A lane change assistance method and system based on lane reference lines

Abstract

The application discloses a lane change assisting method and system for a vehicle based on a lane reference line, which comprises the following steps: combining a surrounding sensing result of the vehicle to calculate a current lane change intention, calculating a driving cost of a current lane and adjacent lanes, and selecting a lane with the minimum cost to drive according to a calculation result. The lane with the minimum cost is judged for lane change rationality, and lane change is performed if the judgment is reasonable. If the lane with the minimum cost is the current lane, straight driving is maintained. When judging the lane change intention, whether a distance between an autonomous vehicle and a front obstacle meets a minimum safety distance, a type of an obstacle in front of the current lane, and an operation state of the obstacle are considered on the basis of speed and acceleration, so that collision between the vehicle and the obstacle is avoided. Filtering processing is added, and the acceleration and the speed of the obstacle vehicle are considered to prevent the distance between the autonomous vehicle and the obstacle from oscillating near the safety distance, so that the lane change result is stable, and the system is robust.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of automatic driving, in particular to a vehicle lane changing assistance method and system based on lane reference line. BACKGROUND

[0002] In the automatic driving system, due to the changing vehicle conditions on the road, the automatic driving vehicle cannot always automatically drive along the lane. For example, the current three-lane will become two-lane; there are interference factors affecting the driving experience in the current lane; at this time, the driver wants to start the ALC function, etc. Therefore, the automatic driving vehicle needs to change lane. Based on these situations, a safe, comfortable and robust lane changing assistance system needs to be designed.

[0003] The existing automatic driving lane changing control method usually compares the driving speed and acceleration of the automatic driving vehicle with the expected value, and if the expected speed and acceleration are not met, it is determined that the vehicle has a lane changing intention. Calculate the distance between the automatic driving vehicle and the front and rear vehicles of the adjacent lane and the distance between the automatic driving vehicle and the front vehicle of the lane, and whether the distance is greater than the safety distance. If it is greater than the safety distance, finally determine whether the lane changing meets the maximum benefit of lane changing. If it meets, change lane. The existing technology has the following shortcomings: the lane changing intention only considers the vehicle speed and acceleration, the lane changing intention judgment condition is not clear, and the lane changing decision condition is not perfect, which leads to the problem that the system lane changing decision result is unstable and the system robustness is poor in the case of unstable lane changing condition.

[0004] For example, a vehicle lane changing system and control method based on MTTC are disclosed in Chinese patent literature, with publication number CN111746540A, which includes a vehicle distance monitoring unit, an acceleration monitoring unit, a vehicle speed monitoring unit, a data processing unit and a vehicle speed control unit. The vehicle distance monitoring unit includes a front radar range finder and a rear radar range finder; the acceleration monitoring unit includes a lateral acceleration sensor, a longitudinal acceleration sensor, a front acceleration measuring instrument and a rear acceleration measuring instrument; the vehicle speed monitoring unit includes a longitudinal vehicle speed sensor, a front speed measuring instrument and a rear speed measuring instrument; the data processing unit includes an MTTC calculation module, a lane changing time calculation module, a first judgment module, a second judgment module, a third judgment module and an information processing module; the main control unit includes a voice prompt device and a control execution device. The above scheme lacks a selection method for the target lane and cannot automatically determine the target lane; by calculating the collision time of the front and rear vehicles of the target lane, the steering wheel angle is controlled, and when the distance between the front and rear vehicles of the target lane is near the critical distance, the situation of back and forth oscillation exists. The problem of unstable lane changing result. SUMMARY

[0005] The present application is to overcome the problems of the prior art vehicle lane changing auxiliary method and system, such as the uncertainty of the lane changing intention judgment condition, the instability of the lane changing result when the distance between the front and rear vehicles in the target lane oscillates near the critical distance, and the poor robustness of the system, and to provide a vehicle lane changing auxiliary method and system based on lane reference lines, which can keep the output of the lane changing system stable and have high robustness.

[0006] To achieve the above-mentioned object, the present application adopts the following technical solutions:

[0007] A vehicle lane changing auxiliary method based on lane reference lines comprises the following steps:

[0008] Collecting the driving state of the vehicle and obtaining the surrounding perception information of the vehicle; comparing the position information of the obstacle with the regional division information near the vehicle according to the collected information to determine the region where the obstacle falls;

[0009] Obtaining that the vehicle is in an automatic driving state or a lane changing auxiliary state from the driving state of the vehicle and the surrounding perception information of the vehicle;

[0010] Judging whether the vehicle in the automatic driving state has a lane changing intention based on the relationship between the actual speed and the expected speed of the automatic driving vehicle, the relationship between the distance of the automatic driving vehicle and the obstacle vehicle and the following safety distance, the type of the obstacle in front of the automatic driving vehicle in the current lane, and the motion state;

[0011] Calculating the lane changing cost of the vehicle in the automatic driving state according to the surrounding perception information of the vehicle, and selecting the lane with the minimum lane changing cost as the target lane; for the vehicle in the lane changing auxiliary state, selecting the lane corresponding to the current turn signal as the target lane;

[0012] The lane line confidence of the target lane is determined whether to meet a threshold, whether the target lane line is a dashed line, whether the distance between the target lane obstacle and the vehicle is within a safe distance, and the target lane is determined whether to meet the reasonable degree condition of lane changing by combining the filtering processing; and the target lane which meets the reasonable degree condition is determined whether to exist intersection between the polyhedron of the target lane obstacle and the polyhedron of the vehicle by using the separation axis theorem, the obstacle collision detection is performed, and the lane changing decision information is output. The automatic driving state is the lane changing auxiliary system controlling the vehicle to change lane autonomously; the lane changing auxiliary state is ALC lever lane changing. When the automatic driving vehicle normally drives, the lane changing intention of the current lane is calculated by combining the surrounding perception results of the vehicle. When the automatic driving vehicle drives in the current lane and generates the lane changing intention, the driving cost of the current lane and the adjacent lane is calculated, and the lane with the minimum cost is selected to drive according to the calculation result. Finally, the lane changing rationality of the lane with the minimum cost is judged, and if the lane changing is reasonable, the lane is changed; if the lane with the minimum cost is the current lane, the vehicle keeps driving straight. If the lane with the minimum cost and the lane with the second minimum cost are both lane changing lanes and the lane changing is not feasible, the vehicle keeps driving straight. Then, the collision detection is performed according to the selected rationality target lane, combining the planned trajectory and the predicted trajectory of the obstacle, and finally the decision result is output.

[0013] Preferably, the position information of the obstacle is compared with the region division information near the vehicle according to the collected information, and the region where the obstacle falls is determined, including: obtaining the longitudinal coordinate value, longitudinal velocity, longitudinal acceleration, transverse coordinate value, transverse velocity and transverse acceleration of the obstacle in the Frenet coordinate system, taking the current lane center line of the vehicle as the longitudinal reference line, taking the line perpendicular to the longitudinal reference line X meters behind the vehicle as the transverse reference line, and taking the obstacle detection region behind the vehicle as the region within X meters behind the vehicle, and the region division information of the obstacle is as follows:

[0014] The obstacle is in front of the current lane automatic driving vehicle;

[0015] The obstacle is behind the current lane automatic driving vehicle;

[0016] The obstacle is in front of the automatic driving vehicle in the left lane;

[0017] The obstacle is behind the automatic driving vehicle in the left lane;

[0018] The obstacle is in front of the automatic driving vehicle in the right lane;

[0019] Obstacle behind the vehicle in the right lane of the autonomous vehicle;

[0020] Wherein: l is the lateral offset of the obstacle, S is the longitudinal offset of the obstacle; l0 is the lateral offset of the vehicle, s0 is the longitudinal offset of the vehicle; w is the lane width. && is the "and" logic symbol in the programming language.

[0021] As preferred, the vehicle in the autonomous driving state whether to have a lane-changing intention is determined based on the relationship between the actual speed of the autonomous vehicle and the expected speed, the relationship between the distance between the autonomous vehicle and the obstacle vehicle and the following safety distance, the type of the obstacle in front of the lane of the autonomous vehicle, and the motion state.

[0022] When the difference between the expected speed and the actual speed of the autonomous vehicle accounts for more than a preset threshold of the expected speed, the autonomous vehicle generates a lane-changing intention;

[0023] When the distance between the obstacle in front of the autonomous vehicle and the vehicle is less than the minimum safety distance, the autonomous vehicle generates a lane-changing intention;

[0024] When the obstacle in front of the autonomous vehicle is a large truck or a stationary obstacle, the autonomous vehicle generates a lane-changing intention.

[0025] As preferred, the lane-changing cost of the vehicle in the autonomous driving state is calculated based on the vehicle surrounding perception information, and the lane-changing cost COST satisfies:

[0026] COST = w1*Cost1 + w2*Cost2 + W3*Cost3 + w4*Cost4

[0027] Wherein, Cost1 is the congestion degree of the target lane, the farther the distance between the obstacle vehicle in front of the target lane and the autonomous vehicle, the more unobstructed the current lane, and the smaller the corresponding Cost1; Cost2 is the change trend of the obstacle in the current lane and the autonomous vehicle, which is calculated based on the speed and acceleration information of the two, such as finding that the two are in a rapid approaching stage, then Cost2 is high; Cost3 is the lane priority selection cost, the left lane is more suitable for lane changing than the right lane, so when the other costs of the left and right lanes are close, the left lane is preferentially selected for lane changing, and the left lane Cost3 is small; Cost4 is the type and motion state of the obstacle vehicle, such as the existence of a large vehicle or an obstacle in the target lane, the Cost4 cost is higher, and w1, w2, w3, and w4 are the weights of the corresponding parameters.

[0028] Wherein, when the distance between the obstacle in front of the autonomous vehicle and the vehicle is less than the minimum safety distance, the autonomous vehicle generates a lane-changing intention; specifically:

[0029] The actual speed of the autonomous vehicle is directly obtained from the chassis information, denoted as Vcur. The expected speed of the autonomous vehicle needs to consider the maximum speed set by the system, the maximum speed that can be achieved to maintain the set following distance, and the speed that can be achieved on the current road curvature, denoted as Vexp. The ratio is

[0030]

[0031] When Vratio is greater than a certain threshold and remains for multiple frames, the autonomous vehicle generates a lane change intention;

[0032] When the distance between the autonomous vehicle and the obstacle in front of the autonomous vehicle is less than the minimum safety distance, the autonomous vehicle generates a lane change intention; specifically:

[0033] The obstacle in front of the autonomous vehicle in the current lane filtered by the obstacle region division module is taken as the reference obstacle, denoted as obj. Therefore, the distance between the autonomous vehicle and the obstacle obj is:

[0034] S = Sobj - SegO - ROH - FOH

[0035] S obj and S ego are the S values of the obstacle obj and the autonomous vehicle ego in the Frenet coordinate system, respectively, ROH is the distance from the rear axle of the obstacle vehicle obj to the tail, and FOH is the distance from the rear axle of the autonomous vehicle ego to the front.

[0036] The minimum safety distance S min = max(S0, S1), where S0 is the distance calculated by the speed and acceleration of the autonomous vehicle and the obstacle vehicle according to the kinematic model, and S1 is the distance calculated by the set following distance. The larger of the two is taken as the minimum safety distance. When the distance S is less than the minimum safety distance S min , the autonomous vehicle generates a lane change intention.

[0037] As a preferred embodiment, the judgment of whether the distance between the target lane obstacle and the vehicle is within the safety distance includes: judging whether the distance S1 between the autonomous vehicle and the front vehicle in the current lane is greater than the lane change safety distance S1_safe, whether the distance S2 between the autonomous vehicle and the front vehicle in the target lane is greater than the lane change safety distance S2_safe, and whether the distance S3 between the autonomous vehicle and the rear vehicle in the target lane is greater than the lane change safety distance S3_safe; only when the following relationship is satisfied, the autonomous vehicle can safely change lanes: the specific relationship is as follows:

[0038] S1 > S1_safe && S2 > S2_safe && S3 > S3_safe;

[0039] S1 = Sobj - SegO - ROH - FOH S1_safe = (Vego - Vobj) * t

[0040] wherein S obj and S ego are the S values of the obstacle obj and ego autonomous vehicle in the Frenet coordinate system respectively, ROH is the distance from the rear axle to the tail of the obstacle vehicle obj, FOH is the distance from the rear axle to the head of the ego autonomous vehicle. Vego is the speed of the ego autonomous vehicle, Vobj is the speed of the obstacle vehicle, t is the lane changing time;

[0041] S2 = Sobj - SegO - ROH - FOH

[0042] S2_safe = (Vego - Vobj) * t

[0043] wherein S obj and S ego are the S values of the obstacle obj and ego autonomous vehicle in the Frenet coordinate system respectively, ROH is the distance from the rear axle to the tail of the obstacle vehicle obj, FOH is the distance from the rear axle to the head of the ego autonomous vehicle. Vego is the speed of the ego autonomous vehicle, Vobj is the speed of the obstacle vehicle, t is the lane changing time;

[0044] S3 = SegO - Sobj - ROH - FOH

[0045] S3_safe = (Vobj - Vego) * t

[0046] wherein S obj and S ego are the S values of the obstacle obj and ego autonomous vehicle in the Frenet coordinate system respectively, FOH is the distance from the rear axle to the head of the obstacle vehicle obj, ROH is the distance from the rear axle to the tail of the ego autonomous vehicle. Vego is the speed of the ego autonomous vehicle, Vobj is the speed of the obstacle vehicle, t is the lane changing time.

[0047] As a preferred, the combination with the filtering processing includes: when S1 > S1_safe + Sbuf, the ego autonomous vehicle and the vehicle in front of the current lane satisfy the minimum lane changing safety distance, and thereafter as long as S1 > S1_safe - Sbuf before the completion of this lane changing, it is considered that the ego autonomous vehicle and the vehicle in front of the current lane satisfy the minimum lane changing safety distance in the lane changing process;

[0048] When S2 > S2_safe + Sbuf, the autonomous vehicle meets the minimum lane-changing safety distance with the front vehicle in the target lane, and thereafter as long as S2 > S2_safe - Sbuf before the completion of this lane-changing, it is considered that the autonomous vehicle meets the minimum lane-changing safety distance with the front vehicle in the target lane during the lane-changing process;

[0049] When S3 > S3_safe + Sbuf, the autonomous vehicle meets the minimum lane-changing safety distance with the rear vehicle in the target lane, and thereafter as long as S3 > S3_safe - Sbuf before the completion of this lane-changing, it is considered that the autonomous vehicle meets the minimum lane-changing safety distance with the rear vehicle in the target lane during the lane-changing process, wherein: Sbuf is a conditional judgment threshold buffer. In the lane-changing process, when the actual distance oscillates around the safety distance, the stability of the lane-changing result is ensured, and the system robustness is improved.

[0050] As a preferred, the lane-changing decision information includes: left lane-changing, center keeping or right lane-changing.

[0051] A vehicle lane-changing assistance system based on lane reference lines, which adopts the vehicle lane-changing assistance method based on lane reference lines according to any one of the above, comprising:

[0052] An input module: collecting vehicle driving state information and obtaining vehicle surrounding perception information; wherein, the vehicle driving state includes: vehicle speed, acceleration, steering wheel angle information, vehicle function level information and vehicle pose information of the autonomous vehicle; the vehicle surrounding perception information includes: lane line information, obstacle information, traffic sign information and obstacle trajectory prediction information;

[0053] An obstacle region division module: calculating the lane attribute of the obstacle, removing the obstacle irrelevant to the autonomous vehicle, judging and outputting the positional relationship of the remaining obstacles relative to the autonomous vehicle, the positional relationship including being located in the front, rear of the autonomous vehicle, being located in the front, rear of the left lane of the autonomous vehicle, being located in the front, rear of the right lane of the autonomous vehicle;

[0054] A lane-changing intention module: judging and outputting the lane-changing intention of the autonomous vehicle based on the relationship between the actual speed and the expected speed of the autonomous vehicle, the relationship between the distance of the autonomous vehicle and the obstacle vehicle and the following safety distance, the type and motion state of the obstacle in front of the autonomous vehicle in the current lane;

[0055] For a vehicle in an autonomous driving state, the lane-changing cost is calculated according to the vehicle surrounding perception information, and the lane with the minimum lane-changing cost is selected as the target lane; for a vehicle in a lane-changing assistance state, the lane corresponding to the current turn signal is selected as the target lane, and the lane-changing intention is outputted;

[0056] The lane changing rationality judgment module judges whether the lane line confidence of the target lane meets a threshold, whether the target lane line is a dashed line, whether the distance between the target lane obstacle and the vehicle is within a safe distance, and determines and outputs whether the target lane meets the rationality condition of lane changing in combination with filtering processing.

[0057] The collision detection and output module judges whether there is an intersection between the polyhedron of the target lane obstacle and the polyhedron of the vehicle using the separation axis theorem for the target lane meeting the rationality condition, performs obstacle collision detection, and outputs lane changing decision information. The function level mainly judges whether the automatic driving vehicle is in an automatic lane changing function or an ALC function.

[0058] Therefore, the present application has the following beneficial effects: (1) When judging the lane changing intention of the automatic driving vehicle, the distance between the automatic driving vehicle and the front obstacle is considered to meet the minimum safe distance, the type of the current lane front obstacle and the obstacle running state, etc., in addition to the speed and acceleration, to avoid collision between the vehicle lane changing and the obstacle.

[0059] (2) The filtering processing is added to consider the possible fluctuation of the obstacle vehicle acceleration and speed, to prevent the distance between the automatic driving vehicle and the obstacle from oscillating near the safe distance, to ensure the stability of the lane changing result and the good system robustness.

[0060] (3) The automatic driving vehicle and the obstacle vehicle are projected to the lane reference line and converted to the Frenet coordinate system for a series of calculations, to simplify the lane changing control method and algorithm used in the actual lane changing process.

[0061] (4) The autonomous lane changing and the ALC function are integrated together to assist lane changing in different driving states of the vehicle, to have the effect of multifunctional lane changing assistance. BRIEF DESCRIPTION OF DRAWINGS

[0062] Figure 1 is a vehicle lane changing assistance system flowchart based on a lane reference line according to an embodiment of the present application.

[0063] Figure 2 is a conversion schematic diagram of a Cartesian coordinate system to a Frenet coordinate system of a vehicle lane changing assistance method based on a lane reference line according to an embodiment of the present application.

[0064] Figure 3 is an obstacle region division schematic diagram in the Frenet coordinate system of a vehicle lane changing assistance method based on a lane reference line according to an embodiment of the present application.

[0065] Figure 4 is a schematic diagram of no collision between two objects during collision detection of a vehicle lane changing assistance method based on a lane reference line according to an embodiment of the present application.

[0066] Figure 5 This is a schematic diagram of a collision between two objects during collision detection in a vehicle lane change assist method based on lane reference lines according to an embodiment of the present invention.

[0067] In the diagram: 1. Autonomous vehicle; 2. Obstacle; 3. Longitudinal reference line. Detailed Implementation

[0068] The present invention will now be further described with reference to the accompanying drawings and specific embodiments.

[0069] Example:

[0070] like Figure 1 The lane change assist system based on lane reference lines shown includes an input module that collects vehicle driving state information and obtains surrounding perception information. The vehicle driving state includes: vehicle speed, acceleration, steering wheel angle information, vehicle function level information, and vehicle position and pose information of the autonomous vehicle 1. The surrounding perception information includes: lane line information, obstacle 2 information, traffic sign information, and obstacle 2 trajectory prediction information. The function level judgment determines whether the autonomous vehicle 1 is in autonomous lane change function or ALC lever lane change function.

[0071] Obstacle 2 area division module: Calculates the lane attributes of obstacle 2, removes obstacles 2 that are unrelated to autonomous vehicle 1, and determines and outputs the positional relationship of the remaining obstacles 2 relative to autonomous vehicle 1. The positional relationship includes being in front of or behind autonomous vehicle 1, in front of or behind the left lane of autonomous vehicle 1, and in front of or behind the right lane of autonomous vehicle 1.

[0072] Lane change intention module: Based on the relationship between the actual speed and the expected speed of autonomous vehicle 1, the relationship between the distance between autonomous vehicle 1 and obstacle 2 and the following safety distance, the type and motion state of obstacle 2 in front of autonomous vehicle 1 in the current lane, the module determines and outputs the lane change intention of the autonomous vehicle.

[0073] For vehicles in autonomous driving mode, the lane change cost is calculated based on the perception information around the vehicle, and the lane with the lowest lane change cost is selected as the target lane; for vehicles in lane change assist mode, the lane corresponding to the current turn signal is selected as the target lane, and the lane change intention is output.

[0074] Lane change rationality judgment module: Determines whether the confidence level of the target lane line meets the threshold, whether the target lane line is a dashed line, and whether the distance between the target lane obstacle 2 and the vehicle is within the safe distance. Combined with filtering processing, it determines and outputs whether the target lane meets the rationality conditions for lane change.

[0075] Collision detection and output module: For target lanes that meet the reasonableness conditions, use the separation axis theorem to determine whether there is an intersection between the polyhedron of obstacle 2 in the target lane and the polyhedron of the vehicle, perform obstacle 2 collision detection, and output lane change decision information.

[0076] This embodiment also discloses a vehicle lane change assistance method based on lane reference lines, including:

[0077] Collect vehicle driving status and obtain surrounding perception information; compare the location information of obstacle 2 with the area division information near the vehicle based on the collected information to determine the area where obstacle 2 fell; specifically including:

[0078] Based on the reference line, such as Figure 2 The diagram shows the conversion of the positions of autonomous vehicle 1 and obstacle vehicle 2 into Frenet coordinates. This yields the corresponding longitudinal coordinate values ​​s, longitudinal velocity s', and longitudinal acceleration s', and the lateral coordinate values ​​l, lateral velocity l', and lateral acceleration l''. For example... Figure 2 As shown, point p in the Cartesian coordinate system is projected onto the reference line as point M. Obtain the longitudinal coordinates, longitudinal velocity, longitudinal acceleration, lateral coordinates, lateral velocity, and lateral acceleration of obstacle 2 in the Frenet coordinate system. Using the current lane centerline as the longitudinal reference line 3, and the line perpendicular to the longitudinal reference line 200 meters behind the vehicle as the lateral reference line, the detection area for obstacle 2 behind the vehicle is within 200 meters. The obstacle 2 area division information is as follows:

[0079] Obstacle 2 is in front of autonomous vehicle 1 in the current lane;

[0080] Obstacle 2 is behind autonomous vehicle 1 in the current lane;

[0081] Obstacle 2 is in front of autonomous vehicle 1 in the left lane;

[0082] Obstacle 2 is behind autonomous vehicle 1 in the left lane;

[0083] Obstacle 2 is in front of autonomous vehicle 1 in the right lane;

[0084] Obstacle 2 is behind autonomous vehicle 1 in the right lane;

[0085] Where: l is the lateral offset of obstacle 2, S is the longitudinal offset of obstacle 2; l0 is the lateral offset of the vehicle, s0 is the longitudinal offset of the vehicle; w is the lane width. && is the "AND" logical operator in programming languages. Figure 3The lateral offset of the vehicle fl of the obstacle 2 in the Frenet coordinate system is l = w, Therefore, the vehicle fl of the obstacle 2 belongs to the left lane of the autonomous vehicle 1 in front. Thus, the obstacle 2 information in 6 directions relative to the front and rear of the current autonomous vehicle 1 and the front and rear of the lane adjacent to the current autonomous vehicle 1 is screened out.

[0086] Obtain that the vehicle is in an autonomous driving state or a lane changing assistance state from the vehicle driving state and the vehicle surrounding perception information; wherein the autonomous driving state is that the lane changing assistance system controls the vehicle to change lanes autonomously, and the lane changing assistance state is that the ALC lever changes lanes.

[0087] Determine whether the vehicle in the autonomous driving state has a lane changing intention based on the relationship between the actual speed of the autonomous vehicle 1 and the expected speed, the relationship between the distance between the autonomous vehicle 1 and the vehicle of the obstacle 2 and the following safety distance, the type of the obstacle 2 in front of the current lane of the autonomous vehicle 1, and the motion state.

[0088] Wherein, the proportion of the actual speed of the autonomous vehicle 1 and the expected speed is calculated as follows:

[0089] The actual speed of the autonomous vehicle 1 is directly obtained from the chassis information, which is represented as Vcur. The expected speed of the autonomous vehicle 1 needs to consider the maximum speed set by the system, the maximum speed that can be reached by maintaining the set following distance, and the speed that can be reached by the current road curvature, which is represented as Vexp. Then the proportion is

[0090]

[0091] When Vratio is greater than a certain threshold and remains for multiple frames, the autonomous vehicle 1 generates a lane changing intention.

[0092] Determine the relationship between the distance between the autonomous vehicle 1 and the vehicle of the obstacle 2 and the minimum safety distance:

[0093] The obstacle 2 in front of the current lane of the autonomous vehicle 1 screened out by the obstacle 2 region division module is taken as the reference obstacle 2, denoted as obj. Therefore, the distance between the autonomous vehicle 1 and the obstacle 2 obj is:

[0094] S = Sobj-Sego-ROH-FOH

[0095] S obj and S ego are the S values of the obstacle 2 obj and the autonomous vehicle 1 ego in the Frenet coordinate system respectively, ROH is the distance from the rear axle of the vehicle obj to the tail, and FOH is the distance from the rear axle of the autonomous vehicle 1 ego to the front.

[0096] The minimum safety distance Smin = max(S0, S1), where S0 is the distance calculated according to the kinematic model of the speed and acceleration of the autonomous vehicle 1 and the obstacle 2 vehicle, and S1 is the distance calculated according to the set following distance. The larger of the two is taken as the minimum safety distance. When the distance S is less than the minimum safety distance S min , the autonomous vehicle 1 generates a lane change intention.

[0097] Determine the type of the obstacle 2 in front of the autonomous vehicle 1 in the current lane:

[0098] When the nearest obstacle 2 in front of the autonomous vehicle 1 in the current lane is a large vehicle or a stationary obstacle 2, the autonomous vehicle 1 generates a lane change intention.

[0099] For a vehicle in an autonomous driving state, calculate the lane change cost according to the vehicle's surrounding perception information, and select the lane with the minimum lane change cost as the target lane. For a vehicle in a lane change assistance state, select the lane corresponding to the current turn signal as the target lane. Assuming that there are three lanes, mainly left lane, current lane, and right lane. When the autonomous vehicle 1 generates a lane change intention, calculate the cost value of the autonomous vehicle 1 driving in the three lanes. The driving cost of the three lanes is mainly calculated according to the type, speed, and acceleration of the obstacle 2 vehicle in different lanes, as well as the speed and acceleration of the autonomous vehicle 1. It is represented as follows:

[0100] COST = w1*Cost1 + w2*Cost2 + W3*Cost3 + w4*Cost4

[0101] where Cost1 is the congestion degree of the target lane. The farther the distance between the obstacle 2 vehicle in front of the target lane and the autonomous vehicle 1, the more unobstructed the current lane, and the smaller the corresponding Cost1. Cost2 is the change trend of the obstacle 2 in the current lane and the autonomous vehicle 1, which is calculated from the speed and acceleration information of the two, such as finding that the two are in a rapid approach stage, then Cost2 is high. Cost3 is the lane priority selection cost. According to the normal lane change logic, the left lane is more suitable for lane change than the right lane, so when the other costs of the left and right lanes are close, the left lane is preferred for lane change, and the Cost3 of the left lane is set to be smaller than the Cost3 of the right lane under the same conditions. Cost4 is the type and motion state of the obstacle 2 vehicle, such as the presence of a large vehicle or a roadblock in the target lane, which has a higher Cost4. w1, w2, w3, and w4 are the weights of the corresponding parameters. The lane with the minimum cost is selected for lane change by considering various situations.

[0102] whether the target lane line confidence of the target lane meets a threshold, whether the target lane line is a dashed line, whether the distance between the target lane obstacle 2 and the ego vehicle is within a safe distance, and in combination with filtering processing, whether the target lane meets the reasonable degree condition of lane changing;

[0103] wherein the target lane line confidence of the target lane is determined:

[0104] directly obtaining a perception result, determining whether it is greater than a certain threshold, and if it is met, it means that the automatic driving vehicle 1 meets the lane changing requirement.

[0105] whether the target lane line of the target lane is a dashed line or a dashed and solid line:

[0106] According to the traffic rules, only when the lane line is a dashed line, can the lane be changed. Therefore, according to the type of lane line obtained by perception, it is determined whether it is a dashed line or a dashed and solid line, and the side close to the automatic driving vehicle 1 is a dashed line, and if it meets the requirement, it means that the automatic driving vehicle 1 meets the lane changing requirement.

[0107] whether the distance between the automatic driving vehicle 1 and the obstacle 2 during the lane changing process is greater than the lane changing safety distance:

[0108] whether the distance S1 between the automatic driving vehicle 1 and the front vehicle of the current lane is greater than the lane changing safety distance S1_safe, whether the distance S2 between the automatic driving vehicle 1 and the front vehicle of the target lane is greater than the lane changing safety distance S2_safe, and whether the distance S3 between the automatic driving vehicle 1 and the rear vehicle of the target lane is greater than the lane changing safety distance S3_safe;

[0109] S1=Sobj-Sego-ROH-FOH

[0110] S1_safe=(Vego-Vobj)*t

[0111] wherein S obj and S ego are the S values of the obstacle 2 obj and the automatic driving vehicle lego in the Frenet coordinate system respectively, ROH is the distance from the rear axle of the obstacle 2 vehicle obj to the tail, FOH is the distance from the rear axle of the automatic driving vehicle lego to the head, Vego is the speed of the automatic driving vehicle 1, Vobj is the speed of the obstacle 2 vehicle, and t is the lane changing time;

[0112] S2=Sobj-Sego-ROH-FOH

[0113] S2_safe=(Vego-Vobj)*t

[0114] wherein S obj and S egoS1 and S2 are the S values of the obstacle 2 obj and the autonomous vehicle 1 ego in the Frenet coordinate system respectively, and FOH is the distance from the rear axle to the front of the obstacle 2 obj, and ROH is the distance from the rear axle to the tail of the autonomous vehicle 1 ego. Vego is the speed of the autonomous vehicle 1, and Vobj is the speed of the obstacle 2 vehicle, and t is the lane changing time.

[0115] S3 = Sego - Sobj - ROH - FOH

[0116] S3_safe = (Vobj - Vego) * t

[0117] wherein S obj and S ego are the S values of the obstacle 2 obj and the autonomous vehicle 1 ego in the Frenet coordinate system respectively, and FOH is the distance from the rear axle to the front of the obstacle 2 obj, and ROH is the distance from the rear axle to the tail of the autonomous vehicle 1 ego. Vego is the speed of the autonomous vehicle 1, and Vobj is the speed of the obstacle 2 vehicle, and t is the lane changing time.

[0118] Only when the following relationship is met, the autonomous vehicle 1 can safely change lanes. The specific relationship is as follows:

[0119] S1 > S1_safe && S2 > S2_safe && S3 > S3_safe.

[0120] Filtering processing is performed:

[0121] including: when S1 > S1_safe + Sbuf, the autonomous vehicle 1 and the vehicle in front of the current lane meet the minimum lane changing safety distance, and thereafter as long as S1 > S1_safe - Sbuf before the completion of this lane changing, it is considered that the autonomous vehicle 1 meets the minimum lane changing safety distance with the vehicle in front of the current lane during the lane changing process;

[0122] when S2 > S2_safe + Sbuf, the autonomous vehicle 1 and the vehicle in front of the target lane meet the minimum lane changing safety distance, and thereafter as long as S2 > S2_safe - Sbuf before the completion of this lane changing, it is considered that the autonomous vehicle 1 meets the minimum lane changing safety distance with the vehicle in front of the target lane during the lane changing process;

[0123] when S3 > S3_safe + Sbuf, the autonomous vehicle 1 and the vehicle behind the target lane meet the minimum lane changing safety distance, and thereafter as long as S3 > S3_safe - Sbuf before the completion of this lane changing, it is considered that the autonomous vehicle 1 meets the minimum lane changing safety distance with the vehicle behind the target lane during the lane changing process, wherein: Sbuf is a conditional judgment threshold buffer amount. In the lane changing process, when the actual distance oscillates around the safety distance, the stability of the lane changing result is ensured, and the system robustness is improved.

[0124] For target lanes that meet the reasonableness criteria, the separation axis theorem is used to determine whether there is an intersection between the polyhedron of obstacle 2 in the target lane and the polyhedron of the vehicle, collision detection of obstacle 2 is performed, and lane change decision information is output:

[0125] Based on the selected target lane, a lane-changing trajectory is planned, and the trajectory of obstacle 2 predicted by the prediction module is used for collision detection. Since the vehicle is a regular rectangle, the separation axis theorem is used for collision detection. Projections are made onto four axes, with the projection lengths of the two rectangle centers denoted as da, the projection length of autonomous vehicle 1 (ego) denoted as db, and the projection length of obstacle 2 (vehicle) denoted as dc.

[0126] Two objects will not collide if their projected lengths on a certain axis satisfy the following relationship: da > db + dc. For example... Figure 4 As shown, among the four projections in this group, the upper right image satisfies this relationship, therefore the two objects do not collide. Figure 5 As shown, none of the four projections in this set satisfy this relationship, so the two objects collide. A, B, C, and D are projections on four different axes.

[0127] When determining the lane-changing intention of autonomous vehicle 1, this application considers not only speed and acceleration, but also factors such as whether the distance between autonomous vehicle 1 and obstacle 2 meets the minimum safe distance, the type of obstacle 2 in the current lane, and the operating state of obstacle 2, to avoid collisions between the vehicle and obstacle 2 during lane changes. A filtering process is added to consider the acceleration and potential speed fluctuations of obstacle 2, preventing the distance between autonomous vehicle 1 and obstacle 2 from oscillating around the safe distance, ensuring stable lane-changing results and good system robustness. The autonomous vehicle 1 and obstacle 2 are projected onto the lane reference line and converted to the Frenet coordinate system for a series of calculations, simplifying the lane-changing control methods and algorithms used in actual lane-changing processes. The autonomous lane-changing function is integrated with ALC, providing lane-changing assistance in different driving states, achieving a multi-functional lane-changing assistance effect.

[0128] The specific embodiments described herein are merely illustrative of the spirit of the invention. Those skilled in the art to which this invention pertains may make various modifications or additions to the described specific embodiments or use similar methods to substitute them, without departing from the spirit of the invention or exceeding the scope defined by the appended claims.

[0129] Although the terms ALC, filtering, obstacles, lane change cost, etc. are used more frequently herein, the possibility of using other terms is not excluded. The use of these terms is merely for the sake of convenience in describing and explaining the nature of the application; they are not to be construed as limiting in any way.

Claims

1. A lane change assist method for a vehicle based on a lane reference line, characterized by, The method comprises the following steps: Collecting vehicle driving state and obtaining vehicle surrounding perception information; Comparing the position information of the obstacle with the region division information near the vehicle according to the collected information to determine the region where the obstacle falls; Obtaining that the vehicle is in an automatic driving state or a lane changing assistance state from the vehicle driving state and the vehicle surrounding perception information; Judging whether the vehicle in the automatic driving state has a lane changing intention based on the relationship between the actual speed and the expected speed of the automatic driving vehicle, the relationship between the distance of the automatic driving vehicle and the obstacle vehicle and the following safety distance, the type and motion state of the obstacle in front of the automatic driving vehicle in the current lane; Calculating the lane changing cost of the vehicle in the automatic driving state according to the vehicle surrounding perception information and selecting the lane with the minimum lane changing cost as the target lane; For the vehicle in the lane changing assistance state, the lane corresponding to the current turn signal is selected as the target lane; Judging whether the lane line confidence of the target lane meets the threshold, whether the target lane line is a dashed line, whether the distance between the obstacle in the target lane and the vehicle is within the safety distance, and determining whether the target lane meets the rationality condition of lane changing by combining the filtering processing; And for the target lane meeting the rationality condition, whether there is an intersection between the polyhedron of the obstacle in the target lane and the polyhedron of the vehicle is judged by using the separation axis theorem to perform obstacle collision detection and output the lane changing decision information.

2. The lane reference line based vehicle lane change assist method according to claim 1, wherein The step of comparing the position information of the obstacle with the region division information near the vehicle according to the collected information comprises the following steps: Obstacle in front of current lane autonomous vehicle; Obstacle behind the current lane autonomous vehicle; Obstacle in front of AV in left lane Obstacle behind AV in left lane; Obstacle in front of right lane autonomous vehicle; Obstacle behind AV in right lane Obtaining the longitudinal coordinate value, longitudinal velocity, longitudinal acceleration, transverse coordinate value, transverse velocity and transverse acceleration of the obstacle in the Frenet coordinate system, taking the current lane center line of the vehicle as the longitudinal reference line, and taking the line perpendicular to the longitudinal reference line at X meters behind the vehicle as the transverse reference line, and the obstacle region division information is as follows:

3. The lane reference line based vehicle lane change assist method according to claim 2, wherein, Wherein: l is the transverse offset of the obstacle, S is the longitudinal offset of the obstacle; l0 is the transverse offset of the vehicle, s0 is the longitudinal offset of the vehicle; w is the lane width. The step of judging whether the vehicle in the automatic driving state has a lane changing intention based on the relationship between the actual speed and the expected speed of the automatic driving vehicle, the relationship between the distance of the automatic driving vehicle and the obstacle vehicle and the following safety distance, the type and motion state of the obstacle in front of the automatic driving vehicle in the current lane comprises the following steps: When the difference between the expected speed and the actual speed of the automatic driving vehicle accounts for more than a preset threshold of the expected speed, the automatic driving vehicle generates a lane changing intention; When the distance between the obstacle in front of the automatic driving vehicle and the vehicle is less than the minimum safety distance, the automatic driving vehicle generates a lane changing intention; 4. The lane reference line based vehicle lane change assist method according to claim 3, wherein, When the obstacle in front of the automatic driving vehicle is a large truck or a stationary obstacle, the automatic driving vehicle generates a lane changing intention. The step of calculating the lane changing cost of the vehicle in the automatic driving state according to the vehicle surrounding perception information comprises the following steps: The lane changing cost COST satisfies: COST = w1*Cost1 + w2*Cost2 + W3*Cost3 + w4*Cost4 Wherein, Cost1 is the congestion degree of the target lane, the farther the distance between the obstacle vehicle in front of the target lane and the autonomous vehicle, the more smooth the current lane, and the smaller the corresponding Cost1; Cost2 is the change trend of the obstacle and the autonomous vehicle, which is calculated by the speed, acceleration and other information of the two, such as finding that the two are in a rapid approach stage, then the Cost2 is high; Cost3 is the lane priority selection cost, the left lane is more suitable for lane changing than the right lane, so when the other costs of the left and right lanes are close, the left lane is preferentially selected for lane changing, and the Cost3 of the left lane is small; Cost4 is the type and motion state of the obstacle vehicle, such as the target lane exists a large vehicle or a roadblock, the Cost4 cost is higher, and w1, w2, w3, w4 are the weights of the corresponding parameters.

5. The lane reference line based vehicle lane change assist method according to claim 4, wherein, The judgment whether the distance between the target lane obstacle and the vehicle is within a safe distance includes: judging whether the distance S1 between the autonomous vehicle and the front vehicle of the current lane is greater than the lane changing safe distance S1_safe, whether the distance S2 between the autonomous vehicle and the front vehicle of the target lane is greater than the lane changing safe distance S2_safe, and whether the distance S3 between the autonomous vehicle and the rear vehicle of the target lane is greater than the lane changing safe distance S3_safe; only when the following relationship is satisfied, the autonomous vehicle can safely change lanes: the specific relationship is as follows: S1>S1_safe&&S2>S2_safe&&S3>S3_safe.

6. The lane reference line based vehicle lane change assist method according to claim 5, wherein The filtering processing includes: when S1>S1_safe+Sbuf, the autonomous vehicle and the vehicle in front of the current lane satisfy the minimum lane changing safe distance, and thereafter as long as S1>S1_safe-Sbuf before the completion of this lane changing, it is considered that the autonomous vehicle and the vehicle in front of the current lane satisfy the minimum lane changing safe distance during the lane changing process; When S2>S2_safe+Sbuf, the autonomous vehicle and the vehicle in front of the target lane satisfy the minimum lane changing safe distance, and thereafter as long as S2>S2_safe-Sbuf before the completion of this lane changing, it is considered that the autonomous vehicle and the vehicle in front of the target lane satisfy the minimum lane changing safe distance during the lane changing process; When S3>S3_safe+Sbuf, the autonomous vehicle and the vehicle behind the target lane satisfy the minimum lane changing safe distance, and thereafter as long as S3>S3_safe-Sbuf before the completion of this lane changing, it is considered that the autonomous vehicle and the vehicle behind the target lane satisfy the minimum lane changing safe distance, wherein: Sbuf is the conditional judgment threshold buffer amount.

7. The lane reference line based vehicle lane change assist method according to claim 1 or 6, characterized in that, The lane changing decision information includes: left lane changing, center keeping or right lane changing.

8. A lane change assist system for a vehicle based on a lane reference line, characterized in that The lane changing auxiliary method based on the lane reference line of any one of claims 1-7, comprising: an input module: collecting vehicle driving state information and obtaining vehicle surrounding perception information; wherein, the vehicle driving state includes: vehicle speed, acceleration, steering wheel angle information, vehicle function level information and vehicle pose information of the autonomous vehicle; the vehicle surrounding perception information includes: lane line information, obstacle information, traffic sign information and obstacle trajectory prediction information; The obstacle region division module calculates the lane attribute of the obstacle, removes the obstacle irrelevant to the autonomous vehicle, judges and outputs the position relationship of the remaining obstacle relative to the autonomous vehicle, and the position relationship includes the front and rear of the autonomous vehicle, the front and rear of the left lane of the autonomous vehicle, and the front and rear of the right lane of the autonomous vehicle; The lane changing intention module judges and outputs the lane changing intention of the autonomous vehicle based on the relationship between the actual speed and the expected speed of the autonomous vehicle, the relationship between the distance of the autonomous vehicle and the obstacle vehicle and the following safety distance, the type and motion state of the obstacle in front of the current lane of the autonomous vehicle; The lane changing cost of the vehicle in the autonomous driving state is calculated according to the vehicle surrounding perception information, and the lane with the minimum lane changing cost is selected as the target lane; for the vehicle in the lane changing assistance state, the lane corresponding to the current turn signal is selected as the target lane, and the lane changing intention is output; The lane changing rationality judgment module judges whether the lane line confidence of the target lane meets the threshold, whether the target lane line is a dashed line, and whether the distance between the target lane obstacle and the vehicle is within the safe distance, and determines and outputs whether the target lane meets the lane changing rationality condition in combination with the filtering processing; The collision detection and output module judges whether there is an intersection between the polyhedron of the target lane obstacle and the polyhedron of the vehicle by using the separation axis theorem, performs obstacle collision detection, and outputs the lane changing decision information.

Images