CN114056425A - Vehicle steering control method and device, vehicle and storage medium - Google Patents
Vehicle steering control method and device, vehicle and storage medium Download PDFInfo
- Publication number
- CN114056425A CN114056425A CN202111395577.3A CN202111395577A CN114056425A CN 114056425 A CN114056425 A CN 114056425A CN 202111395577 A CN202111395577 A CN 202111395577A CN 114056425 A CN114056425 A CN 114056425A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- steering
- sample
- sample data
- opposite
- 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.)
- Granted
Links
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B62—LAND VEHICLES FOR TRAVELLING OTHERWISE THAN ON RAILS
- B62D—MOTOR VEHICLES; TRAILERS
- B62D6/00—Arrangements for automatically controlling steering depending on driving conditions sensed and responded to, e.g. control circuits
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Chemical & Material Sciences (AREA)
- Combustion & Propulsion (AREA)
- Transportation (AREA)
- Mechanical Engineering (AREA)
- Traffic Control Systems (AREA)
Abstract
The present disclosure relates to a vehicle steering control method, apparatus, vehicle, and storage medium, the method comprising: the method comprises the steps of obtaining detection data of a vehicle to be steered at the current intersection, wherein the detection data comprises information of opposite vehicles, and controlling steering operation of the vehicle to be steered at the current intersection based on the detection data and a pre-trained steering decision model.
Description
Technical Field
The present disclosure relates to the field of autonomous driving vehicles, and in particular, to a method and an apparatus for controlling vehicle steering, a vehicle, and a storage medium.
Background
In recent years, the technology of the automatic driving automobile is rapidly developed, and is also closely related to the daily life of people in the future. In the automatic driving process of the automatic driving automobile, a steering decision is one of the most difficult key problems to solve, such as left turning. How to realize efficient and safe left turning at the intersection according to some information of the oncoming vehicle has important significance for the whole automatic driving system.
Disclosure of Invention
In order to achieve the above object, the present disclosure provides a vehicle steering control method, device, vehicle and storage medium, which are used for solving the problem of steering decision during automatic driving of the vehicle.
According to a first aspect of an embodiment of the present disclosure, there is provided a vehicle steering control method including:
acquiring detection data of a vehicle to be steered at a current intersection, wherein the detection data comprises information of an opposite vehicle of the vehicle to be steered at the current intersection, and the opposite vehicle is a vehicle which runs opposite to the vehicle to be steered on an opposite lane of the vehicle to be steered and is closest to the vehicle to be steered;
controlling the turning operation of the vehicle to be turned at the current intersection based on the detection data and a pre-trained turning decision model; the steering decision model is obtained by training a plurality of groups of sample data of a sample vehicle at a target intersection, wherein the sample data comprises information of the vehicle opposite to the sample vehicle at the target intersection.
Optionally, before the acquiring the detection data of the vehicle to be steered at the current intersection, the method comprises: collecting the multiple groups of sample data of the sample vehicle at the target intersection; training the steering decision model based on the plurality of sets of sample data.
Optionally, the vehicle information of the oncoming vehicle includes vehicle speed, vehicle position, vehicle acceleration information of the oncoming vehicle; training the steering decision model based on the plurality of sets of sample data comprises: calculating the collision time of the opposite vehicle to the position of the sample vehicle based on the vehicle speed, the vehicle position and the vehicle acceleration information of the opposite vehicle; training a steering decision model based on the time-to-collision.
Optionally, the sample data includes first sample data and second sample data, where the first sample data is sample data in a scene where the sample vehicle is suggested to turn to a first direction, and the second sample data is sample data in a scene where the sample vehicle is not suggested to turn to the first direction.
Optionally, the vehicle information of the oncoming vehicle further includes a vehicle number of the oncoming vehicle; training a steering decision model based on the sample data comprises: calculating a turn probability of the sample vehicle based on a first fault tolerance coefficient and the first sample data; the turn probability characterizes a probability that the sample vehicle can turn in the first direction; the first fault-tolerant coefficient is an offset coefficient determined based on a training result of the first sample data; the calculation formula of the steering probability comprises:
wherein λ is the first fault-tolerance coefficient, and n is the number of vehicles; ttcnThe collision time of the subtended vehicle n in the first sample data to the current position of the sample vehicle; u. ofnThe average value of the collision time of the oncoming vehicle on an oncoming lane n in the first sample data, the oncoming lane n being an oncoming lane in which the oncoming vehicle n is located; sigman 2Is the variance of the collision time on the opposite lane n in the first sample data;
calculating a non-steering probability of the sample vehicle based on a second fault-tolerance coefficient and the second sample data; the non-steering probability characterizes a probability that the sample vehicle cannot steer in the first direction; the second fault-tolerant coefficient is an offset coefficient determined based on a training result of the second sample data; the calculation formula of the non-steering probability comprises:
wherein λ' is the second fault tolerance coefficient, and n is the number of vehicles; ttc'nThe arrival place of the opposite vehicle n in the second sample dataA time of collision of the current location of the sample vehicle; u'nThe average value of the collision time of the opposite vehicle on an opposite lane n in the second sample data, wherein the opposite lane n is the opposite lane where the opposite vehicle n is located; sigma'n 2The variance of the collision time on the opposite lane n in the second sample data is obtained; and respectively updating the first fault-tolerant coefficient and the second fault-tolerant coefficient according to the steering probability and the non-steering probability, and performing the steps of calculating the steering probability of the sample vehicle based on the first fault-tolerant coefficient and the first sample data and calculating the non-steering probability of the sample vehicle based on the second fault-tolerant coefficient and the second sample data again by using the updated first fault-tolerant coefficient and the updated second fault-tolerant coefficient until a preset training condition is met to obtain the steering decision model.
Optionally, the vehicle information of the oncoming vehicle includes a vehicle number and a vehicle type of the oncoming vehicle; training a steering decision model based on the sample data comprises: calculating the steering probability of the vehicle to be steered based on the collision risk coefficient, the first fault-tolerant coefficient and the first sample data; the turn probability characterizes a probability that the sample vehicle can turn in the first direction; the collision risk coefficient is a constant determined based on the vehicle type and the vehicle speed; the first fault-tolerant coefficient is an offset coefficient determined based on a training result of the first sample data; the calculation formula of the steering probability comprises:
wherein, TnThe collision risk coefficient of the opposite vehicle n is lambda, the first fault-tolerant coefficient is obtained, and n is the number of the vehicles; ttcnThe collision time of the subtended vehicle n in the first sample data to the current position of the sample vehicle; u. ofnIs an average value of collision times of the oncoming vehicles on an oncoming lane n in the first sample data, the oncoming lane n being the oncoming vehicleAn opposite lane in which the vehicle n is located; sigman 2Is the variance of the collision time on the oncoming lane n in the first sample data.
Optionally, the step of controlling the steering operation of the vehicle to be steered at the current intersection based on the detection data and a pre-trained steering decision model comprises: acquiring information of a target opposite vehicle of the vehicle to be steered at the current intersection; the target opposite vehicle is a vehicle which runs opposite to the vehicle to be steered on an opposite lane of the vehicle to be steered, and the information of the target opposite vehicle comprises the number of vehicles, the type of the vehicles, the speed of the vehicles, the position of the vehicles and the acceleration information of the vehicles; calculating the steering probability and the non-steering probability of the vehicle to be steered in a first direction by utilizing the steering decision model based on the information of the target opposite vehicle; and controlling the steering operation of the vehicle to be steered according to the steering probability and the non-steering probability of the vehicle to be steered.
According to a second aspect of the embodiments of the present disclosure, there is provided a vehicle steering control apparatus, the apparatus including: the system comprises an acquisition module, a control module and a display module, wherein the acquisition module is used for acquiring detection data of a vehicle to be steered at a current intersection, the detection data comprises information of an opposite vehicle of the vehicle to be steered at the current intersection, and the opposite vehicle is a vehicle which runs opposite to the vehicle to be steered on an opposite lane of the vehicle to be steered;
the processing module is used for controlling the steering operation of the vehicle to be steered at the current intersection based on the detection data and a pre-trained steering decision model; the steering decision model is obtained by training a plurality of groups of sample data of a sample vehicle at a target intersection, wherein the sample data comprises information of the vehicle opposite to the sample vehicle at the target intersection.
According to a third aspect of embodiments of the present disclosure, there is provided a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the aforementioned vehicle steering control method.
According to a fourth aspect of the embodiments of the present disclosure, there is provided a vehicle including: a memory having a computer program stored thereon; a processor for executing the computer program in the memory to implement the steps of the aforementioned vehicle steering control method.
The technical scheme provided by the embodiment of the disclosure can have the following beneficial effects: the method and the device can acquire the detection data of the vehicle to be steered at the current intersection, the detection data comprises the information of the opposite vehicles, and the steering operation of the vehicle to be steered at the current intersection is controlled based on the detection data and the pre-trained steering decision model, so that whether the automatic driving automobile can perform the steering operation or not can be effectively judged, and the driving safety of the automatic driving system is improved.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure.
Drawings
The accompanying drawings, which are included to provide a further understanding of the disclosure and are incorporated in and constitute a part of this specification, illustrate embodiments of the disclosure and together with the description serve to explain the disclosure without limiting the disclosure. In the drawings:
fig. 1 is a flowchart illustrating a vehicle steering control method according to an exemplary embodiment of the present disclosure.
Fig. 2 is a scene schematic diagram of a sample vehicle at a target intersection, shown in an exemplary embodiment of the present disclosure.
Fig. 3 is a flowchart illustrating sub-steps of step S202 according to an exemplary embodiment of the present disclosure.
Fig. 4 is a flowchart illustrating sub-steps of step S2022 according to an exemplary embodiment of the present disclosure.
Fig. 5 is a flowchart illustrating sub-steps of step S204 according to an exemplary embodiment of the present disclosure.
Fig. 6 is a block diagram of a vehicle steering control apparatus shown in an exemplary embodiment of the present disclosure.
Fig. 7 is a block diagram of a vehicle shown in an exemplary embodiment of the present disclosure.
Description of the reference numerals
10-an oncoming vehicle; 20-an oncoming vehicle; 30-sample vehicle; 200-vehicle steering control means; 203-an obtaining module; 205-a processing module; 1900-vehicle; 1922-a processor; 1932-memory; 1926-power supply components; 1950-a communication component; 1958-input/output interface.
Detailed Description
The following detailed description of specific embodiments of the present disclosure is provided in connection with the accompanying drawings. It should be understood that the detailed description and specific examples, while indicating the present disclosure, are given by way of illustration and explanation only, not limitation.
When the following description refers to the accompanying drawings, like numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the exemplary embodiments below are not intended to represent all implementations consistent with the present disclosure. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present disclosure, as detailed in the appended claims.
Referring to fig. 1, fig. 1 is a flowchart illustrating a method for controlling steering of a vehicle according to an exemplary embodiment of the present disclosure, where the method is applicable to an electronic device, which may be an electronic device in a vehicle, which may be an autonomous vehicle. The vehicle steering control method shown in fig. 1 includes the steps of:
in step S201, a plurality of sets of sample data of the sample vehicle at the target intersection are collected.
Referring to fig. 2, fig. 2 is a schematic view of a scene of a sample vehicle at a target intersection according to an exemplary embodiment of the disclosure; in fig. 2, a sample vehicle waits for a left turn at an intersection, and can be driven by a driver with abundant driving experiences to acquire sample data, and more than one driver with abundant driving experiences, such as two or three drivers, can be allocated on the sample vehicle (automatic driving vehicle) to improve the accuracy of the sample data; the driver judges whether the vehicle can turn left according to the information of the opposite vehicles on opposite lanes, wherein the opposite vehicles are vehicles which run opposite to the sample vehicle and are closest to the sample vehicle on each opposite lane, as shown in fig. 2, the opposite vehicles comprise opposite vehicles 10 and opposite vehicles 20, the opposite vehicle 10 is the vehicle which is closest to the sample vehicle 30 on the lane where the opposite vehicle 10 is located, and the opposite vehicle 20 is the vehicle which is closest to the sample vehicle 30 on the lane where the opposite vehicle 20 is located; only two oncoming vehicles are shown in fig. 2, but in one embodiment the oncoming vehicles may also be one, three, etc. The driver-driven sample vehicle 30 can acquire the information of the oncoming vehicle, which includes the number of oncoming vehicles, the vehicle type, the vehicle speed, the vehicle position, the vehicle acceleration information. The information of the oncoming vehicles may be acquired by a networking device or a sensor such as a radar or a camera, which is not specifically required by the present disclosure.
The sample data includes, but is not limited to, information of the oncoming vehicle of the sample vehicle. Sample tags for the sample data are recorded while the sample data is being acquired (the sample tags may be determined based on the driver's judgment), the sample tags including a recommendation that the sample vehicle turn in a first direction and a non-recommendation that the sample vehicle turn in the first direction, which in one embodiment is a left turn direction.
It should be noted that, in this step, there may be a plurality of sample vehicles and a plurality of target intersections, for example, one sample vehicle may go to a plurality of target intersections to acquire a plurality of sets of sample data, or a plurality of sample vehicles may go to the same target intersection to acquire a plurality of sets of sample data, or one sample vehicle may repeatedly acquire a plurality of sets of sample data at one target intersection, and the plurality of sets of sample data may be acquired by any one or more of the above manners as needed, which is not limited in this disclosure.
In step S202, a steering decision model is trained based on multiple sets of sample data.
It should be noted that, since the driver determines whether the vehicle can turn left according to the information of the oncoming vehicle on the oncoming lane, first sample data and second sample data are obtained, where the first sample data is sample data in a scene in which the sample vehicle is suggested to turn to the first direction, and the second sample data is sample data in a scene in which the sample vehicle is not suggested to turn to the first direction, and in one embodiment, the first direction is a left-turn direction; in step S202, the same number of first sample data and second sample data are selected for training the steering decision model.
Step S202 includes sub-steps S2021 and S2022, and the specific manner of training the steering decision model will be described in detail in the sub-step of step S202. Referring to fig. 3, fig. 3 is a flowchart illustrating sub-steps of step S202 according to an exemplary embodiment of the present disclosure.
In sub-step S2021, the collision time at which the oncoming vehicle reaches the present position of the sample vehicle is calculated based on the vehicle speed, the vehicle position, and the vehicle acceleration information of the oncoming vehicle.
And calculating the Time-To-Collision (TTC) of the opposite vehicle reaching the position of the sample vehicle according To the selected first sample data and the second sample data. Specifically, a first collision time ttc of the oncoming vehicle to the location of the sample vehicle is calculated from the vehicle speed, the vehicle position, and the vehicle acceleration information of the oncoming vehicle in the first sample datanAnd calculating a second collision time ttc 'when the opposite vehicle reaches the position of the sample vehicle according to the vehicle speed, the vehicle position and the vehicle acceleration information of the opposite vehicle in the second sample data'n。
According to the first collision time ttcnAnd a second time to collision ttc'nThe turning probability and the non-turning probability of the sample vehicle can be calculated, the turning probability represents the probability that the sample vehicle can turn towards the first direction, the non-turning probability represents the probability that the sample vehicle cannot turn towards the first direction, and the first direction is the left-turning direction in one embodiment. For example, referring to fig. 2, the turning probability represents the probability that the sample vehicle 30 can make a left turn in the current states of the oncoming vehicle 10 and the oncoming vehicle 20, and the non-turning probability represents the probability that the sample vehicle 30 cannot make a left turn in the current states of the oncoming vehicle 10 and the oncoming vehicle 20.
The steering probability p can be obtained from the first sample data1The calculation formula (2) includes:
wherein n is the number of vehicles in the opposite direction; ttcnAs the collision time of the oncoming vehicle n in the first sample data to the current position of the sample vehicle, for example, refer to fig. 2, in which there are two oncoming vehicles, ttc, in fig. 21May represent the time to collision, ttc, of the subject vehicle 10 to the current location of the sample vehicle2May represent the time of collision for the subject vehicle 20 to reach the current position of the sample vehicle, when the value of n is 2; u. ofnThe average value of the collision time of the oncoming vehicle on the oncoming lane n in the first sample data is obtained, wherein the oncoming lane n is the oncoming lane in which the oncoming vehicle n is located; sigman 2Is the variance of the collision time on the oncoming lane n in the first sample data.
According to the second sample data, the non-steering probability p can be obtained2The calculation formula (2) includes:
wherein n is the number of vehicles in the opposite direction; ttc'nFor the time to collision of the oncoming vehicle n to the current location of the sample vehicle in the second sample data, for example, refer to fig. 2, where there are two oncoming vehicles, ttc'1The time to collision, ttc ', of the arrival of the subject vehicle 10 at the current location of the sample vehicle may be represented'2May represent the time of collision for the subject vehicle 20 to reach the current position of the sample vehicle, when the value of n is 2; u'nThe average value of the collision time of the oncoming vehicle on the oncoming lane n in the second sample data is obtained, wherein the oncoming lane n is the oncoming lane in which the oncoming vehicle n is located; sigma'n 2Is the variance of the collision time on the oncoming lane n in the second sample data.
In sub-step S2022, a steering decision model is trained based on the time of collision.
For an automatic driving system, driving safety must be the first place, so that an established steering decision model can tolerate the condition that 'the vehicle can turn left' is wrongly judged as 'the vehicle can not turn left', but cannot tolerate the condition that 'the vehicle can not turn left' is wrongly judged as 'the vehicle can turn left', namely, the problem is solved based on cost-sensitive classification, and the automatic driving safety is improved.
In one embodiment, the present disclosure introduces a cost-characterizing fault tolerance coefficient for biasing the output results of the steering decision model, e.g., biasing the result of "vehicle can turn left" toward "vehicle cannot turn left".
Step S2022 includes sub-steps S20221 and S20222, and the specific manner of training the steering decision model will be described in detail in the sub-step of step S2022. Referring to fig. 4, fig. 4 is a flowchart illustrating sub-steps of step S2022 according to an exemplary embodiment of the present disclosure.
In sub-step S20221, a turning probability of the sample vehicle is calculated based on the first fault-tolerance coefficient and the first sample data.
Illustratively, a first fault-tolerant coefficient is introduced, and the turning probability of the sample vehicle is calculated based on the first sample data, wherein the first fault-tolerant coefficient is an offset coefficient determined based on the training result of the first sample data; at this time, the calculation formula of the steering probability includes:
wherein, λ is a first fault-tolerant coefficient and its value range is 0<λ<1，n、ttcn、unAnd sigman 2The meaning of (a) is the same as that described in the foregoing embodiment, and the foregoing embodiment may be specifically referred to, and is not repeated herein.
In sub-step S20222, a non-turning probability of the sample vehicle is calculated based on the second fault tolerance coefficient and the second sample data.
Illustratively, a second fault-tolerant coefficient is introduced, the turning probability of the sample vehicle is calculated based on second sample data, and the second fault-tolerant coefficient is an offset coefficient determined based on a training result of the second sample data; in this case, the calculation formula of the non-steering probability includes:
wherein λ 'is a second fault-tolerant coefficient, and the value range is λ'>1，n、ttc′n、u′nAnd sigma'n 2The meaning of (a) is the same as that described in the foregoing embodiment, and the foregoing embodiment may be specifically referred to, and is not repeated herein.
And respectively updating the first fault-tolerant coefficient and the second fault-tolerant coefficient according to the steering probability and the non-steering probability, executing the substep S20221 again by using the updated first fault-tolerant coefficient and the updated second fault-tolerant coefficient to calculate the steering probability of the sample vehicle, and executing the substep S20222 again to calculate the non-steering probability of the sample vehicle until the preset training condition is met to obtain a steering decision model. The preset training condition is a preset reasonable value in the field, and aims to enable the accuracy of the steering decision model to reach a preset value, and can be obtained based on experience.
It should be noted that, the first fault-tolerant coefficient and the second fault-tolerant coefficient are updated to shift the boundary of the steering decision model to the direction of "the vehicle cannot turn left", that is, to reduce the value of the steering probability and increase the value of the non-steering probability, so as to reduce the risk of erroneously determining that the sample vehicle "the vehicle cannot turn left" as "the vehicle can turn left".
In addition, for the automatic driving system, the magnitude of the risk brought by the vehicle collision needs to be considered, for example, in the case that the opposite vehicle is a common small vehicle and the opposite vehicle is a large vehicle and the opposite vehicle is at a high speed, the result of the automatic driving vehicle after the collision is completely different, and for this reason, the present disclosure designs another training method of the steering decision model based on the severity of the result which may be caused by the vehicle collision, which will be described in detail below.
In the embodiment, a collision risk coefficient T for representing the collision risk grade of the vehicle is introducednThe collision risk coefficient is a constant determined based on the type of vehicle and the vehicle speed. Firstly, the collision danger grades are classified according to the severity of the vehicle collision consequences:
for example, when the oncoming vehicle is a normal small-sized vehicle and is traveling at a traveling speed lower than a predetermined threshold value, the collision risk level is 4; when the opposite vehicle is a common small vehicle and the running speed is higher than the preset threshold value, the collision danger level is 3 grade; when the opposite vehicle is a large vehicle and the running speed is lower than the preset threshold value, the collision danger level is 2 grade; when the oncoming vehicle is a large vehicle and the running speed is higher than the predetermined threshold value, the collision risk level is level 1. In one embodiment, the predetermined threshold may be, but is not limited to, 50 kilometers per hour. Referring to fig. 2, the oncoming vehicle in fig. 2 is used for explanation, the oncoming lane in which the oncoming vehicle 10 is located is the adjacent oncoming lane of the sample vehicle, and for the oncoming vehicles on the adjacent oncoming lane, when the collision risk levels are 1 level, 2 level, 3 level, and 4 level, respectively, the corresponding T isnValues of (a) may be, but are not limited to, 0.9, 0.6, 0.3, 0; the opposite lane in which the opposite vehicle 20 is located is a far-adjacent opposite lane of the sample vehicle, because the opposite lane in which the opposite vehicle 20 is located is separated from the lane in which the sample vehicle is located by one lane, for the opposite vehicles on the far-adjacent opposite lane, when the collision risk levels are 1 level, 2 levels, 3 levels and 4 levels respectively, the corresponding T isnCan be, but is not limited to, 1.2, 0.8, 0.4, 0.
The calculation formula of the steering probability is further optimized based on the collision risk coefficient, and the calculation formula of the steering probability of the vehicle to be steered based on the collision risk coefficient, the first fault-tolerant coefficient and the first sample data comprises the following steps:
wherein, Tnλ, n, ttc as the collision risk coefficient of the oncoming vehicle nn、unAnd sigman 2The same as in the previous embodimentThe meanings are consistent, and reference may be made to the foregoing embodiments in particular, which are not described herein again.
The collision risk coefficient can be obtained according to the behavior data and habits of the user in driving the vehicle at ordinary times. For example, for some users who are very cautious to drive at ordinary times, the value of the collision risk coefficient can be increased appropriately.
The optimized steering decision model changes the random variable ttcnSubject to mathematically expected positions by increasing the safety time TnThe safety of personnel on the autonomous vehicle can be further ensured.
For example, before optimizing the steering decision model, when the collision risk level of the adjacent opposite lanes is 1-4, it can be determined that the vehicle turns left assuming that the critical collision time TTC is 5 seconds; after the adjustment, when the collision grade of the adjacent opposite lane is 2 grade, the vehicle is judged to turn left until the critical collision time TTC reaches 5.6 seconds, and when the collision grade of the adjacent opposite lane is 1 grade, the vehicle is judged to turn left until the critical collision time TTC reaches 5.9 seconds. Obviously, the optimized steering decision model can be adjusted according to the severity of the consequences caused by vehicle collision so as to improve the driving safety.
In step S203, detection data of the vehicle to be steered at the current intersection is acquired.
In this step, the detection data is information of the oncoming vehicle of the vehicle to be steered at the current intersection, the information of the oncoming vehicle including the number of vehicles of the oncoming vehicle, the vehicle type, the vehicle speed, the vehicle position, the vehicle acceleration information.
In step S204, based on the detection data and the pre-trained steering decision model, the steering operation of the vehicle to be steered at the current intersection is controlled.
The step S204 includes a substep S2041, a substep S2042 and a substep S2043, and a specific manner of controlling the steering operation of the vehicle to be steered at the current intersection will be described in detail in the substep of the step S204. Referring to fig. 5, fig. 5 is a flowchart illustrating sub-steps of step S204 according to an exemplary embodiment of the present disclosure.
In sub-step S2041, information of a target oncoming vehicle of the vehicle to be steered at the current intersection is acquired.
The target counter vehicle is a vehicle which runs opposite to the vehicle to be steered on each counter lane and is closest to the vehicle to be steered, and the information of the target counter vehicle comprises the number of vehicles, the type of vehicles, the speed of the vehicles, the position of the vehicles and the acceleration information of the vehicles. The information of the target oncoming vehicle is obtained from the detection data acquired in step S203.
In sub-step S2042, based on the information of the target oncoming vehicle, a turning probability and a non-turning probability of the vehicle to be turned in the first direction are calculated using the turning decision model.
Information of the target opposite vehicle is input into any one of the trained steering decision models in step S202 to calculate a steering probability and a non-steering probability of the vehicle to be steered in a first direction, which may be a left-turn direction in one embodiment.
In sub-step S2043, the steering operation of the vehicle to be steered is controlled according to the steering probability and the non-steering probability of the vehicle to be steered.
And controlling the vehicle to be steered to steer to a first direction when the steering probability of the vehicle to be steered is greater than the non-steering probability, wherein the first direction can be a left-turning direction in one embodiment.
And when the steering probability of the vehicle to be steered is less than or equal to the non-steering probability, controlling the vehicle to be steered not to perform steering operation.
In summary, the present disclosure can obtain the detection data of the vehicle to be steered at the current intersection, where the detection data includes information of oncoming vehicles, and control the steering operation of the vehicle to be steered at the current intersection based on the detection data and a pre-trained steering decision model, so as to effectively determine whether the autonomous vehicle can perform the steering operation, thereby improving the driving safety of the autonomous driving system. The boundary of the steering decision model is adjusted through the fault-tolerant coefficient, so that the fault-tolerant rate of the automatic driving system is improved; the driving safety is improved by adjusting the collision danger coefficient based on the vehicle type.
Fig. 6 is a block diagram of a vehicle steering control apparatus shown in an exemplary embodiment of the present disclosure. Referring to fig. 6, the apparatus 20 includes an acquisition module 203 and a processing module 205.
The obtaining module 203 is configured to obtain detection data of a vehicle to be steered at a current intersection, where the detection data includes information of an oncoming vehicle of the vehicle to be steered at the current intersection, and the oncoming vehicle is a vehicle that runs on an oncoming lane of the vehicle to be steered in a direction opposite to the vehicle to be steered.
The processing module 205 is configured to control a steering operation of the vehicle to be steered at the current intersection based on the detection data and a pre-trained steering decision model; the steering decision model is obtained by training a plurality of groups of sample data of a sample vehicle at a target intersection, wherein the sample data comprises information of the vehicle opposite to the sample vehicle at the target intersection.
Optionally, the obtaining module 203 is further configured to collect the multiple sets of sample data of the sample vehicle at the target intersection.
Optionally, the processing module 205 is further configured to train the steering decision model based on the plurality of sets of sample data.
Optionally, the processing module 205 includes:
and the time calculation submodule is used for calculating the collision time of the opposite vehicle reaching the position of the sample vehicle based on the vehicle speed, the vehicle position and the vehicle acceleration information of the opposite vehicle.
And the model training submodule is used for training a steering decision model based on the collision time.
Optionally, the model training sub-module is further configured to calculate a turning probability of the sample vehicle based on a first fault-tolerance coefficient and the first sample data; the turn probability characterizes a probability that the sample vehicle can turn in the first direction; the first fault-tolerant coefficient is an offset coefficient determined based on a training result of the first sample data;
the calculation formula of the steering probability comprises:
wherein λ is the first fault-tolerance coefficient, and n is the number of vehicles; ttcnThe collision time of the subtended vehicle n in the first sample data to the current position of the sample vehicle; u. ofnThe average value of the collision time of the oncoming vehicle on an oncoming lane n in the first sample data, the oncoming lane n being an oncoming lane in which the oncoming vehicle n is located; sigman 2Is the variance of the collision time on the oncoming lane n in the first sample data.
The model training submodule is also used for calculating the non-steering probability of the sample vehicle based on a second fault-tolerant coefficient and the second sample data; the non-steering probability characterizes a probability that the sample vehicle cannot steer in the first direction; the second fault-tolerant coefficient is an offset coefficient determined based on a training result of the second sample data;
the calculation formula of the non-steering probability comprises:
wherein λ' is the second fault tolerance coefficient, and n is the number of vehicles; ttc'nThe collision time of the opposite vehicle n reaching the current position of the sample vehicle in the second sample data is obtained; u'nThe average value of the collision time of the opposite vehicle on an opposite lane n in the second sample data, wherein the opposite lane n is the opposite lane where the opposite vehicle n is located; sigma'n 2Is the variance of the collision time on the oncoming lane n in the second sample data.
And the model training submodule is further used for respectively updating the first fault-tolerant coefficient and the second fault-tolerant coefficient according to the steering probability and the non-steering probability, and performing the steps of calculating the steering probability of the sample vehicle based on the first fault-tolerant coefficient and the first sample data and calculating the non-steering probability of the sample vehicle based on the second fault-tolerant coefficient and the second sample data again by using the updated first fault-tolerant coefficient and the updated second fault-tolerant coefficient until a preset training condition is met, so that the steering decision model is obtained.
Optionally, the model training submodule is further configured to calculate a steering probability of the vehicle to be steered based on a collision risk coefficient, a first fault-tolerant coefficient and the first sample data; the turn probability characterizes a probability that the sample vehicle can turn in the first direction; the collision risk coefficient is a constant determined based on the vehicle type and the vehicle speed; the first fault-tolerant coefficient is an offset coefficient determined based on a training result of the first sample data;
the calculation formula of the steering probability comprises:
wherein, TnThe collision risk coefficient of the opposite vehicle n is lambda, the first fault-tolerant coefficient is obtained, and n is the number of the vehicles; ttcnThe collision time of the subtended vehicle n in the first sample data to the current position of the sample vehicle; u. ofnThe average value of the collision time of the oncoming vehicle on an oncoming lane n in the first sample data, the oncoming lane n being an oncoming lane in which the oncoming vehicle n is located; sigman 2Is the variance of the collision time on the oncoming lane n in the first sample data.
With regard to the apparatus in the above-described embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment related to the method, and will not be elaborated here.
Optionally, the processing module 205 is further configured to obtain information of a target oncoming vehicle of the vehicle to be steered at the current intersection; the target opposite vehicle is a vehicle which runs opposite to the vehicle to be steered on an opposite lane of the vehicle to be steered, and the information of the target opposite vehicle comprises the number of vehicles, the type of the vehicles, the speed of the vehicles, the position of the vehicles and the acceleration information of the vehicles.
And the steering decision model is also used for calculating the steering probability and the non-steering probability of the vehicle to be steered in the first direction based on the information of the target opposite vehicle.
And the steering control system is also used for controlling the steering operation of the vehicle to be steered according to the steering probability and the non-steering probability of the vehicle to be steered.
Fig. 7 is a block diagram illustrating a vehicle 1900 according to an exemplary embodiment of the present disclosure, where the vehicle 1900 may be an autonomous vehicle. Referring to fig. 7, a vehicle 1900 includes a processor 1922, which may be one or more in number, and a memory 1932 for storing computer programs executable by the processor 1922. The computer program stored in memory 1932 may include one or more modules that each correspond to a set of instructions. Further, the processor 1922 may be configured to execute the computer program to perform the vehicle steering control method described above.
Additionally, the vehicle 1900 may also include a power component 1926 and a communication component 1950, the power component 1926 may be configured to perform power management of the vehicle 1900, and the communication component 1950 may be configured to enable communication, e.g., wired or wireless communication, of the vehicle 1900. In addition, the vehicle 1900 may also include an input/output (I/O) interface 1958. The vehicle 1900 may operate based on an operating system, such as Windows Server, stored in memory 1932TM，Mac OS XTM，UnixTM，LinuxTMAnd so on.
In another exemplary embodiment, a computer readable storage medium comprising program instructions which, when executed by a processor, implement the steps of the vehicle steering control method described above is also provided. For example, the non-transitory computer readable storage medium may be the memory 1932 described above that includes program instructions executable by the processor 1922 of the vehicle 1900 to perform the vehicle steering control method described above.
In another exemplary embodiment, a computer program product is also provided, which comprises a computer program executable by a programmable apparatus, the computer program having code portions for performing the vehicle steering control method described above when executed by the programmable apparatus.
The preferred embodiments of the present disclosure are described in detail with reference to the accompanying drawings, however, the present disclosure is not limited to the specific details of the above embodiments, and various simple modifications may be made to the technical solution of the present disclosure within the technical idea of the present disclosure, and these simple modifications all belong to the protection scope of the present disclosure.
It should be noted that, in the foregoing embodiments, various features described in the above embodiments may be combined in any suitable manner, and in order to avoid unnecessary repetition, various combinations that are possible in the present disclosure are not described again.
In addition, any combination of various embodiments of the present disclosure may be made, and the same should be considered as the disclosure of the present disclosure, as long as it does not depart from the spirit of the present disclosure.
Claims (10)
1. A vehicle steering control method characterized by comprising:
acquiring detection data of a vehicle to be steered at a current intersection, wherein the detection data comprises information of an opposite vehicle of the vehicle to be steered at the current intersection, and the opposite vehicle is a vehicle which runs opposite to the vehicle to be steered on an opposite lane of the vehicle to be steered and is closest to the vehicle to be steered;
controlling the turning operation of the vehicle to be turned at the current intersection based on the detection data and a pre-trained turning decision model; the steering decision model is obtained by training a plurality of groups of sample data of a sample vehicle at a target intersection, wherein the sample data comprises information of the vehicle opposite to the sample vehicle at the target intersection.
2. The method of claim 1, wherein prior to said obtaining detection data of a vehicle to be steered at a current intersection, the method comprises:
collecting the multiple groups of sample data of the sample vehicle at the target intersection;
training the steering decision model based on the plurality of sets of sample data.
3. The method according to claim 2, wherein the vehicle information of the oncoming vehicle includes vehicle speed, vehicle position, vehicle acceleration information of the oncoming vehicle;
training the steering decision model based on the plurality of sets of sample data comprises:
calculating the collision time of the opposite vehicle to the position of the sample vehicle based on the vehicle speed, the vehicle position and the vehicle acceleration information of the opposite vehicle;
training a steering decision model based on the time-to-collision.
4. The method of claim 3, wherein the sample data comprises an equal number of first sample data and second sample data, the first sample data being sample data in a scenario in which steering of the sample vehicle to a first direction is recommended, the second sample data being sample data in a scenario in which steering of the vehicle to be sampled to the first direction is not recommended.
5. The method according to claim 4, wherein the vehicle information of the oncoming vehicle further includes a vehicle number of the oncoming vehicle;
training a steering decision model based on the sample data comprises:
calculating a turn probability of the sample vehicle based on a first fault tolerance coefficient and the first sample data; the turn probability characterizes a probability that the sample vehicle can turn in the first direction; the first fault-tolerant coefficient is an offset coefficient determined based on a training result of the first sample data;
the calculation formula of the steering probability comprises:
wherein λ is the first fault-tolerance coefficient, and n is the number of vehicles; ttcnThe collision time of the subtended vehicle n in the first sample data to the current position of the sample vehicle; u. ofnThe average value of the collision time of the oncoming vehicle on an oncoming lane n in the first sample data, the oncoming lane n being an oncoming lane in which the oncoming vehicle n is located; sigman 2Is the variance of the collision time on the opposite lane n in the first sample data;
calculating a non-steering probability of the sample vehicle based on a second fault-tolerance coefficient and the second sample data; the non-steering probability characterizes a probability that the sample vehicle cannot steer in the first direction; the second fault-tolerant coefficient is an offset coefficient determined based on a training result of the second sample data;
the calculation formula of the non-steering probability comprises:
wherein λ' is the second fault tolerance coefficient, and n is the number of vehicles; ttc'nThe collision time of the opposite vehicle n reaching the current position of the sample vehicle in the second sample data is obtained; u'nThe average value of the collision time of the opposite vehicle on an opposite lane n in the second sample data, wherein the opposite lane n is the opposite lane where the opposite vehicle n is located;
the variance of the collision time on the opposite lane n in the second sample data is obtained;
and respectively updating the first fault-tolerant coefficient and the second fault-tolerant coefficient according to the steering probability and the non-steering probability, and performing the steps of calculating the steering probability of the sample vehicle based on the first fault-tolerant coefficient and the first sample data and calculating the non-steering probability of the sample vehicle based on the second fault-tolerant coefficient and the second sample data again by using the updated first fault-tolerant coefficient and the updated second fault-tolerant coefficient until a preset training condition is met to obtain the steering decision model.
6. The method according to claim 4, wherein the vehicle information of the oncoming vehicle includes a vehicle number and a vehicle type of the oncoming vehicle;
training a steering decision model based on the sample data comprises:
calculating the steering probability of the vehicle to be steered based on the collision risk coefficient, the first fault-tolerant coefficient and the first sample data; the turn probability characterizes a probability that the sample vehicle can turn in the first direction; the collision risk coefficient is a constant determined based on the vehicle type and the vehicle speed; the first fault-tolerant coefficient is an offset coefficient determined based on a training result of the first sample data;
the calculation formula of the steering probability comprises:
wherein, TnThe collision risk coefficient of the opposite vehicle n is lambda, the first fault-tolerant coefficient is obtained, and n is the number of the vehicles; ttcnThe collision time of the subtended vehicle n in the first sample data to the current position of the sample vehicle; u. ofnThe average value of the collision time of the oncoming vehicle on an oncoming lane n in the first sample data, the oncoming lane n being an oncoming lane in which the oncoming vehicle n is located; sigman 2Is the variance of the collision time on the opposite lane n in the first sample data;
and updating the collision risk coefficient according to the steering probability, and executing the step of calculating the steering probability of the sample vehicle based on the collision risk coefficient, the first fault-tolerant coefficient and the first sample data again by using the updated collision risk coefficient until a preset training condition is met to obtain the steering decision model.
7. The method of claim 1, wherein the step of controlling the steering operation of the vehicle to be steered at the current intersection based on the detection data and a pre-trained steering decision model comprises:
acquiring information of a target opposite vehicle of the vehicle to be steered at the current intersection; the target opposite vehicle is a vehicle which runs opposite to the vehicle to be steered on an opposite lane of the vehicle to be steered, and the information of the target opposite vehicle comprises the number of vehicles, the type of the vehicles, the speed of the vehicles, the position of the vehicles and the acceleration information of the vehicles;
calculating the steering probability and the non-steering probability of the vehicle to be steered in a first direction by utilizing the steering decision model based on the information of the target opposite vehicle;
and controlling the steering operation of the vehicle to be steered according to the steering probability and the non-steering probability of the vehicle to be steered.
8. A vehicle steering control device, characterized in that,
the system comprises an acquisition module, a control module and a display module, wherein the acquisition module is used for acquiring detection data of a vehicle to be steered at a current intersection, the detection data comprises information of an opposite vehicle of the vehicle to be steered at the current intersection, and the opposite vehicle is a vehicle which runs opposite to the vehicle to be steered on an opposite lane of the vehicle to be steered;
the processing module is used for controlling the steering operation of the vehicle to be steered at the current intersection based on the detection data and a pre-trained steering decision model; the steering decision model is obtained by training a plurality of groups of sample data of a sample vehicle at a target intersection, wherein the sample data comprises information of the vehicle opposite to the sample vehicle at the target intersection.
9. A non-transitory computer readable storage medium having stored thereon a computer program, characterized in that the program, when executed by a processor, implements the steps of the vehicle steering control method according to any one of claims 1-7.
10. A vehicle, characterized by comprising:
a memory having a computer program stored thereon;
a processor for executing the computer program in the memory to implement the steps of the vehicle steering control method of any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111395577.3A CN114056425B (en) | 2021-11-23 | 2021-11-23 | Vehicle steering control method and device, vehicle and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111395577.3A CN114056425B (en) | 2021-11-23 | 2021-11-23 | Vehicle steering control method and device, vehicle and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114056425A true CN114056425A (en) | 2022-02-18 |
| CN114056425B CN114056425B (en) | 2023-07-18 |
Family
ID=80275436
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111395577.3A Active CN114056425B (en) | 2021-11-23 | 2021-11-23 | Vehicle steering control method and device, vehicle and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114056425B (en) |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2010092424A (en) * | 2008-10-10 | 2010-04-22 | Denso Corp | Driving support apparatus and driving support method of automobile |
| JP2014241036A (en) * | 2013-06-11 | 2014-12-25 | 三菱電機株式会社 | Vehicle driving assist system |
| US10019011B1 (en) * | 2017-10-09 | 2018-07-10 | Uber Technologies, Inc. | Autonomous vehicles featuring machine-learned yield model |
| US20180297592A1 (en) * | 2017-04-12 | 2018-10-18 | Hrl Laboratories, Llc | Cognitive behavior prediction system for autonomous systems |
| US10235882B1 (en) * | 2018-03-19 | 2019-03-19 | Derq Inc. | Early warning and collision avoidance |
| CN109901574A (en) * | 2019-01-28 | 2019-06-18 | 华为技术有限公司 | Automatic Pilot method and device |
| CN110298122A (en) * | 2019-07-03 | 2019-10-01 | 北京理工大学 | Automatic driving vehicle urban intersection left-hand rotation decision-making technique based on conflict resolution |
| CN110723207A (en) * | 2019-09-23 | 2020-01-24 | 江苏大学 | Intelligent automobile model prediction steering controller based on model reconstruction and control method thereof |
| US20200125994A1 (en) * | 2018-10-17 | 2020-04-23 | Automotive Research & Testing Center | Intelligent driving method for passing intersections based on support vector machine and intelligent driving system thereof |
| CN112365710A (en) * | 2020-10-19 | 2021-02-12 | 南京航空航天大学 | Intelligent vehicle lane change decision-making method based on probability output model |
| CN113200086A (en) * | 2021-06-21 | 2021-08-03 | 盐城工学院 | Intelligent vehicle steering control system and control method thereof |
-
2021
- 2021-11-23 CN CN202111395577.3A patent/CN114056425B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2010092424A (en) * | 2008-10-10 | 2010-04-22 | Denso Corp | Driving support apparatus and driving support method of automobile |
| JP2014241036A (en) * | 2013-06-11 | 2014-12-25 | 三菱電機株式会社 | Vehicle driving assist system |
| US20180297592A1 (en) * | 2017-04-12 | 2018-10-18 | Hrl Laboratories, Llc | Cognitive behavior prediction system for autonomous systems |
| US10019011B1 (en) * | 2017-10-09 | 2018-07-10 | Uber Technologies, Inc. | Autonomous vehicles featuring machine-learned yield model |
| US10235882B1 (en) * | 2018-03-19 | 2019-03-19 | Derq Inc. | Early warning and collision avoidance |
| US20200125994A1 (en) * | 2018-10-17 | 2020-04-23 | Automotive Research & Testing Center | Intelligent driving method for passing intersections based on support vector machine and intelligent driving system thereof |
| CN109901574A (en) * | 2019-01-28 | 2019-06-18 | 华为技术有限公司 | Automatic Pilot method and device |
| CN110298122A (en) * | 2019-07-03 | 2019-10-01 | 北京理工大学 | Automatic driving vehicle urban intersection left-hand rotation decision-making technique based on conflict resolution |
| CN110723207A (en) * | 2019-09-23 | 2020-01-24 | 江苏大学 | Intelligent automobile model prediction steering controller based on model reconstruction and control method thereof |
| CN112365710A (en) * | 2020-10-19 | 2021-02-12 | 南京航空航天大学 | Intelligent vehicle lane change decision-making method based on probability output model |
| CN113200086A (en) * | 2021-06-21 | 2021-08-03 | 盐城工学院 | Intelligent vehicle steering control system and control method thereof |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114056425B (en) | 2023-07-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111506980B (en) | Method and device for generating traffic scene for virtual driving environment | |
| EP3533681B1 (en) | Method for detecting safety of driving behavior, apparatus and storage medium | |
| US11214280B2 (en) | Autonomous vehicle providing driver education | |
| US11940790B2 (en) | Safe hand-off between human driver and autonomous driving system | |
| KR102373430B1 (en) | Method and device for providing personalized and calibrated adaptive deep learning model for the user of an autonomous vehicle | |
| US10168697B2 (en) | Assistance for an autonomous vehicle using crowd-sourced responses | |
| Shin et al. | Randomized adversarial imitation learning for autonomous driving | |
| CN113291308A (en) | Vehicle self-learning lane-changing decision-making system and method considering driving behavior characteristics | |
| US20220204020A1 (en) | Toward simulation of driver behavior in driving automation | |
| US20220410933A1 (en) | Implementing manoeuvres in autonomous vehicles | |
| CN111666307A (en) | Unmanned safety judgment system for intuitive inference according to scene observation | |
| CN113085873A (en) | Method and device for acquiring driving strategy, computer equipment and storage medium | |
| WO2021028533A1 (en) | Method, device, medium, and vehicle for providing individual driving experience | |
| CN116331221A (en) | Driving assistance method, driving assistance device, electronic equipment and storage medium | |
| CN111583716A (en) | Vehicle obstacle avoidance method and device, electronic equipment and storage medium | |
| EP3751465A1 (en) | Methods, apparatuses and computer programs for generating a reinforcement learning-based machine-learning model and for generating a control signal for operating a vehicle | |
| CN114056425B (en) | Vehicle steering control method and device, vehicle and storage medium | |
| CN116653957A (en) | Speed changing and lane changing method, device, equipment and storage medium | |
| CN115092181A (en) | Vehicle control method and device, storage medium and processor | |
| CN110932932A (en) | Reliability determination method for automatic driving cloud service platform and related device | |
| WO2023230740A1 (en) | Abnormal driving behavior identification method and device and vehicle | |
| EP3989031B1 (en) | Systems and methods for fusing road friction data to enhance vehicle maneuvering | |
| CN110378460B (en) | Decision making method | |
| CN114026574A (en) | Running a trainable module while monitoring for departure from a trained application scope | |
| Zhao et al. | Evaluation of a semi-autonomous lane departure correction system using naturalistic driving data |
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 |