Self-learning lane-changing decision control system based on driver habits

Abstract

The present invention discloses a self-learning lane change decision control system based on driver habits, the system includes: a perception module, which is used to obtain the road information of the current vehicle and the surrounding vehicle information; a human-vehicle system cognition module, which is used to collect the driving data of the driver of the vehicle, and compare it with the driving data of different preset excellent drivers to extract driving data that meets the preferences of the driver of the vehicle and is close to the driving habits of excellent drivers; a signal processing module, which is used to make a lane change instruction according to the extracted driving data, road information and surrounding vehicle information; an automatic lane change execution module, which is used to control the relevant actuator to complete the lane change according to the lane change instruction, taking into account the driver's driving habits and lane change requirements. The present invention can realize a lane change operation that is more in line with the driving habits of the driver of the vehicle according to the different driving states of the vehicles around the vehicle and the driving habits of the driver of the vehicle, and can improve the intelligence and comfort of the automatic lane change system.

Description

Technical Field

[0001] The present invention belongs to the field of vehicle lane changing, and in particular relates to a self-learning lane changing decision control system based on driver habits. Background Art

[0002] Intelligence is one of the important trends in the development of the automotive industry today. It is expected that in the next decade, intelligent driving technology and systems will develop rapidly around the world. Intelligent driving can not only greatly reduce the fatigue of manual driving, but also effectively reduce the occurrence of traffic accidents, and has been widely favored.

[0003] The intelligent driving system mainly consists of three parts: perception, decision-making and control. Among them, the environmental perception system is the basis for realizing intelligent driving, and multi-sensor fusion is the inevitable trend for realizing intelligent driving environmental perception. At present, some vehicles are equipped with automatic lane change systems. The vehicle monitors the status of surrounding vehicles through the perception system, and the decision control system determines whether to change lanes. This is to a certain extent limited by the perfection of the decision control system, which affects the timeliness, stability and accuracy of the vehicle's lane change decision.

[0004] At the same time, different drivers have different driving habits, and the automatic lane changing system is difficult to adapt to the driving habits of different drivers. Summary of the invention

[0005] The purpose of the present invention is to provide a self-learning lane change decision control system based on driver habits, so that the lane change operation of the lane change decision control system conforms to the driving habits of the driver of the vehicle.

[0006] The technical solution adopted by the present invention is as follows:

[0007] Obtain lane change information by comparing surrounding vehicle information with system pre-set parameters;

[0008] Execute corresponding lane changing operation according to lane changing information;

[0009] The perception system collects the relative position, longitudinal speed, lateral speed, yaw rate and change rate of the vehicles in front, on the right and behind the vehicle, and transmits the information to the decision-making system;

[0010] The decision-making system makes lane-changing decisions based on the relative position between the vehicle and the ego vehicle, longitudinal velocity, lateral velocity, yaw rate and rate of change, the driving status of the ego vehicle, and the driving data of excellent drivers in the database;

[0011] The decision-making system will constantly compare the driving habits of the driver with the operation data of excellent drivers preset in the system, continuously update the algorithm of the decision-making system, obtain the decision algorithm that is closest to the driving habits of the driver, and apply it to the lane change decision system of the vehicle.

[0012] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0013] The present invention can realize a lane changing operation that is more in line with the driving habits of the driver of the own vehicle according to different driving conditions of vehicles around the own vehicle and the driving habits of the driver of the own vehicle, and can improve the intelligence and comfort of the automatic lane changing system. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 This is a block diagram of a self-learning lane change decision control system based on driver habits;

[0015] Figure 2 It is the data collection flow chart of the human-vehicle system cognitive module;

[0016] Figure 3 It is a schematic diagram of the vehicle lane changing application scenario. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0018] The present invention proposes a self-learning lane change decision and control method and system based on driver habits, so that the vehicle control system can learn the driving habits of excellent drivers in advance and add relevant parameters to the lane change decision and control system model. At the same time, in subsequent driving, the system can continuously perform self-learning in the background to optimize and improve the algorithm.

[0019] The present invention obtains surrounding vehicle information through an on-board sensing system (including but not limited to a camera, a laser radar, and a millimeter-wave radar);

[0020] Obtain lane change information by comparing surrounding vehicle information with system pre-set parameters;

[0021] Execute corresponding lane changing operation according to lane changing information;

[0022] The perception system collects the relative position, longitudinal speed, lateral speed, yaw rate and change rate of the vehicles in front, on the right and behind the vehicle, and transmits the information to the decision-making system;

[0023] The decision-making system makes lane-changing decisions based on the relative position between the vehicle and the ego vehicle, longitudinal velocity, lateral velocity, yaw rate and rate of change, the driving status of the ego vehicle, and the driving data of excellent drivers in the database;

[0024] The decision-making system will constantly compare the driving habits of the driver with the operation data of excellent drivers preset in the system, continuously update the algorithm of the decision-making system, obtain the decision algorithm that is closest to the driving habits of the driver, and apply it to the lane change decision system of the vehicle.

[0025] Embodiment 1:

[0026] Figure 1 It is a block diagram of a vehicle automatic lane change decision and control system disclosed in an embodiment of the present invention. The automatic lane change system mainly includes a perception module, a human-vehicle system cognition module, a signal processing module, and an automatic lane change execution module.

[0027] The perception modules are at least one of the sensors such as the camera and lidar installed in the front of the vehicle, the corner lidar, millimeter-wave radar, ultrasonic radar, etc. installed at the four corners of the vehicle. The automatic lane change system can obtain the road information of the current vehicle and the vehicle information around the vehicle traveling on the current road section through the above sensors; the above information mainly includes but is not limited to: lane information, vehicle position, vehicle longitudinal speed, acceleration, number of target vehicles, yaw angular velocity, angular velocity change rate, longitudinal acceleration, etc.

[0028] The input end of the signal processing module is connected to the perception module and the human-vehicle system cognitive module, and the output end is connected to the input end of the automatic lane change execution module. The signal processing module receives information from the human-vehicle system cognitive module and the perception module, and makes a lane change strategy based on the lateral and longitudinal distances, longitudinal acceleration, lateral acceleration and change rate of the surrounding vehicles and the self-vehicle, combined with the driver's driving habits and the driving habits of excellent drivers preset by the system. The perception system transmits the kinematic information of the surrounding vehicles to the signal processing module; the built-in excellent driver data and the collected driving data of the self-vehicle driver are transmitted to the human-vehicle system cognitive module. Through data comparison, the data that meets the preferences of the self-vehicle driver and is close to the driving habits of excellent drivers is extracted to the signal processing module. Combined with the surrounding environment, the signal processing module transmits signals to the relevant actuators to execute lane changes, so that the driver's driving habits and lane change needs can be taken into account at the same time.

[0029] The automatic lane change execution module controls the relevant actuators including but not limited to EPS, ESC, EMS and other control units to complete the lane change of the vehicle according to the instructions of the signal processing module.

[0030] Embodiment 2:

[0031] Figure 2 This is a flow chart of data collection of a human-vehicle system cognitive module disclosed in an embodiment of the present invention. This module mainly includes:

[0032] A driving data collection unit, used to obtain the driver's driving information;

[0033] As an optional implementation, the driving data acquisition unit can obtain the current driver's driving data through at least the following control units and actuators: accelerator pedal opening, engine ECU, brake pedal opening, ESC, steering wheel torque, steering wheel angle, EPS, etc.;

[0034] The cognitive control unit obtains the current driver's driving data and the data of the perception module. The perception module can obtain the road information of the current vehicle and the driving information of vehicles around the vehicle on the current road section through the vehicle's sensors; the above information mainly includes but is not limited to: lane information, vehicle position, vehicle longitudinal speed, acceleration, number of target vehicles, yaw angular velocity, angular velocity change rate, longitudinal acceleration, etc.

[0035] The driving data of different excellent drivers are collected and preset in the system, as shown in Table 1, mainly including but not limited to the following scenarios:

[0036] Table 1 Scenario table

[0037] Serial number Scenario 1 Highway (day / night / rain / fog / snow) 2 City roads (day / night / rain / fog / snow) 3 Closed area (day / night / rain / fog / snow) …… ……

[0038] Currently, information scenes may include: scenes can be classified into single-target scenes, two-target scenes, three-target scenes and multi-target scenes according to target types; they can be divided into pedestrian scenes, vehicle scenes and animal scenes according to categories; according to target behaviors, they can be classified into target acceleration scenes, target deceleration scenes, target entry scenes, target exit scenes, target accelerated driving scenes, target parking scenes, etc. Other meteorological and lighting scenes are also within the protection scope of this embodiment.

[0039] The system collects the current driver's driving data based on the driver's driving record, including the timing of the driver's lane change, the priority of lane change, etc. in the above scenarios.

[0040] The cognitive control unit combines the current scene input from the perception module and compares the current driver with the system's excellent driver data.

[0041] The driving scene data is classified based on the classification conditions of the preset driving scenes and the driver model, wherein the driver model is generated based on the driving behavior data under different preset driving scenes. In this embodiment, there may be multiple preset driving scenes, and each preset driving scene has corresponding classification conditions. Pre-classification is performed, and the driver model establishes a driving model corresponding to the preset driving scene through the driver's behavior data, and the model reflects which driving scene the current driving scene data belongs to through the driver's behavior data. When it is impossible to confirm which driving scene the currently acquired driving scene belongs to based on the classification conditions of the preset driving scenes, the driving scene can be classified based on the driver model in another dimension. The classification of the preset driving scenes has been explained in the above table.

[0042] The preset driving scenes are divided into several first main scenes and several sub-scenes. Among them, the first main scene is based on road conditions, weather, lighting and target information. The sub-scenes are divided according to the movement state of the vehicle and the relationship between other vehicles and the vehicle. Including but not limited to pedestrian scenes, vehicle scenes, animal scenes, acceleration scenes, cut-in and cut-out scenes, etc.

[0043] The human-vehicle system cognitive module compares the driving data of the driver of the vehicle and the preset data of excellent drivers. It can first compare based on the conditions of several first scenes, and then compare several sub-scenes under the divided first main scene. When the degree of fit is greater than the preset value, it is considered that the driving scene is relatively similar to the preset scene. When it exists, the first main scene and sub-scene with the highest degree of fit are selected for matching and directly output to the signal processing module. And the current matching scene model is continuously iterated to improve the driver model. The cognitive control unit compares the current driver with the preset excellent driving data of the system, combines the data of the current perception module, completes the online identification of the driver's habits, completes the data output of the driver's personalized habits, and saves it in the data storage unit.

[0044] When it does not exist, the perception module data is directly output to the signal processing module, and the system directly determines the current scene and outputs the execution information.

[0045] Embodiment three:

[0046] Figure 3 Schematic diagram of a vehicle lane change application scenario disclosed in an embodiment of the present invention. Figure 3 As shown, the automatic lane changing system mainly includes the following steps:

[0047] The perception system obtains driving information of surrounding vehicles;

[0048] 1) When the obstacle vehicle is directly in front of the vehicle:

[0049] According to the longitudinal speed V1, acceleration a1, longitudinal distance D1 of the obstacle vehicle; the longitudinal speed V0, acceleration a0 of the vehicle, the time T1 required for the current vehicle to hit the obstacle vehicle in front can be obtained

[0050] The signal processing module compares the data of the cognitive control unit to obtain a safety threshold. When T1 is greater than the safety threshold, the automatic lane change execution module can execute the lane change operation;

[0051] 2) When the obstacle vehicle is on the right side of the vehicle:

[0052] According to the obstacle vehicle's longitudinal velocity V2, longitudinal acceleration a2, yaw angular velocity w2, yaw angular velocity change rate ζ2, longitudinal distance D2, lateral distance H2; the vehicle's longitudinal velocity V0, acceleration a0 and lane line information;

[0053] The automatic lane change system determines whether △V≥V* is satisfied;

[0054] △V=V2-V0, V2 is the speed of the obstacle vehicle on the right, V0 is the current speed of the vehicle; V* is the speed threshold;

[0055] If △V≥V*, we can further determine whether △a≥a* is satisfied;

[0056] △a=a2-a0, a2 is the acceleration of the obstacle vehicle on the right, a0 is the current acceleration of the vehicle; a* is the acceleration threshold;

[0057] If △a≥a*, we can further determine whether △w≥w* is satisfied;

[0058] △w=w2-w0, w2 is the yaw rate of the obstacle vehicle on the right, a0 is the current yaw rate of the vehicle; a* is the yaw rate threshold;

[0059] If △w≥w*, we can further determine whether △ζ≥ζ* is satisfied;

[0060] △ζ=ζ2-ζ0, ζ2 is the yaw rate change rate of the obstacle vehicle on the right, ζ0 is the current yaw rate change rate of the vehicle; ζ* is the yaw rate change rate threshold;

[0061] If △ζ≥ζ*, the time T2 required for the ego vehicle to collide with the obstacle vehicle can be further determined;

[0062] The signal processing module compares the driver's driving data collected by the cognitive control unit to obtain the safety threshold T* that is most suitable for the current driver's operation. When T2≥T*, the automatic lane change execution module can execute the lane change operation.

[0063] In summary, the present invention completes the comparison between excellent driver data and vehicle driver behavior data, and continuously iterates to improve the driver model library; combined with the algorithm, it can output the driving strategy that best suits the current driver's habits, thereby improving the intelligence and comfort of the system.

[0064] It will be easily understood by those skilled in the art that the above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A self-learning lane change decision control system based on driver habits, characterized in that: The system includes a signal processing module and a perception module connected to the signal processing module, a human-vehicle system recognition module and an automatic lane change execution module; The perception module is used to obtain the road information of the current vehicle and the surrounding vehicles; The human-vehicle system recognition module is used to collect the driving data of the driver of the vehicle and compare it with the driving data of different preset excellent drivers to extract driving data that meets the preferences of the driver of the vehicle and is close to the driving habits of excellent drivers; A signal processing module, used to make lane change instructions based on the extracted driving data, road information and surrounding vehicle information; The automatic lane change execution module is used to control the relevant actuators to complete the lane change according to the lane change command of the signal processing module, taking into account the driver's driving habits and lane change needs; Among them, making a lane change instruction includes: 1) When the obstacle vehicle is directly in front of the vehicle: According to the longitudinal speed V1 of the obstacle vehicle, the longitudinal distance D1 and the longitudinal speed V0 of the vehicle, the time T1 required for the current vehicle to collide with the obstacle vehicle in front is obtained: When T1 is greater than the safety threshold, the automatic lane change execution module executes the lane change operation; the safety threshold is obtained by comparing the driving data of the driver of the vehicle with the driving data of different preset excellent drivers; 2) When the obstacle vehicle is on the right side of the vehicle: According to the lateral velocity V2, acceleration a2, yaw angular velocity w2, yaw angular velocity change rate ζ2, lateral distance H2 and lane line information of the obstacle vehicle; The signal processing module determines whether △V≥V* is satisfied; △V=V2-V, V2 is the lateral speed of the obstacle vehicle on the right, V is the lateral speed of the vehicle, and V* is the speed threshold; If △V≥V*, further determine whether △a≥a* is satisfied; △a=a2-a0, a2 is the acceleration of the obstacle vehicle on the right, a0 is the current acceleration of the vehicle, and a* is the acceleration threshold; If △a≥a*, further determine whether △w≥w* is satisfied; △w=w2-w0, w2 is the yaw rate of the obstacle vehicle on the right, w0 is the current yaw rate of the vehicle, and a* is the yaw rate threshold; If △w≥w*, further determine whether △ζ≥ζ* is satisfied; △ζ=ζ2-ζ0, ζ2 is the yaw rate change rate of the obstacle vehicle on the right, ζ0 is the current yaw rate change rate of the vehicle, and ζ* is the yaw rate change rate threshold; If △ζ≥ζ*, further determine the time T2 required for the vehicle to collide with the obstacle vehicle: When T2 ≥ T*, the automatic lane change execution module executes the lane change operation; wherein T* represents a preset time threshold, which is obtained by comparing the driving data of the driver of the vehicle with the driving data of different preset excellent drivers.

2. The self-learning lane change decision control system based on driver habits according to claim 1, characterized in that: The human-vehicle system cognitive module includes a driving data acquisition unit, a cognitive control unit, and a data storage unit; A driving data collection unit, used to collect driving data of the driver of the vehicle; Cognitive control unit, used to preset driving data of different excellent drivers; A data storage unit, used to store driving data of the vehicle and different excellent drivers; The driving data of the vehicle and different excellent drivers include the position of the obstacle vehicle, the longitudinal speed V1, acceleration a1, longitudinal distance D1 of the obstacle vehicle, the longitudinal speed V0, acceleration a0 of the vehicle, and the time T0 when the driver takes the lane change action; When storing driving data, the above driving data is stored in an array; When the driving scenario conditions are met, the driving data of the driver of the vehicle is compared with the driving data of different preset excellent drivers to determine the timing of changing lanes.

3. The self-learning lane change decision control system based on driver habits according to claim 1, characterized in that: The perception module includes a camera installed at the front of the vehicle, a laser radar, and at least one of a corner laser radar, a millimeter-wave radar, and an ultrasonic radar installed at the four corners of the vehicle.

4. The self-learning lane change decision control system based on driver habits according to claim 1, characterized in that: Road information and surrounding vehicle information include lane information, vehicle position, vehicle longitudinal speed, acceleration, number of target vehicles, yaw rate, angular velocity change rate and longitudinal acceleration.

5. The self-learning lane change decision control system based on driver habits according to claim 1, characterized in that: Driving data includes accelerator pedal opening, engine ECU data, brake pedal opening, ESC data, steering wheel torque, steering wheel angle and EPS data.

Images