Motor vehicle tracking method based on road interaction

Abstract

The invention discloses a motor vehicle tracking method based on road interaction, and the method comprises the steps: building a vehicle motion model frame through the interaction logic of a vehicle with other vehicles and lane lines in the transverse and lane directions; establishing a lane direction acceleration model according to the distance and speed relationship between the vehicle and other vehicles in the same lane; establishing a target transverse lane change resistance model according to the relation of distance, speed and the like between adjacent lanes and vehicles in front of and behind the lane; establishing a transverse lane resistance model according to the relative distance between the vehicle and the lane line; a measurement model is established through radar scattering point characteristics, and target tracking is realized through a filtering algorithm. According to the scheme, the technical problem that an existing motor vehicle target tracking algorithm is insufficient in robustness and accuracy is solved.

Description

Technical Field

[0001] This invention relates to the field of mobile target tracking technology, and in particular to a method for tracking motor vehicles based on road interaction. Background Technology

[0002] With the shift of modern warfare towards informatization and intelligence, higher demands are being placed on the real-time performance, accuracy, and robustness of vehicle target tracking.

[0003] Since Rekalman proposed the Kalman filter algorithm in 1960, it has been widely used in target tracking as an efficient and reliable estimator. However, traditional motion models face many challenges when dealing with motor vehicle targets in complex road environments. This is because, from the perspective of motion models, vehicles do not follow a simple, singular motion pattern when driving on the road. Vehicle movement is influenced by road conditions (such as curves, slopes, intersections, etc.) and traffic rules (such as speed limits, traffic lights, and yield rules); in addition, there are complex interactions between vehicles, such as following behavior between vehicles traveling in the same direction and lane-changing interactions between vehicles in different lanes.

[0004] Therefore, Chandler, Song D, Kesting A, and others have studied vehicle following and lane-changing models, proposing a series of vehicle motion models such as GHR. However, existing models have significant shortcomings. First, existing lane-changing models do not comprehensively consider the impact of vehicles in front and behind in the target lane, failing to fully quantify the lane-changing resistance brought by vehicles in front and behind in the target lane, making it difficult to accurately predict vehicle lane-changing decisions and behaviors. Second, in terms of modeling vehicle driving characteristics, traditional models ignore the behavioral patterns of drivers under the visual guidance of lane lines—drivers usually tend to drive in the middle of the lane and will not drive across the line for a long time. This key characteristic directly affects the lateral movement trajectory and lane-changing process of the vehicle, but existing research has not incorporated it into the lane-changing model construction system, causing the model to deviate from the actual vehicle motion characteristics, which greatly weakens the adaptability and prediction accuracy of target tracking algorithms to complex vehicle behaviors.

[0005] In summary, existing target tracking technologies are insufficient to meet the demands for efficient and accurate tracking of motor vehicle targets in complex road environments. This invention provides a powerful tool for solving the problem of motor vehicle tracking in complex scenarios. By simultaneously considering both car-following and lane-changing models, this invention innovatively improves upon the interaction resistance between vehicles in the target lane and the driver's lane driving characteristics in the lane-changing model, constructing a novel motor vehicle target tracking algorithm based on road interaction. Summary of the Invention

[0006] This invention proposes a road-interaction-based vehicle target tracking algorithm, which is an innovative improvement on traditional car-following and lane-changing models. Building upon the synergistic effect of the car-following and lane-changing models, the algorithm improves the lane-changing model based on the MOBIL model, fully considering the lane-changing resistance from vehicles in front and behind in the target lane, constructing a quantitative model to accurately characterize the vehicle's lane-changing decision-making process. Simultaneously, it innovatively incorporates the driver's tendency to drive in the middle of the lane under the influence of lane lines, optimizing the prediction of the vehicle's lateral movement trajectory. Through multi-dimensional modeling and algorithm fusion, efficient and accurate tracking of motor vehicle targets in complex road environments is achieved. To achieve the above objectives, the technical solution of this invention is as follows:

[0007] A method for tracking motor vehicles based on road interaction includes the following steps:

[0008] Step 1: Establish the motion model of the vehicle target: Assume there are N targets at time k. k A flight path is defined, with the x-direction representing the lane direction and the y-direction representing the lateral direction perpendicular to the lane. The motion state of the i-th vehicle at time k is defined as follows: satisfy in, Let x and y be the coordinates of the i-th vehicle at time k, respectively. Let v be the magnitude of the velocity components of the i-th vehicle in the x and y directions at time k. k It is Gaussian noise, F k Let Γ be the state transition matrix. k This is the process noise distribution matrix. The additional acceleration term is caused by the coupling of motion states between vehicles and lane lines, and includes the influence of all road interactions on motion. Determined by surrounding vehicles, to meet in, Additional acceleration in the direction of the lane is determined by the vehicle in front in the same lane; This is the additional acceleration in the lateral direction, determined by vehicles in adjacent lanes and those in the current lane, and consists of two parts: in, The lane-changing resistance generated by a vehicle attempting to change lanes when the lane-changing conditions are not met is determined jointly by vehicles in adjacent lanes and those adjacent to the current lane. Furthermore, drivers typically tend to drive in the middle of their lanes and avoid prolonged driving along the lane lines; this key characteristic directly affects the vehicle's lateral movement trajectory and the lane-changing process. This refers to the additional lateral acceleration that propels the vehicle toward the center equilibrium point of each lane.

[0009] Step 2: Establish additional acceleration in the lane direction Model: Based on the distance l in front e Whether there is a car inside depends on two situations, which need to be determined. Values ​​are taken when there are no cars. When there is a vehicle, a car-following model (CFM) can be used for modeling, such as the Gazis-Herman-Rothery (GHR) model.

[0010] Step 3: Establish a lateral lane change resistance model: To address the lane change resistance generated by vehicles attempting to change lanes under conditions where lane change conditions are not met, an improved MOBIL model is adopted. This model comprehensively assesses the distance and speed of vehicles in the target lane and the lane in front and behind, thus establishing a comprehensive model for lateral lane change resistance.

[0011] Step 4: Establish the lateral lane resistance model. (Introduction) Describes the additional lateral acceleration that propels the vehicle toward the center equilibrium point of each lane. In the area near the center of the lane The direction should be towards the center. The closer the vehicle is to the lane line, the more... The larger it is; the closer it is to the center line, The smaller. Established by the lateral position of the vehicle. Model.

[0012] Step 5: Tracking based on filtering algorithms. Assume the vehicle is a point target with a certain spatial volume but occupying only a single sensor resolution unit. At each moment, the radar receives measurements generated from the vehicle's equivalent scattering center; that is, at each moment, the same vehicle can only generate one measurement. Then, the i-th vehicle has the following measurement model:

[0013] in, w is the j-th measurement acquired by the radar at time k. k With zero measurement noise and h(g) being a nonlinear observation function, the motion model established in steps one through four is used for filtering via the EKF algorithm. The beneficial effects of this invention are as follows:

[0014] (1) In view of the shortcomings of the traditional MOBIL model in not quantifying the impact of the distance between the front and rear vehicles in the target lane on the lane change resistance, this invention introduces a quantitative evaluation mechanism of the distance between the front and rear vehicles in the target lane on the basis of the MOBIL model. By constructing a mathematical model of lane change resistance, the interaction between the vehicle and the vehicle in the target lane during lane change is accurately characterized, which significantly improves the accuracy of lane change decision prediction.

[0015] (2) This invention innovatively incorporates the driver’s behavior characteristics of being guided by lane lines and tending to drive in the middle of the lane into the lane change model. By establishing a lateral motion trajectory model constrained by the lane center line, it effectively corrects the misjudgment of the vehicle’s driving over the lane line by the traditional model and improves the accuracy of the vehicle’s lateral position prediction.

[0016] (3) By integrating the car-following model and the improved lane-changing model, this invention constructs a motor vehicle target tracking algorithm system that is closer to the actual road interaction behavior, enhances the adaptability of the algorithm to dynamic behaviors such as lane changing and car-following in complex military scenarios, and provides more reliable technical support for accurate tracking of battlefield targets. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of vehicle relationships in the method of the present invention;

[0018] Figure 2 This is a schematic diagram of the lateral lane resistance in the method of the present invention;

[0019] Figure 3 A schematic diagram illustrating the relationship between lateral lane resistance and lateral distance in the method of this invention;

[0020] Figure 4 This is a schematic diagram of the overall method flow of the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this specification clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments in this specification, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments in this application without creative effort are within the scope of protection of this document. Specific Implementation Example 1:

[0023] A method for tracking motor vehicles based on road interaction includes the following steps:

[0024] Step 1: Modeling the target motion framework. Assume there are N targets at time k. k A trajectory, defining the motion state of the i-th vehicle at time k as... in, Let x and y be the coordinates of the i-th vehicle at time k, respectively. Let x and y be the magnitudes of the velocity components of the i-th vehicle at time k along the x and y axes, respectively. Assume the x-direction represents the lane direction and the y-direction represents the lateral direction perpendicular to the lane. The motion model should then satisfy:

[0025]

[0026] Among them, v k It is zero-mean Gaussian noise with a covariance of Q. k ; This refers to the additional acceleration term caused by the coupling of motion states between vehicles, as well as lane markings, etc.; Fk Let Γ be the state transition matrix. k Let be the process noise distribution matrix, satisfying:

[0027]

[0028] Where T is the sampling interval.

[0029] It includes the impact of the interaction between vehicles on motion. Determined by surrounding vehicles, satisfying:

[0030]

[0031] in, The additional acceleration in the x-direction is determined by the vehicle in front in the same lane; The additional acceleration in the y-direction is determined by vehicles in adjacent lanes and vehicles adjacent to the current lane. This is the additional acceleration in the y-direction, determined by the vehicle's lane-changing behavior, lane markings, and vehicles in adjacent lanes. It consists of two parts:

[0032]

[0033] in, The lane-changing resistance generated by a vehicle attempting to change lanes when the lane-changing conditions are not met is determined jointly by vehicles in adjacent lanes and those adjacent to the current lane. Furthermore, drivers typically tend to drive in the middle of their lanes and avoid prolonged driving along the lane lines; this key characteristic directly affects the vehicle's lateral movement trajectory and the lane-changing process. This refers to the additional lateral acceleration that propels the vehicle toward the center equilibrium point of each lane.

[0034] Step 2: Additional acceleration in lane direction Modeling. Based on the distance ahead l e Whether there is a car inside depends on two situations:

[0035] 1. No vehicles are present. At this time, the vehicle is freely moving in the x-direction.

[0036] 2. There is a car. At this point, the vehicle's movement in the x-direction is mainly influenced by the car in front, and a car-following model (CFM) can be used for modeling. The Gazis-Herman-Rothery (GHR) model is a classic car-following model that describes motion dependence by modeling the acceleration of the following vehicle as a function of its motion relative to the car in front. Introducing variables... This indicates that the vehicle preceding the i-th vehicle at time k is the j-th vehicle. According to the GHR model, we can obtain:

[0037]

[0038] Where c, β, and γ are undetermined parameters, and τ is the reaction time.

[0039] Stop lines, obstacles, and other targets can affect the vehicle's motion. In this case, such targets can be considered equivalent to a virtual stationary vehicle in front, whose motion state is as follows: in, It is the target's location information.

[0040] Step 3: Lateral lane change resistance Modeling. The lane-changing resistance generated by a vehicle attempting to change lanes when the lane-changing conditions are not met is determined jointly by vehicles in adjacent lanes and those adjacent to the current lane. Vehicles typically exist in two states in the lateral direction: stable and lane-changing. When a vehicle is in a stable state, its lateral velocity... Smaller, always staying in the same lane; when the vehicle is changing lanes, the lateral speed... Larger vehicles can cross one or more lanes in a short period of time. Lane changes require certain conditions to be met; defining these conditions is crucial for constructing a lane-changing model. The MOBIL model is a classic lane-changing model that divides lane-changing conditions into two parts: safety criteria and incentive criteria. The safety criteria require that after the lane change, the new following vehicle (… Figure 1 The deceleration of the car in lane o does not exceed a certain safety threshold; the incentive criterion is based on the following vehicles before and after lane change ( Figure 1 The degree of change in the acceleration of the vehicle (o, r) is used to measure the feasibility of lane changing.

[0041] Introducing variables Let represent the vehicles following the i-th vehicle in its current lane at time k (a lane change is only considered complete when the vehicle has fully entered another lane). Figure 1 The middle vehicle (r) and the following vehicle in the target lane after changing lanes ( Figure 1 (Vehicle o), where g = 1 indicates changing lanes to the left, and g = -1 indicates changing lanes to the right. When the vehicle satisfies:

[0042]

[0043] At that time, a lane change will be made. Among them, Let b be the acceleration of the i-th vehicle in the x-axis direction before and after the lane change. safe For safe acceleration, Let be the excitation acceleration of the i-th vehicle, p be the courtesy coefficient that determines the degree of influence of other vehicles on the lane change, and Δa be the acceleration of the i-th vehicle. thThis is the incentive threshold. However, in reality, when p is not too small, equation (2) already includes equation (1), so only the influence of equation (2) on lane changing is considered subsequently. The MOBIL model does not consider the influence of the distance between the vehicle in front and the vehicle behind in the target lane on lane changing, so the MOBIL model is improved. Based on equation (2), the vehicle also needs to satisfy:

[0044] l1≥l th1 ,l2≥l th2 (3)

[0045] Only then can you change lanes. Here, l1 and l2 represent the distances to the vehicle behind and in front, respectively, when changing lanes. th1 ,l th2 This is a safety distance threshold. When surrounding vehicles do not meet the lane-changing conditions (2) and (3), they will create resistance to the vehicle attempting to change lanes, preventing the lane-changing behavior. This resistance is defined as... When g satisfies (Only considering) Resistance of the direction lane, such as when At that time, only the resistance generated by the left lane to lane changing is considered. )hour, satisfy:

[0046]

[0047] Where f(·) is the resistance function, it should satisfy the condition related to the vehicle's lateral velocity. The directions are opposite, and the numerical values ​​are the same as... l th1 -l1 and l th2 -l2 is positively correlated. Let x be the excitation acceleration of the i-th vehicle at time k in the x direction.

[0048] Step 4: Lateral Lane Resistance Modeling. Vehicles must adhere to lane markings and traffic rules, and to avoid collisions with vehicles in adjacent lanes, drivers tend to maintain a stable position in the center of the lane. When a lane change is needed, the vehicle quickly crosses the lane markings without lingering on them. Therefore, in the lateral direction, a vehicle is only stable at the center line of each lane; at other locations, the vehicle tends to move to the position where it is easiest to achieve equilibrium. In the longitudinal direction, each position is equivalent.

[0049] Introduction The additional lateral acceleration that propels the vehicle toward the center equilibrium point of each lane is described above. As can be seen from the above analysis, at the center of each lane... In the area near the center of the lane The direction should point towards the center. For example... Figure 2 As shown in the figure, Δ represents half the lane width.

[0050] The closer a vehicle gets to the lane line, the more immediately it should move away. The larger it is, the closer it is to the center line, and the closer it is to a state of equilibrium. The smaller. Assuming Lateral position The relationship is linear. Not only with lateral position It should also be related to lateral velocity. Related. To prevent vehicles from being hit by vehicles traveling in the opposite direction in the same lane when changing lanes. Constraints cause slow steering, therefore in Downward direction The region in the opposite direction should become smaller, while... The size also varies with Increase and decrease to reduce the impact on lane changes. Lateral position The relationship is roughly as follows Figure 3 As shown in the figure This indicates the vehicle's offset from the center of the lane.

[0051] Based on the above analysis, we can summarize as follows:

[0052]

[0053] Where j = -∞,L,-1,1,L,∞, Let be the slope, satisfying The lane center offset satisfies

[0054] Step 5: Track using a filtering algorithm. Assume the vehicle is a point target with a certain spatial volume but occupying only a single sensor resolution unit. At each moment, the radar receives measurements generated from the vehicle's equivalent scattering center; that is, at each moment, the same vehicle can only generate one measurement. Then, the i-th vehicle has the following measurement model:

[0055]

[0056] in, Let j be the j-th measurement acquired by the radar at time k. These represent the radial distance, azimuth angle, and radial velocity measured at time k, respectively; w k The measurement noise is zero-mean Gaussian, and its covariance is R. k h(g) is a nonlinear observation function, defined as follows:

[0057]

[0058] Based on the aforementioned motion and measurement models, filtering is performed using the EKF algorithm. Since EKF is not the focus of this invention and has no particularity, this part is omitted. Specific Implementation Example 2:

[0060] In a first aspect, the present invention provides a method for tracking motor vehicles based on road interaction, the method comprising:

[0061] By establishing the interaction logic between the vehicle under test and other vehicles, and the interaction logic between the vehicle under test and the lane lines, a vehicle motion model framework is established.

[0062] The acceleration model in lane direction is determined by the distance and speed relationship between the vehicle under test and other vehicles in the same lane; the target lateral lane change resistance model is determined by the distance and speed relationship between other vehicles in adjacent lanes; and the lateral lane resistance model is determined by the relative distance between the vehicle under test and the lane lines.

[0063] A measurement model is established based on the characteristics of radar scattering points, and the vehicle under test is tracked using a filtering algorithm.

[0064] Furthermore, the establishment of a vehicle motion model framework through the interaction logic between the vehicle under test and other vehicles, and the interaction logic between the vehicle under test and lane lines, includes:

[0065] The vehicle motion model framework satisfies:

[0066]

[0067] Among them, v k It is zero-mean Gaussian noise. For the additional acceleration term caused by the coupling of motion states between vehicles and lane lines, Γ k The process noise distribution matrix is... Let F be the motion state of the i-th vehicle at time k. k Here is the state transition matrix. Let represent the motion state of the i-th vehicle at time k+1.

[0068] Furthermore, the process noise distribution matrix satisfies:

[0069]

[0070] Where T is the sampling interval.

[0071] The additional acceleration term satisfies:

[0072]

[0073] in, The additional acceleration in the x-direction is determined by the vehicle in front in the same lane; The additional acceleration in the y-direction is determined by vehicles in adjacent lanes and vehicles adjacent to the current lane. It is divided into two parts:

[0074]

[0075] in, The lane-changing resistance generated by a vehicle attempting to change lanes when the lane-changing conditions are not met is determined jointly by vehicles in adjacent lanes and vehicles adjacent to the current lane. This refers to the additional lateral acceleration that propels the vehicle toward the center equilibrium point of each lane.

[0076] Furthermore, the acceleration model for determining the lane direction by analyzing the distance and speed relationship between the vehicle under test and other vehicles in the same lane includes:

[0077] Additional acceleration in the x direction Based on the distance ahead l e Whether there is a car inside depends on two situations:

[0078] When there are no vehicles ahead, the vehicle is in a free-moving state in the x-direction.

[0079] When there is a vehicle ahead, the vehicle's movement in the x-direction is mainly influenced by the vehicle in front; a following model is used for modeling. Variables are introduced. This indicates that at time k, the vehicle in front of the i-th vehicle is the j-th vehicle. According to the following vehicle model, we can obtain:

[0080]

[0081] Where c, β, and γ are undetermined parameters, and τ is the reaction time.

[0082] Furthermore, determining the target lateral lane change resistance model based on the distance and speed relationships of other vehicles in adjacent lanes includes:

[0083] When g = 1, it indicates that the vehicle is changing lanes to the left; when g = -1, it indicates that the vehicle is changing lanes to the right. Let represent the lateral velocity of the i-th vehicle at time k, when At that time, the lane-changing resistance generated by a vehicle attempting to change lanes when the lane-changing conditions are not met. satisfy:

[0084]

[0085] in, Let Δa be the excitation acceleration of the i-th vehicle at time k. thTo set the incentive threshold, l1 and l2 are the distances to the new rear and front vehicles, respectively, during a lane change. th1 ,l th2 Let f(·) be the drag function, representing the safe distance threshold, and satisfy the condition related to the vehicle's lateral velocity. The directions are opposite, and the numerical values ​​are the same as... l th1 -l1 and l th2 -l2 is positively correlated.

[0086] Furthermore, the lane change of the vehicle under test must meet both safety and incentive criteria:

[0087] Let i be the vehicle following the i-th vehicle in its current lane at time k. Let g represent the vehicle following the i-th vehicle in the target lane after the i-th vehicle changes lanes at time k; where g = 1 indicates a lane change to the left, and g = -1 indicates a lane change to the right. When the vehicles satisfy:

[0088]

[0089] l1≥l th1 ,l2≥l th2

[0090] At that time, a lane change will be made; among them, Let b be the acceleration of the i-th vehicle in the x-axis direction before and after the lane change. safe For safety acceleration, p is the courtesy coefficient, which determines the degree of influence of other vehicles on lane changing.

[0091] Furthermore, the lateral lane resistance model is determined by the relative distance between the vehicle under test and the lane lines;

[0092] Additional lateral acceleration that propels the vehicle toward the center equilibrium point of each lane. for:

[0093]

[0094] Where j = -∞,L,-1,1,L,∞, Let be the slope, satisfying The lane center offset satisfies Δ indicates half the lane width.

[0095] Furthermore, the i-th vehicle has the following measurement model:

[0096]

[0097] in, Let j be the j-th measurement acquired by the radar at time k. These represent the radial distance, azimuth angle, and radial velocity measured at time k, respectively; w k The measurement noise is zero-mean Gaussian, and its covariance is R. k h(g) is a nonlinear observation function, defined as follows:

[0098]

[0099] Based on the above motion model and measurement model, the EKF algorithm is used for filtering.

[0100] In a second aspect, a road-interactive motor vehicle tracking system includes a control device and a radar detection device, wherein the control device controls the radar detection device to perform the method described in the first aspect.

[0101] Thirdly, an electronic device mounted on a radar detection device, the device including a processor and a memory electrically connected to the processor, the memory for storing a computer program, and the processor for calling the computer program to perform the method described in the first aspect.

[0102] Regarding the explanation of formulas, the formula notation in this invention adopts the industry-standard formula notation method. Unmarked formula symbols have the same meaning as the corresponding symbols in textbooks in this field. Any symbol in this invention can replace other symbols of the same nature; the meaning of the original formula symbol remains unchanged after the replacement.

[0103] Example embodiments have been disclosed herein, and while specific terminology has been used, it is for general illustrative purposes only and should not be construed as limiting. In some embodiments, it will be apparent to those skilled in the art that features, characteristics, and / or elements described in connection with particular embodiments may be used alone, or in combination with features, characteristics, and / or elements described in connection with other embodiments, unless otherwise expressly indicated. Therefore, those skilled in the art will understand that various changes in form and detail may be made without departing from the scope of the invention as set forth by the appended claims.

Claims

1. A method for tracking motor vehicles based on road interaction, characterized in that, The method includes: By establishing the interaction logic between the vehicle under test and other vehicles, and the interaction logic between the vehicle under test and the lane lines, a vehicle motion model framework is established. The vehicle motion model framework is filled with the lane direction acceleration model, the target heading lane change resistance model, and the lateral lane resistance model to obtain the vehicle motion model of the vehicle under test. The lane direction acceleration model is determined by the distance and speed relationship between the vehicle under test and other vehicles in the same lane. The target heading lane change resistance model is determined by the distance and speed relationship between other vehicles in adjacent lanes. The lateral lane resistance model is determined by the relative distance between the vehicle under test and the lane line. A measurement model is established based on the characteristics of radar scattering points. The measurement model is used to reflect the real-time radial distance, azimuth angle and radial velocity of the vehicle under test. The data from the measurement model is substituted into the vehicle motion model, and the vehicle under test is tracked through a filtering algorithm.

2. The method for tracking motor vehicles based on road interaction as described in claim 1, characterized in that, The process of establishing a vehicle motion model framework through the interaction logic between the vehicle under test and other vehicles, and the interaction logic between the vehicle under test and lane lines, includes: The vehicle motion model framework satisfies: Among them, v k It is zero-mean Gaussian noise. For the additional acceleration term caused by the coupling of motion states between vehicles and lane lines, Γ k The process noise distribution matrix is... Let F be the motion state of the i-th vehicle at time k. k Here is the state transition matrix. Let represent the motion state of the i-th vehicle at time k+1.

3. The method for tracking motor vehicles based on road interaction as described in claim 2, characterized in that, The process noise distribution matrix satisfies: Where T is the sampling interval; the additional acceleration term satisfies: in, The additional acceleration in the x-direction is determined by the vehicle in front in the same lane; The additional acceleration in the y-direction is determined by vehicles in adjacent lanes and vehicles adjacent to the current lane. It is divided into two parts: in, The lane-changing resistance generated by a vehicle attempting to change lanes when the lane-changing conditions are not met is determined jointly by vehicles in adjacent lanes and vehicles adjacent to the current lane. This refers to the additional lateral acceleration that propels the vehicle toward the center equilibrium point of each lane.

4. The method for tracking motor vehicles based on road interaction as described in claim 3, characterized in that, The lane-direction acceleration model is determined by the distance and speed relationship between the vehicle under test and other vehicles in the same lane, including: Additional acceleration in the x direction Based on the distance ahead l e Whether there is a car inside depends on two situations: When there are no vehicles ahead, the vehicle is in a free-moving state in the x-direction. When there is a vehicle ahead, the vehicle's movement in the x-direction is mainly influenced by the vehicle in front; a following model is used for modeling. Variables are introduced. This indicates that at time k, the vehicle in front of the i-th vehicle is the j-th vehicle. According to the following vehicle model, we can obtain: Where c, β, and γ are undetermined parameters, and τ is the reaction time.

5. The method for tracking motor vehicles based on road interaction as described in claim 4, characterized in that, The target heading lane change resistance model is determined by the distance and speed relationship between other vehicles in adjacent lanes, including: When g = 1, it indicates that the vehicle is changing lanes to the left; when g = -1, it indicates that the vehicle is changing lanes to the right. Let represent the lateral velocity of the i-th vehicle at time k, when At that time, the lane-changing resistance generated by a vehicle attempting to change lanes when the lane-changing conditions are not met. satisfy: in, Let Δa be the excitation acceleration of the i-th vehicle at time k. th To set the incentive threshold, l1 and l2 are the distances to the new rear and front vehicles, respectively, during a lane change. th1 ,l th2 Let f(·) be the drag function, representing the safe distance threshold, and satisfy the condition related to the vehicle's lateral velocity. The directions are opposite, and the numerical values ​​are the same as... l th1 -l1 and l th2 -l2 is positively correlated.

6. The method for tracking motor vehicles based on road interaction as described in claim 5, characterized in that, The lane change of the vehicle under test must meet both safety and incentive criteria: Let i be the vehicle following the i-th vehicle in its current lane at time k. Let i be the vehicle following the i-th vehicle in the target lane after the i-th vehicle changes lanes at time k; when the vehicle satisfies: l1≥l th1 ,l2≥l th2 At that time, a lane change will be made; among them, Let b be the acceleration of the i-th vehicle in the x-axis direction before and after the lane change. safe For safety acceleration, p is the courtesy coefficient, which determines the degree of influence of other vehicles on lane changing.

7. The method for tracking motor vehicles based on road interaction as described in claim 6, characterized in that, The lateral lane resistance model is determined by the relative distance between the vehicle under test and the lane lines, including: Additional lateral acceleration that propels the vehicle toward the center equilibrium point of each lane. for: Where j = -∞,L,-1,1,L,∞, Let be the y-coordinates of the i-th vehicle at time k. Let be the slope, satisfying The lane center offset satisfies 8. The method for tracking motor vehicles based on road interaction as described in claim 7, characterized in that, The i-th vehicle has the following measurement model: in, Let j be the j-th measurement acquired by the radar at time k. These represent the radial distance, azimuth angle, and radial velocity measured at time k, respectively; w k The measurement noise is zero-mean Gaussian, and its covariance is R. k h(g) is a nonlinear observation function, defined as follows: Based on the above motion model and measurement model, the EKF algorithm is used for filtering.

9. A motor vehicle tracking system based on road interaction, characterized in that, The system includes a control device and a radar detection device, wherein the control device controls the radar detection device to perform the method as described in claim 1.

10. An electronic device, characterized in that, The device is mounted on a radar detection system, the system including a processor and a memory electrically connected to the processor, the memory for storing a computer program, and the processor for calling the computer program to execute the method as described in claim 1.

Images