent lane to the left, and a vehicle B behind, with a lane line L for the vehicle a1、L2And L3Are each x1、x2And x3. The condition of the end of lane change is | x1-x2< Kw and | h-hmAnd | < C. W is the lane width of the lane on the left side, K is a constant, the degree of the vehicle running in the middle of the lane is represented, and the smaller K is, the closer the vehicle is to the lane center line LmRunning, generally taking K as 0-0.1, h as the vehicle course angle, hmThe left lane navigation angle is C, which is a constant and is generally 0-0.1.
Referring to fig. 2, the figure is a schematic flow chart of a lane change control method according to an embodiment of the present invention. In the embodiment of the present invention, the left lane changing of the controlled vehicle is taken as an example for description, and the control method for the right lane changing of the controlled vehicle is similar, which is specifically referred to the relevant description of the left lane changing, and is not described in detail in the embodiment of the present invention.
The lane-changing control method provided by the embodiment comprises the following steps S201-S204.
S201: and acquiring current environment data of the environment around the controlled vehicle at the current moment.
In this embodiment, the environment data of the controlled vehicle may be acquired by the environment sensing module, where the environment data includes vehicle information around the controlled vehicle, lane line information, and the like.
For example, the environment sensing module is responsible for acquiring an overhead view image around a controlled vehicle, and specifically, the overhead view image may be formed by point cloud data scanned by a multi-line laser radar, or the overhead view image may be obtained by stitching video images acquired by a front camera, a rear camera, and left and right panoramic cameras. It should be noted that, in order to ensure the safety of automatic lane changing, the detection distance of the camera needs to be greater than 100 meters.
S202: and obtaining the score of the lane changing action at the current moment according to the action evaluation model obtained in advance and the current environment data.
Here, the lane change operation includes an acceleration and a lane change angle (or a steering wheel angle) of the vehicle, and the lane change operation may be represented by a (α, θ), where α is the acceleration and θ is the lane change angle. For example, a (1, -2) represents that the controlled vehicle accelerates at 1m/s, and the lane change angle of the controlled vehicle is 2 degrees for left turning; a (-1,2) represents that the controlled vehicle decelerates at 1m/s, and the lane change angle of the controlled vehicle is 2 degrees to the right.
It can be understood that, in the existing automatic lane changing technology, when the control planning is performed on the longitudinal direction (i.e. acceleration) and the lateral direction (i.e. steering), the longitudinal direction and the lateral direction are respectively controlled in a planning manner, and are mutually coupled, and the longitudinal direction and the lateral direction are mutually influenced, so that the robustness of the lane changing control is poor, and the comfort of passengers is not high. In the embodiment of the invention, longitudinal control and transverse control are integrated (namely lane change action), and longitudinal and transverse planning of the vehicle is integrally planned, so that the robustness and the comfort are improved.
In the embodiment, the action evaluation model is related to the environmental data around the controlled vehicle and the lane-changing action performed by the controlled vehicle, and is used for evaluating the similarity between the lane-changing action performed by the controlled vehicle and the real driving operation of the driver in the current state, and the higher the score is, the higher the similarity is. The action evaluation model can be obtained by utilizing convolutional neural network training according to the lane changing action of the driver in the actual driving lane changing process which is collected in advance.
S203: and determining the lane changing action corresponding to the highest score as the selected lane changing action at the current moment.
S204: and controlling the controlled vehicle to act according to the selected lane changing action.
Through the steps S201 and S202, a score function of the lane change executed by the controlled vehicle under the environment data is obtained, that is, the similarity between the lane change executed by the current controlled vehicle and the actual driving operation of the driver is obtained, and the score is positively correlated with the similarity, so that the lane change corresponding to the highest score is determined as the selected lane change, and the controlled vehicle is controlled to execute the selected lane change, so that the lane change process of the controlled vehicle has a human-like driving effect and has better comfort.
In the embodiment, the environmental data around the controlled vehicle is collected firstly, the combined score of the acceleration and the steering angle at the current moment is obtained according to the environmental data and the action evaluation model obtained by pre-training, namely the similarity between the acceleration and the steering angle and the real driving action in the current state is obtained, and the lane changing action corresponding to the highest score is adopted to control the action of the controlled vehicle, so that the lane changing process of the controlled vehicle is close to the real lane changing action of a driver in the current state, and the comfort level of passengers is ensured. The embodiment of the invention comprehensively considers the acceleration and the lane change angle in the lane change process, avoids the stability problem caused by forced decoupling of longitudinal and transverse plans, and increases the robustness and the comfort level of passengers.
The following illustrates how the action evaluation model is specifically established.
Referring to fig. 3, in the embodiment of the present invention, the motion evaluation model may be specifically obtained by training through the method described in the following steps S301 to S302.
S301: and acquiring a lane change action training set.
The lane changing action training set comprises a plurality of groups of lane changing data, and each group of lane changing data comprises each lane changing action executed by a driver controlling a first vehicle in a lane changing process, and vehicle speed data of the first vehicle and environment data of the surrounding environment of the first vehicle when the lane changing action is executed.
It can be understood that the action evaluation model is trained based on the actual lane changing action of the driver, so that the lane changing action with the highest score has higher similarity with the lane changing operation actually executed by the driver in the current environment during actual control, and the robustness of the lane changing process and the comfort of passengers are improved.
It should be noted that, in practical applications, due to differences in vehicle types, lane change actions actually performed may be different, and in order to ensure accuracy and comfort of lane change, a lane change action training set used in training an action evaluation model needs to be related to an actual controlled vehicle, that is, a first vehicle and the controlled vehicle have the same vehicle type (or similar size).
S302: training a training target function shown in the following formula (1) based on a convolutional neural network according to a lane changing action training set, and obtaining an action evaluation model q (s, a) after the training is converged;
wherein, a is the current lane changing action, s is the current environment data, a 'is the lane changing action at the next moment, s' is the environment data at the next moment, γ is the learning rate, and R(s) is the instant report at the current moment.
The training process of the convolutional neural network is shown in fig. 4, and a person skilled in the art can specifically set parameters of the convolutional layer, the pooling layer and the full-link layer according to actual conditions, and the specific training process of the convolutional neural network is not described again here.
In practical applications, the learning rate may be equal to 0.9.
In a possible implementation manner of this embodiment, in the lane change process, the instant report is positively correlated with the first comfort data and the second comfort data, and the instant report is also positively correlated with the first safety data and/or the second safety data; at the end of the lane change, the real-time report is 100.
Wherein the first comfort data is inversely related to the variation degree of the lateral acceleration of the first vehicle (for example, vehicle a in fig. 1) within a preset time period; the second comfort data is inversely related to the change degree of the longitudinal acceleration of the first vehicle in the preset time period; the first safety data is positively correlated with a first distance and the vehicle speed of a second vehicle (such as a vehicle B in fig. 1), the second vehicle is in front of the first vehicle and is positioned in a lane to be turned, the first distance is the longitudinal distance between the second vehicle and the first vehicle, and the first safety data is also negatively correlated with the vehicle speed, the maximum braking deceleration and the reaction time delay of the first vehicle; the second safety data is positively correlated with a second distance and the vehicle speed of the first vehicle, the second distance is a longitudinal distance between a third vehicle (such as the vehicle C in fig. 1) and the first vehicle, the third vehicle is behind the first vehicle and is positioned in a lane to be turned, and the second safety data is also negatively correlated with the vehicle speed, the maximum braking deceleration and the reaction time delay of the third vehicle.
As an example, during the lane change process, the immediate reward is specifically given by the following equation (2):
wherein, [ f ]0,f1]Is a lateral acceleration spectrum, [ f ]2,f3]Is a spectrum of longitudinal acceleration, axFor the lateral acceleration of the first vehicle over a predetermined period of time, ayLongitudinal acceleration of the first vehicle over a predetermined time period, y1Is a first distance, y2Is a second distance, v1Speed, v, of the first vehicle2Speed, v, of the second vehicle3Is the speed of the third vehicle, amaxK corresponds to the maximum braking deceleration of the vehicle, tau corresponds to the reaction time delay of the vehicle,
then, steps S202 to S203 in the foregoing embodiment may specifically include:
inputting the current environment into an action evaluation model q (s, a) to obtain a scoring function q (a) of the lane changing action at the current moment; determining the lane change action corresponding to the highest score as the current timeThe selected lane changing action specifically includes: determining the selected track action a according to the formula (4)Is selected by。
Based on the lane change control method provided by the embodiment, the embodiment of the invention also provides a lane change control device.
Referring to fig. 5, the figure is a schematic structural diagram of a lane change control device according to an embodiment of the present invention.
The lane change control device provided by the embodiment comprises: an acquisition unit 100, a scoring unit 200, a determination unit 300, and a control unit 400;
an obtaining unit 100, configured to obtain current environment data of a surrounding environment of a controlled vehicle at a current time;
the scoring unit 200 is used for obtaining a score of the lane changing action at the current moment according to a pre-obtained action evaluation model and current environment data, wherein the lane changing action comprises an acceleration and a steering angle;
a determining unit 300, configured to determine the lane change action corresponding to the highest score as the selected lane change action at the current time;
and the control unit 400 is used for controlling the controlled vehicle to act according to the selected lane changing action.
In some possible implementations of this embodiment, the method further includes: a model training unit;
a model training unit specifically configured to:
acquiring a lane change action training set, wherein the lane change action training set comprises a plurality of groups of lane change data, and each group of lane change data comprises each lane change action executed by a driver controlling a first vehicle in a lane change process and vehicle speed data of the first vehicle and environmental data of the surrounding environment thereof when the lane change action is executed;
training a training target function shown in the following formula (1) based on a convolutional neural network according to a lane changing action training set, and obtaining an action evaluation model q (s, a) after the training is converged;
wherein, a is the current lane changing action, s is the current environment data, a 'is the lane changing action at the next moment, s' is the environment data at the next moment, γ is the learning rate, and R(s) is the instant report at the current moment.
In some possible implementation manners of this embodiment, in the lane changing process, the instant report is positively correlated with the first comfort data and the second comfort data, and the instant report is also positively correlated with the first safety data and/or the second safety data;
the first comfort data is inversely related to the change degree of the lateral acceleration of the first vehicle in a preset time period; the second comfort data is inversely related to the change degree of the longitudinal acceleration of the first vehicle in the preset time period; the first safety data are positively correlated with a first distance and the vehicle speed of a second vehicle, the second vehicle is in front of the first vehicle and is positioned in a lane to be turned, the first distance is the longitudinal distance between the second vehicle and the first vehicle, and the first safety data are also negatively correlated with the vehicle speed, the maximum braking deceleration and the reaction time delay of the first vehicle; the second safety data are positively correlated with a second distance and the vehicle speed of the first vehicle, the second distance is the longitudinal distance between the third vehicle and the first vehicle, the third vehicle is behind the first vehicle and is positioned in a lane to be turned, and the second safety data are also negatively correlated with the vehicle speed, the maximum braking deceleration and the reaction time delay of the third vehicle;
at the end of the lane change, the real-time report is 100.
In some possible implementations of this embodiment, during the lane change process, the immediate reward is specifically given by the following formula (2):
wherein, [ f ]0,f1]Is a lateral acceleration spectrum, [ f ]2,f3]Is a spectrum of longitudinal acceleration, axFor the lateral acceleration of the first vehicle over a predetermined period of time, ayLongitudinal acceleration of the first vehicle over a predetermined time period, y1Is a first distance, y2Is a second distance, v1Speed, v, of the first vehicle2Speed, v, of the second vehicle3Is the speed of the third vehicle, amaxK corresponds to the maximum braking deceleration of the vehicle, tau corresponds to the reaction time delay of the vehicle,
in some possible implementation manners of this embodiment, the scoring unit is specifically configured to:
inputting the current environment into an action evaluation model q (s, a) to obtain a scoring function q (a) of the lane changing action at the current moment;
determining the lane change action corresponding to the highest score as the selected lane change action at the current moment, specifically comprising:
determining the selected lane action a according to the following formula (4)Is selected by。
In the embodiment, the environmental data around the controlled vehicle is collected firstly, the combined score of the acceleration and the steering angle at the current moment is obtained according to the environmental data and the action evaluation model obtained by pre-training, namely the similarity between the acceleration and the steering angle and the real driving action in the current state is obtained, and the lane changing action corresponding to the highest score is adopted to control the action of the controlled vehicle, so that the lane changing process of the controlled vehicle is close to the real lane changing action of a driver in the current state, and the comfort level of passengers is ensured. The embodiment of the invention comprehensively considers the acceleration and the lane change angle in the lane change process, avoids the stability problem caused by forced decoupling of longitudinal and transverse plans, and increases the robustness and the comfort level of passengers.
Based on the lane change control method and device provided by the above embodiments, the embodiment of the present invention further provides a computer readable storage medium. The computer-readable storage medium having stored thereon a computer program that, when executed, performs the steps of:
acquiring current environment data of the surrounding environment of the controlled vehicle at the current moment;
obtaining the score of the lane changing action at the current moment according to a pre-obtained action evaluation model and the current environment data, wherein the lane changing action comprises acceleration and a steering angle;
determining the lane changing action corresponding to the highest score as the selected lane changing action at the current moment;
and controlling the controlled vehicle to act according to the selected lane changing action.
Based on the lane change control method and device provided by the embodiment, the embodiment of the invention also provides a vehicle control unit. This vehicle control unit includes: a memory and a processor;
the memory for storing a computer program that when executed by the processor is capable of performing the steps of:
acquiring current environment data of the surrounding environment of the controlled vehicle at the current moment;
obtaining the score of the lane changing action at the current moment according to a pre-obtained action evaluation model and the current environment data, wherein the lane changing action comprises acceleration and a steering angle;
determining the lane changing action corresponding to the highest score as the selected lane changing action at the current moment;
and controlling the controlled vehicle to act according to the selected lane changing action.
It should be noted that, in the present specification, the embodiments are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments may be referred to each other. The device disclosed by the embodiment corresponds to the method disclosed by the embodiment, so that the description is simple, and the relevant part can be referred to the method part for description.
It is further noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
The steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module may reside in Random Access Memory (RAM), memory, Read Only Memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.
The foregoing is merely a preferred embodiment of the invention and is not intended to limit the invention in any manner. Although the present invention has been described with reference to the preferred embodiments, it is not intended to be limited thereto. Those skilled in the art can make numerous possible variations and modifications to the present teachings, or modify equivalent embodiments to equivalent variations, without departing from the scope of the present teachings, using the methods and techniques disclosed above. Therefore, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention are still within the scope of the protection of the technical solution of the present invention, unless the contents of the technical solution of the present invention are departed.
Claims (10)
1. A lane change control method, comprising:
acquiring current environment data of the surrounding environment of the controlled vehicle at the current moment;
obtaining the score of the lane changing action at the current moment according to a pre-obtained action evaluation model and the current environment data, wherein the lane changing action comprises acceleration and a steering angle;
determining the lane changing action corresponding to the highest score as the selected lane changing action at the current moment;
controlling the controlled vehicle to act according to the selected lane changing action;
the method for establishing the action evaluation model specifically comprises the following steps:
acquiring a lane change action training set, wherein the lane change action training set comprises a plurality of groups of lane change data, and each group of lane change data comprises each lane change action executed by a driver controlling a first vehicle in a lane change process and vehicle speed data of the first vehicle and environmental data of the surrounding environment thereof when the lane change action is executed;
training an objective function based on a convolutional neural network pair according to the lane change action training set
Training is carried out, and after the training is converged, the action evaluation model q (s, a) is obtained;
wherein, a is the current lane changing action, s is the current environment data, a 'is the lane changing action at the next moment, s' is the environment data at the next moment, γ is the learning rate, and R(s) is the instant report at the current moment.
2. The lane change control method according to claim 1,
in the lane changing process, the instant return is positively correlated with the first comfort data and the second comfort data, and the instant return is also positively correlated with the first safety data and/or the second safety data;
wherein the first comfort data is inversely related to a degree of change of a lateral acceleration of the first vehicle over a preset time period; the second comfort data is inversely related to a degree of change of the longitudinal acceleration of the first vehicle over the preset time period; the first safety data are positively correlated with a first distance and the vehicle speed of a second vehicle, the second vehicle is in front of the first vehicle and is positioned in a lane to be turned, the first distance is the longitudinal distance between the second vehicle and the first vehicle, and the first safety data are also negatively correlated with the vehicle speed, the maximum braking deceleration and the reaction time delay of the first vehicle; the second safety data are positively correlated with a second distance and the vehicle speed of the first vehicle, the second distance is the longitudinal distance between a third vehicle and the first vehicle, the third vehicle is behind the first vehicle and is positioned in the lane to be turned, and the second safety data are also negatively correlated with the vehicle speed, the maximum braking deceleration and the reaction time delay of the third vehicle;
at the end of the lane change, the instant return is 100.
3. The lane change control method according to claim 2, wherein the instant report specifically comprises:
wherein, [ f ]0,f1]Is a lateral acceleration spectrum, [ f ]2,f3]Is a spectrum of longitudinal acceleration, axFor the lateral acceleration of the first vehicle within the predetermined time period, ayLongitudinal acceleration, y, of the first vehicle in the preset time period1Is the first distance, y2Is said second distance, v1Speed, v, of said first vehicle2Speed, v, of said second vehicle3Is the speed of the third vehicle,
4. The lane-change control method according to claim 1, wherein the obtaining a score of the lane-change action at the current time according to a pre-obtained action evaluation model and the current environment data specifically comprises:
inputting the current environment into the action evaluation model q (s, a) to obtain a scoring function q (a) of the lane-changing action at the current moment;
the determining the lane change action corresponding to the highest score as the selected lane change action at the current time specifically includes:
according to the formula
Determining the selected lane action aIs selected by。
5. A lane-change control apparatus, comprising: the device comprises an acquisition unit, a scoring unit, a determination unit and a control unit;
the acquisition unit is used for acquiring current environment data of the surrounding environment of the controlled vehicle at the current moment;
the scoring unit is used for obtaining the score of the lane changing action at the current moment according to a pre-obtained action evaluation model and the current environment data, wherein the lane changing action comprises acceleration and a steering angle;
the determining unit is used for determining the lane changing action corresponding to the highest score as the selected lane changing action at the current moment;
the control unit is used for controlling the controlled vehicle to act according to the selected lane changing action;
the device, still include: a model training unit; the model training unit is specifically configured to:
acquiring a lane change action training set, wherein the lane change action training set comprises a plurality of groups of lane change data, and each group of lane change data comprises each lane change action executed by a driver controlling a first vehicle in a lane change process and vehicle speed data of the first vehicle and environmental data of the surrounding environment thereof when the lane change action is executed;
training an objective function based on a convolutional neural network pair according to the lane change action training set
Training is carried out, and after the training is converged, the action evaluation model q (s, a) is obtained;
wherein, a is the current lane changing action, s is the current environment data, a 'is the lane changing action at the next moment, s' is the environment data at the next moment, γ is the learning rate, and R(s) is the instant report at the current moment.
6. The lane change control apparatus according to claim 5,
in the lane changing process, the instant return is positively correlated with the first comfort data and the second comfort data, and the instant return is also positively correlated with the first safety data and/or the second safety data;
wherein the first comfort data is inversely related to a degree of change of a lateral acceleration of the first vehicle over a preset time period; the second comfort data is inversely related to a degree of change of the longitudinal acceleration of the first vehicle over the preset time period; the first safety data are positively correlated with a first distance and the vehicle speed of a second vehicle, the second vehicle is in front of the first vehicle and is positioned in a lane to be turned, the first distance is the longitudinal distance between the second vehicle and the first vehicle, and the first safety data are also negatively correlated with the vehicle speed, the maximum braking deceleration and the reaction time delay of the first vehicle; the second safety data are positively correlated with a second distance and the vehicle speed of the first vehicle, the second distance is the longitudinal distance between a third vehicle and the first vehicle, the third vehicle is behind the first vehicle and is positioned in the lane to be turned, and the second safety data are also negatively correlated with the vehicle speed, the maximum braking deceleration and the reaction time delay of the third vehicle;
at the end of the lane change, the instant return is 100.
7. The apparatus according to claim 6, wherein the real-time report comprises:
wherein, [ f ]0,f1]Is a lateral acceleration spectrum, [ f ]2,f3]Is a spectrum of longitudinal acceleration, axFor the lateral acceleration of the first vehicle within the predetermined time period, ayLongitudinal acceleration, y, of the first vehicle in the preset time period1Is the first distance, y2Is said second distance, v1Speed, v, of said first vehicle2Speed, v, of said second vehicle3Is the speed of the third vehicle,
8. The lane change control device according to claim 5, wherein the scoring unit is specifically configured to:
inputting the current environment into the action evaluation model q (s, a) to obtain a scoring function q (a) of the lane-changing action at the current moment;
the determining the lane change action corresponding to the highest score as the selected lane change action at the current time specifically includes:
according to the formula
Determining the selected lane action aIs selected by。
9. A computer-readable storage medium, characterized in that a computer program is stored thereon for performing the method of any of claims 1-4.
10. A vehicle control unit, comprising: a memory and a processor;
the memory for storing a computer program for executing the method of any one of claims 1-4 by the processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711064941.1A CN109760681B (en) | 2017-11-02 | 2017-11-02 | Channel changing control method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711064941.1A CN109760681B (en) | 2017-11-02 | 2017-11-02 | Channel changing control method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109760681A CN109760681A (en) | 2019-05-17 |
| CN109760681B true CN109760681B (en) | 2020-09-29 |
Family
ID=66448667
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711064941.1A Active CN109760681B (en) | 2017-11-02 | 2017-11-02 | Channel changing control method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109760681B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112124310B (en) * | 2019-06-24 | 2021-12-24 | 魔门塔(苏州)科技有限公司 | Vehicle path transformation method and device |
| CN111081046B (en) * | 2020-01-03 | 2022-01-14 | 阿波罗智能技术(北京)有限公司 | Method, device, electronic equipment and medium for automatically changing lane of driving vehicle |
| WO2021189210A1 (en) * | 2020-03-23 | 2021-09-30 | 华为技术有限公司 | Vehicle lane changing method and related device |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7363140B2 (en) * | 2004-06-17 | 2008-04-22 | Robert Bosch Gmbh | Lane changing assistant for motor vehicles |
| CN106874597A (en) * | 2017-02-16 | 2017-06-20 | 北理慧动（常熟）车辆科技有限公司 | A kind of highway passing behavior decision-making technique for being applied to automatic driving vehicle |
| CN107134173A (en) * | 2017-06-15 | 2017-09-05 | 长安大学 | A kind of lane-change early warning system and method recognized with driving habit |
| CN107161155A (en) * | 2017-04-27 | 2017-09-15 | 大连理工大学 | A kind of vehicle collaboration lane-change method and its system based on artificial neural network |
| CN107215339A (en) * | 2017-06-26 | 2017-09-29 | 地壳机器人科技有限公司 | The lane-change control method and device of automatic driving vehicle |
-
2017
- 2017-11-02 CN CN201711064941.1A patent/CN109760681B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7363140B2 (en) * | 2004-06-17 | 2008-04-22 | Robert Bosch Gmbh | Lane changing assistant for motor vehicles |
| CN106874597A (en) * | 2017-02-16 | 2017-06-20 | 北理慧动（常熟）车辆科技有限公司 | A kind of highway passing behavior decision-making technique for being applied to automatic driving vehicle |
| CN107161155A (en) * | 2017-04-27 | 2017-09-15 | 大连理工大学 | A kind of vehicle collaboration lane-change method and its system based on artificial neural network |
| CN107134173A (en) * | 2017-06-15 | 2017-09-05 | 长安大学 | A kind of lane-change early warning system and method recognized with driving habit |
| CN107215339A (en) * | 2017-06-26 | 2017-09-29 | 地壳机器人科技有限公司 | The lane-change control method and device of automatic driving vehicle |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109760681A (en) | 2019-05-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113165652B (en) | Verifying predicted trajectories using a mesh-based approach | |
| EP3572896B1 (en) | Automatic driving method and device | |
| CN108919795B (en) | Automatic driving automobile lane change decision method and device | |
| CN108647638B (en) | Vehicle position detection method and device | |
| Min et al. | Deep Q learning based high level driving policy determination | |
| US20200086861A1 (en) | Systems and methods for predicting vehicle trajectory | |
| Okumura et al. | Challenges in perception and decision making for intelligent automotive vehicles: A case study | |
| CN113291308B (en) | Vehicle self-learning lane-changing decision-making system and method considering driving behavior characteristics | |
| CN110588653A (en) | Control system, control method and controller for autonomous vehicle | |
| CN111301419A (en) | Reinforcement learning based method for SAE4 level automated lane change | |
| US10919543B2 (en) | Learning method and learning device for determining whether to switch mode of vehicle from manual driving mode to autonomous driving mode by performing trajectory-based behavior analysis on recent driving route | |
| CN109760681B (en) | Channel changing control method and device | |
| US10885358B2 (en) | Method for detecting traffic signs | |
| KR20200101517A (en) | Method for autonomous cooperative driving based on vehicle-road infrastructure information fusion and apparatus for the same | |
| Jeon et al. | Rear vehicle tracking on a bicycle using active sensor orientation control | |
| CN107209997A (en) | The travel controlling system and travel control method of vehicle | |
| JP2016215790A (en) | Lane change plan generating device, lane change plan generating method | |
| Dong et al. | Smooth behavioral estimation for ramp merging control in autonomous driving | |
| US20200302645A1 (en) | Vehicle object tracking | |
| CN113071515B (en) | Movable carrier control method, device, movable carrier and storage medium | |
| CN109297502A (en) | Laser projection pointing method and device based on image procossing Yu GPS navigation technology | |
| KR101738425B1 (en) | Apparatus and method for setting region of interest for detecting traffic lanes | |
| CN112542061A (en) | Lane borrowing and overtaking control method, device and system based on Internet of vehicles and storage medium | |
| CN110608752A (en) | Vehicle interaction method and device, vehicle and computer-readable storage medium | |
| CN116343174A (en) | Target detection method, device, vehicle and storage medium |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |