Multi-compartment chain type vehicle anti-collision early warning system and method

Abstract

The invention discloses an anti-collision early warning system and method for a multi-compartment chained vehicle. The system is deployed on a vehicle head and a compartment, wherein the vehicle head part comprises a data acquisition module, a communication module, a map module, a pose estimation module, a situation estimation module and a collision early warning module; and the compartment part comprises a data acquisition and communication module. The data acquisition module is used for acquiring pose information and environment information of the chained vehicle; the communication module is used for the headstock to receive the pose information of the compartment; the map module is used for reading a known environment map; the pose estimation module is used for a vehicle kinematic model to calculate and predict the trajectory of a carriage; the situation estimation module is used for constructing a vehicle and environment target situation field and performing prediction; and the collision early warning module is used for discretizing and binarizing the chained vehicle and the environment target situation field, judging a potential collision risk area of the vehicle according to an intersection result of the chained vehicle and the environment target situation field in a global grid coordinate system, and outputting early warning information. According to the invention, the safety of the chained vehicle in a complex environment can be effectively improved, and the control difficulty is reduced.

Description

Technical Field

[0001] The invention belongs to the field of automation and computer technology, in particular to the field of anti-collision technology for multi-carriage chain-type vehicles, and specifically relates to an anti-collision warning system and method for multi-carriage chain-type vehicles. Background Art

[0002] In modern transportation systems, multi-carriage chained vehicles are widely used in public transportation and logistics transportation. Due to their structural characteristics, these vehicles are easily affected by various factors, leading to collision accidents. Therefore, designing an anti-collision warning system for multi-carriage chained vehicles has important application value.

[0003] Chinese patent application: A collision prediction method for autonomous driving of an articulated vehicle (application number: CN202211437902.2) discloses a collision avoidance prediction method for an articulated vehicle. A grid map is established to project the vehicle and obstacles onto the grid map. The projection of the vehicle posture information at future moments on the grid map is calculated through a kinematic model, and it is determined whether it will overlap with the projection of the obstacle. This scheme only considers the predicted vehicle posture information, and the vehicle has a low tolerance for sensor errors. Summary of the invention

[0004] The present invention aims to solve the above problems of the prior art. A method is proposed. The technical solution of the present invention is as follows:

[0005] A multi-carriage chain vehicle anti-collision warning system is deployed in the front and carriages. The front part includes a front data acquisition module, a front communication module, a map module, a posture estimation module and a situation estimation module, and a collision warning module. The carriage part includes a carriage data acquisition module and a carriage communication module, wherein:

[0006] The front data acquisition module is used to obtain the front position information of the chain vehicle and collect the surrounding environment target data; the carriage data acquisition module is used to obtain the data of the inertial measurement unit IMU of the carriage;

[0007] The front communication module is used for the front of the chained vehicle to receive the carriage IMU data; the carriage communication module is used for sending the carriage IMU data to the front of the chained vehicle;

[0008] The map module is used to read a known environment map and obtain known environment target information in the environment map;

[0009] The posture estimation module is used to calculate the current posture information of the vehicle head and each compartment through the vehicle kinematic model, and predict the vehicle posture data at several moments in the future;

[0010] The situation estimation module is used to construct corresponding situation fields for the environmental target and the chained vehicle respectively, and predict the situation fields of the chained vehicle situation field at several moments in the future according to the vehicle posture estimation information, and predict the target situation field according to the motion state estimation of the environmental target, and fuse the predicted situation fields respectively to obtain the corresponding total situation field;

[0011] The collision warning module is used to discretize and binarize the total situation field corresponding to the chained vehicle and the environmental target to obtain the corresponding local grid map, and finally convert the local grid map coordinates of the chained vehicle and the environmental target situation field to the global grid map coordinate system to calculate the intersection, determine whether the vehicle has a collision risk, and output collision warning information when there is a collision risk.

[0012] A multi-carriage chain vehicle anti-collision warning method based on the system comprises the following steps:

[0013] 2.1 Data Collection:

[0014] Vehicle head data collection: Vehicle head data collection module 0 obtains vehicle head position information And obtain the surrounding environment information in real time, obtain environmental target information The speed and position coordinates of the target information are represented by a five-tuple in represents the x, y coordinates of the target at time k, represents the linear velocity of the target x, y at time k, Represents the direction angle of the target at time k.

[0015] Carriage data collection: The data collection module of carriage i collects and obtains the IMU data of each carriage, where i = 1, 2, ..., n, representing the i-th carriage, and the front car number is 0;

[0016] 2.2 Communication: The communication module of car i sends the data information of the IMU of the car to the communication module of the front car, and the communication module of the front car receives it;

[0017] 2.3 Map import: The map module imports the known environment map and obtains the environment map target information;

[0018] 2.4 Vehicle pose estimation: Based on the kinematic model of the chained vehicle, and using the pose information obtained from the front sensor and the IMU data of the carriage, the pose information of the k front and each carriage of the chained vehicle at the current moment is estimated, and the pose of the chained vehicle at the future moment is predicted;

[0019] 2.5 Situation estimation: Based on the posture information of the chained vehicle, the situation fields of the front and carriages of the chained vehicle are constructed respectively, and the situation fields of each carriage of the vehicle at the future time τ are predicted based on the posture estimation of the chained vehicle, and then the situation fields of each carriage are integrated into the overall chained vehicle situation field. Based on the known environmental target information and target detection information, the situation fields of static, fixed and dynamic targets in the environment are constructed, and the motion estimation of the target predicts the environmental target situation field at the future time τ, and all target situation fields are integrated to obtain the overall target situation field;

[0020] 2.6 Collision warning: Set the chain vehicle collision safety threshold, discretize the vehicle situation field and the environmental target situation field, and binarize them through the safety threshold to obtain the corresponding local grid map. Finally, convert the local grid map coordinates of the chain vehicle and the environmental target situation field to the global grid map coordinate system to calculate the intersection, and use the collision warning function F alert To calculate whether the vehicle will collide, if a collision occurs, output collision warning information.

[0021] Furthermore, the posture estimation method of step 2.4 specifically includes the following steps:

[0022] 3.1 Vehicle posture calculation: According to the kinematic model of the chain vehicle, the posture relationship between carriage i+1 and carriage i is as follows;

[0023]

[0024] Where i = 0, 1, 2…; l0 is the wheelbase of the front of the vehicle; l i represents the wheelbase of carriage i; represents the turning angle between carriage i and carriage i+1 at time k, is the heading angle of carriage i at time k;

[0025] Get the front position information of the chain vehicle at time k according to the sensor Represents the position coordinates of the vehicle head at time k; represents the speed of the vehicle head at time k; represents the orientation angle of the train head at time k. The position and posture of each carriage are calculated in turn according to formula (1).

[0026] 3.2 Vehicle posture prediction:

[0027] The vehicle head position at time k+τ: The vehicle head position information obtained at time k Predict the future position at time τ

[0028]

[0029] Among them, δ is the front wheel turning angle of the vehicle;

[0030] The carriage position at time k+τ: according to the body parameters and hinge angle of the chain vehicle Given the k+τ head position information, the kinematic model is used to predict the position information of car i. The prediction of the carriage’s posture information is the same as formula (1).

[0031] Furthermore, the situation assessment method of step 2.5 specifically includes the following steps:

[0032] 4.1 Constructing the environment target situation field:

[0033] Construct static target situation field: construct the corresponding static situation field according to the target position and target characteristics respectively

[0034]

[0035] Among them, k obstp Represents the scale factor of the target feature; P obstp Indicates the target feature size; Indicates the location of the target.

[0036] And there are:

[0037]

[0038] Constructing the fixed target situation field: same as static target;

[0039] Construct dynamic target potential field: Construct dynamic target potential field according to target position and target kinematic characteristics Dynamic target situation field function:

[0040]

[0041] Among them, k obs represents the scaling factor of the target speed; m represents the compensation coefficient, which is used to compensate for the potential field changes caused by the target type; The target position is the vector pointing to the vehicle position; the target speed is remember and The angle is is a unit vector;

[0042] Estimated based on the current state of the target Use the Kalman filter to predict the position at the future k+τ time and construct the dynamic target potential field at the future time τ

[0043]

[0044] in, represents the speed of the vehicle at time τ in the future.

[0045] 4.2 Constructing the chain vehicle situation field:

[0046] Construction of the front situation field: According to the position and kinematic characteristics of the chain vehicle, the dynamic situation field of the front of the chain vehicle can be expressed as:

[0047]

[0048] Among them, k car A scaling factor representing the vehicle speed;

[0049] According to the chain vehicle posture prediction, the situation field of the vehicle head at time k+τ is calculated:

[0050]

[0051] Construction of carriage situation field: The movement of the carriage relative to the front carriage is constrained by the hinge. The carriage of the chain vehicle is basically stationary relative to the front of the vehicle. The static situation field of carriage i is constructed:

[0052]

[0053] Where i = 1, 2, ...; k cartp The proportional factor representing the characteristics of the carriage; P cartp Indicates the characteristic size of the carriage.

[0054] And there are:

[0055]

[0056] 4.3 Situation Field Fusion:

[0057] Chain vehicle situation field: Chain vehicles construct situation fields for the locomotive and each carriage separately, and then merge the situation fields of each carriage into the situation field of the entire vehicle

[0058]

[0059] Where n is the total number of chain vehicle carriages;

[0060] Environmental situation field: The situation fields of static, fixed and moving targets in the environment are fused to obtain the overall target situation field.

[0061]

[0062] Furthermore, the steps of the collision warning method in step 2.6 are as follows:

[0063] 5.1 Situation Field Discretization: The chain vehicle situation field and the environmental target situation field are projected onto the multi-valued local grid map respectively; the local grid map generated by the chain vehicle situation field projection is Grid veh,local ={Grid(i,j)|-M≤i≤M,-N≤j≤N} represents; the local grid map generated by the projection of the environmental target situation field is Grid obs,local ={Grid(i,j)∣-M≤i≤M,-N≤j≤N};

[0064] 5.2 Binarization of situation field: Set the safety threshold S0, and use the safety threshold to binarize the grid map of the chain vehicle situation field and the environmental target situation field:

[0065]

[0066] 5.3 Collision Warning: Design of Collision Warning Function F alert The calculation method is: convert the local grid map coordinates of the chained vehicle itself and the environmental target situation field to the global grid map coordinate system. veh,global With Grid obs,global Calculate the intersection of the two map grid coordinates and add them together;

[0067] Collision warning function F alert :

[0068]

[0069] If F alert ≥1, the intersection of the two maps is considered not to be an empty set, the vehicles may collide, and a collision warning message is output. Otherwise, F alert =0, the intersection of the two grid maps is an empty set and there is no risk of collision.

[0070] The advantages and beneficial effects of the present invention are as follows:

[0071] The present invention proposes a collision warning system and method for multi-carriage chained vehicles, which estimates and predicts the position and posture of each car, and combines the situation estimation and prediction of the chained vehicle and environmental targets, and then fuses the situation fields of the chained vehicle and environmental targets to obtain the corresponding total situation field, discretizes and binarizes the situation field of the chained vehicle and the situation field of the environmental target, calculates the intersection of each car of the vehicle and the environmental target in the global coordinate system of the grid map, and determines whether there is a collision risk based on the intersection result. If there is a collision risk, an alarm is issued, thereby reducing the risk of accidents. The system can significantly improve the driving safety of multi-carriage chained vehicles in complex environments. The system adds a forward prediction method, which can avoid risks in advance, reduce the difficulty of vehicle operation, and improve the transportation efficiency of chained vehicles. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 It is a structural diagram of a vehicle anti-collision warning system for a multi-carriage chain-type vehicle according to a preferred embodiment of the present invention.

[0073] Figure 2 Schematic diagram of the kinematic model of the chained vehicle in the embodiment of the present invention.

[0074] Figure 3 It is a flow chart of early warning judgment of the present invention.

[0075] Figure 4 It is a chain vehicle model diagram of the present invention.

[0076] Figure 5 It is an example of the situation field formed by the objectives of the present invention and the carriage. DETAILED DESCRIPTION

[0077] The following will describe the technical solutions in the embodiments of the present invention in detail in conjunction with the accompanying drawings in the embodiments of the present invention. The described embodiments are only part of the embodiments of the present invention.

[0078] The technical solution of the present invention to solve the above technical problems is:

[0079] Figure 1 The figure shows a structure diagram of a vehicle anti-collision warning system for a multi-carriage chain-type vehicle of the present invention. The method comprises:

[0080] The data acquisition module is used to obtain the position information of the front of the chain vehicle and collect the target data of the surrounding environment. The data acquisition module of the carriage is used to obtain the data of the IMU of the carriage;

[0081] The communication module is used for the front of the chain vehicle to receive the IMU data information of the carriage;

[0082] The map module is used to read a known environment map and obtain known environment target information in the environment map;

[0083] The posture estimation module is used to calculate the current posture information of the vehicle head and each compartment through the vehicle kinematic model, and predict the vehicle posture data at several moments in the future;

[0084] The situation estimation module is used to construct corresponding situation fields for the environmental target and the chained vehicle respectively, and predict the situation field of the chained vehicle according to the vehicle posture estimation information, and predict the target situation field according to the motion state estimation of the environmental target, and fuse the predicted situation fields to obtain the corresponding total situation field;

[0085] The collision warning module is used to discretize and binarize the total situation field of the chained vehicle and the environmental target respectively to obtain the corresponding local grid map, and finally convert the local grid map coordinates of the chained vehicle and the environmental target situation field to the global grid map coordinate system to calculate the intersection, determine whether the vehicle has a collision risk, and output collision warning information when there is a collision risk.

[0086] Figure 2 This is a model diagram of a chain vehicle of the present invention, which is composed of a locomotive and a plurality of carriages, wherein the locomotive is a four-wheel Ackerman structure, and the carriages are four-wheel structure carriages, and the locomotive and the carriages are connected by hinges.

[0087] Figure 3 This is a flow chart of the vehicle early warning judgment process of the present invention. The chain vehicle posture estimate is calculated based on the kinematic model and the vehicle posture estimate at time k+τ is obtained by prediction. The vehicle situation field and target situation field at time k+τ are then calculated. The situation field is discretized and binarized to obtain the corresponding local grid map. Finally, the grid map coordinates of the chain vehicle itself and the local environment target situation field are converted to the global grid map coordinate system to calculate the intersection and determine whether the vehicle has a collision threat. The process includes the following steps:

[0088] (1) The vehicle's posture is estimated and predicted through the chain vehicle kinematic model to obtain the vehicle's posture information.

[0089] (2) Construction of situation field: Based on the obtained posture information of the chain vehicle, the situation field of the front and carriage of the chain vehicle is constructed separately, and then the situation field of each carriage is integrated into the situation field of the entire vehicle; the target situation field is constructed based on the known environmental target information and target detection information.

[0090] (2.1) Environmental target situation field

[0091] Construct static target situation field: construct the corresponding static situation field according to the target position and target characteristics respectively

[0092]

[0093] Static target situation field at the future k+τ time:

[0094]

[0095] Construct a fixed target situation field:

[0096]

[0097] Fixed target situation field at the future k+τ time:

[0098]

[0099] Construct dynamic target situation field: estimate the target state at time k Use the Kalman filter to predict the position at the next k+τ moment:

[0100]

[0101] Among them, F k is the state transfer matrix, specifically

[0102] According to the target location and target kinematic characteristics to construct dynamic target potential field Dynamic target situation field function:

[0103]

[0104] (2.2) Construct the vehicle's own situation field:

[0105] Construction of the front situation field: According to the position and kinematic characteristics of the chain vehicle, the dynamic situation field of the front of the chain vehicle can be expressed as:

[0106]

[0107] Construction of carriage situation field: The movement of the carriage relative to the front carriage is constrained by the hinge. The carriage of the chain vehicle is basically stationary relative to the front of the vehicle. The static situation field of carriage i is constructed:

[0108]

[0109] According to the chain vehicle posture prediction, the situation field of the carriage at time k+τ is calculated:

[0110]

[0111] (2.4) Situation Field Fusion:

[0112] Chain vehicle situation field: Chain vehicles construct situation fields for the locomotive and each carriage separately, and then merge the situation fields of each carriage into the situation field of the entire vehicle Same as formula (11).

[0113] Environmental situation field: The situation fields of static, fixed and moving targets in the environment are fused to obtain the overall target situation field. Same as formula (12).

[0114] (3) Collision Warning

[0115] (3.1) Situation field discretization: The chain vehicle situation field and the environmental target situation field are projected onto the multi-valued local grid map. The local grid map generated by the chain vehicle situation field projection is Grid veh,local ={Grid(i,j)|-M≤i≤M,-N≤j≤N} represents; the local grid map generated by the projection of the environmental target situation field is Grid obs,local ={Grid(i,j)∣-M≤i≤M,-N≤j≤N}.

[0116] (3.2) Situation field binarization: Set the safety threshold S0, and binarize the local grid maps of the chain vehicle situation field and the environmental target situation field respectively using formula (13):

[0117] (3.3) Collision warning: Design collision warning function F alert , transform the local grid map coordinates of the chained vehicle itself and the environmental target situation field into the global grid map coordinate system, and then calculate the two grid maps Grid in the global coordinate system veh,global With Grid obs,global The intersection of is calculated as shown in formula (14). If F alert ≥1, the intersection of the two grid maps is not an empty set, it is judged that the vehicle may collide, and the collision warning information is output, otherwise F alert =0, no risk of collision.

[0118] Figure 4 This is a schematic diagram of a kinematic model of a multi-carriage chain vehicle in an embodiment of the present invention. The multi-carriage chain vehicle abstracts each two car bodies into the L shown in the figure below. i and L i+1 In this way, the kinematic calculation model of multi-carriage vehicles is simplified, the position of the vehicle head at future moments is calculated by prediction, and the trajectory points of the chain vehicle carriages are calculated recursively according to the kinematic formula. The specific steps are as follows.

[0119] (1) Posture estimation: The vehicle head data acquisition module 0 obtains the vehicle head posture information through sensors And obtain the surrounding environment information in real time, obtain environmental target information The speed and position coordinates of the target information are represented by a five-tuple

[0120] (2) Calculate the turning angle between carriages based on the chain vehicle IMU information

[0121]

[0122] in, represents the turning angle between carriage i and carriage i+1 at time k, a i,yawrepresents the angular acceleration of carriage i, w i represents the angular velocity value of carriage i, j is the number of sampling points, and Δt is the sampling time interval.

[0123] (3) Obtain the front position information of the chain vehicle at time k based on the sensor information Then carriage i obtains the posture information according to the relative relationship The calculation is the same as formula (1).

[0124] (4) Vehicle posture prediction:

[0125] The vehicle head position at time k+τ: The vehicle head position information obtained at time k Predict the future position at time τ The calculation is the same as formula (2).

[0126] The carriage position at time k+τ: according to the body parameters and hinge angle of the chain vehicle Given the k+τ head position information, the kinematic model is used to predict the position information of car i. The calculation is the same as formula (1).

[0127] Figure 5 It is an example of the environmental target and chain vehicle situation field of the present invention.

[0128] Figure a shows the static situation field of the chain vehicle. The carriage is fixed by hinges, so the carriage is stationary relative to the vehicle during the movement, and it also shows the situation field of the carriage. The static situation field of the stationary target in Figure b is characterized by being evenly distributed around the target. The potential field size is the same at the same distance from the target, and the farther away from the target, the smaller the potential field.

[0129] Figure c is the dynamic situation field of the front of the chain vehicle, and Figure d is the dynamic situation field of the environmental target. The distribution of the situation field is no longer uniform, but is enhanced in the direction of speed. The greater the absolute value of the speed, the greater the degree of enhancement.

[0130] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions.

[0131] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0132] The above embodiments should be understood to be only used to illustrate the present invention and not to limit the protection scope of the present invention. After reading the contents of the present invention, technicians can make various changes or modifications to the present invention, and these equivalent changes and modifications also fall within the scope defined by the claims of the present invention.

Claims

1. A multi-carriage chain vehicle anti-collision warning system, deployed on the front and carriages, characterized in that: The front part includes the front data acquisition module, the front communication module, the map module, the posture estimation module and the situation estimation module, and the collision warning module. The carriage part includes the carriage data acquisition module and the carriage communication module, among which: The front data acquisition module is used to obtain the front position information of the chain vehicle and collect the surrounding environment target data; the carriage data acquisition module is used to obtain the data of the inertial measurement unit IMU of the carriage; The front communication module is used for the front of the chained vehicle to receive the carriage IMU data; the carriage communication module is used for sending the carriage IMU data to the front of the chained vehicle; The map module is used to read a known environment map and obtain known environment target information in the environment map; The posture estimation module is used to calculate the current posture information of the vehicle head and each compartment through the vehicle kinematic model, and predict the vehicle posture data at several moments in the future; The situation estimation module is used to construct corresponding situation fields for the environmental target and the chained vehicle respectively, and predict the situation fields of the chained vehicle situation field at several moments in the future according to the vehicle posture estimation information, and predict the target situation field according to the motion state estimation of the environmental target, and fuse the predicted situation fields respectively to obtain the corresponding total situation field; The collision warning module is used to discretize and binarize the total situation field corresponding to the chained vehicle and the environmental target to obtain the corresponding local grid map, and finally convert the local grid map coordinates of the chained vehicle and the environmental target situation field to the global grid map coordinate system to calculate the intersection, determine whether the vehicle has a collision risk, and output collision warning information when there is a collision risk.

2. A multi-carriage chain vehicle anti-collision warning method based on the system of claim 1, characterized in that: The following steps are involved: 2.1 Data Collection: Vehicle head data collection: Vehicle head data collection module 0 obtains vehicle head position information And obtain the surrounding environment information in real time, obtain environmental target information The speed and position coordinates of the target information are represented by a five-tuple in represents the x, y coordinates of the target at time k, represents the linear velocity of the target x, y at time k, Represents the direction angle of the target at time k. Carriage data collection: The data collection module of carriage i collects and obtains the IMU data of each carriage, where i = 1, 2, ..., n, representing the i-th carriage, and the front car number is 0; 2.2 Communication: The communication module of car i sends the data information of the IMU of the car to the communication module of the front car, and the communication module of the front car receives it; 2.3 Map import: The map module imports the known environment map and obtains the environment map target information; 2.4 Vehicle pose estimation: Based on the kinematic model of the chained vehicle, and using the pose information obtained from the front sensor and the IMU data of the carriage, the pose information of the k front and each carriage of the chained vehicle at the current moment is estimated, and the pose of the chained vehicle at the future moment is predicted; 2.5 Situation estimation: Based on the posture information of the chained vehicle, the situation fields of the front and carriages of the chained vehicle are constructed respectively, and the situation fields of each carriage of the vehicle at the future time τ are predicted based on the posture estimation of the chained vehicle, and then the situation fields of each carriage are integrated into the overall chained vehicle situation field. Based on the known environmental target information and target detection information, the situation fields of static, fixed and dynamic targets in the environment are constructed, and the motion estimation of the target predicts the environmental target situation field at the future time τ, and all target situation fields are integrated to obtain the overall target situation field; 2.6 Collision warning: Set the chain vehicle collision safety threshold, discretize the vehicle situation field and the environmental target situation field, and binarize them through the safety threshold to obtain the corresponding local grid map. Finally, convert the local grid map coordinates of the chain vehicle and the environmental target situation field to the global grid map coordinate system to calculate the intersection, and use the collision warning function F alert To calculate whether the vehicle will collide, if a collision occurs, output collision warning information.

3. The early warning method according to claim 2, characterized in that: The posture estimation method of step 2.4 specifically comprises the following steps: 3.1 Vehicle posture calculation: According to the kinematic model of the chain vehicle, the posture relationship between carriage i+1 and carriage i is as follows; Where i = 0, 1, 2…; l0 is the wheelbase of the front of the vehicle; l i represents the wheelbase of carriage i; represents the turning angle between carriage i and carriage i+1 at time k, represents the orientation angle of carriage i at time k; Get the front position information of the chain vehicle at time k according to the sensor Represents the position coordinates of the vehicle head at time k; represents the speed of the vehicle head at time k; represents the orientation angle of the train head at time k. According to formula (1), the position and posture of each carriage are calculated in turn; 3.2 Vehicle posture prediction: The vehicle head position at time k+τ: The vehicle head position information obtained at time k Predict the future position at time τ Among them, δ is the front wheel turning angle of the vehicle; The carriage position at time k+τ: according to the body parameters and hinge angle of the chain vehicle Given the k+τ head position information, the kinematic model is used to predict the position information of car i. The prediction of the carriage’s position information is the same as formula (1).

4. The early warning method according to claim 2 is characterized in that the situation estimation method in step 2.5 specifically comprises the following steps: 4.1 Constructing the environment target situation field: Construct static target situation field: construct the corresponding static situation field according to the target position and target characteristics respectively Among them, k obstp Represents the scale factor of the target feature; P obstp Indicates the target feature size; Indicates the location coordinates of the target. And there are: Constructing the fixed target situation field: same as static target; Construct dynamic target potential field: Construct dynamic target potential field according to target position and target kinematic characteristics Dynamic target situation field function: Among them, k obs represents the scaling factor of the target speed; m represents the compensation coefficient, which is used to compensate for the potential field changes caused by the target type; The vector representing the target position pointing to the vehicle position; represents the target speed at the current k moment; and The angle is is a unit vector; represents the speed of the vehicle at the current time k. Estimated based on the current state of the target The Kalman filter is used to predict the target position at the future k+τ time: Construct the dynamic target potential field at the future time τ in, represents the speed of the vehicle at the future time k+τ; represents the target speed at the future k+τ time. 4.2 Constructing the chain vehicle situation field: Construction of the front situation field: According to the position and kinematic characteristics of the chain vehicle, the dynamic situation field of the front of the chain vehicle can be expressed as: in, k car A scaling factor representing the vehicle speed; According to the chain vehicle posture prediction, the situation field of the vehicle head at time k+τ is calculated: Construction of carriage situation field: The movement of the carriage relative to the front carriage is constrained by the hinge. The carriage of the chain vehicle is basically stationary relative to the front of the vehicle. The static situation field of carriage i is constructed: Where i = 1, 2, ...; k cartp The proportional factor representing the characteristics of the carriage; P cartp Indicates the characteristic size of the carriage. And there are: 4.3 Situation Field Fusion: Chain vehicle situation field: Chain vehicles construct situation fields for the locomotive and each carriage separately, and then merge the situation fields of each carriage into the situation field of the entire vehicle Where n is the total number of chain vehicle carriages; Environmental situation field: The situation fields of static, fixed and moving targets in the environment are fused to obtain the overall target situation field.

5. The early warning method according to claim 2, characterized in that: The steps of the collision warning method in step 2.6 are as follows: 5.1 Situation Field Discretization: The chain vehicle situation field and the environmental target situation field are projected onto the multi-valued local grid map respectively; the local grid map generated by the chain vehicle situation field projection is Grid veh,local ={Grid(i,j)|-M≤i≤M,-N≤j≤N}; The local grid map generated by the projection of the environmental target situation field is Grid obs,local ={Grid(i,j)∣-M≤i≤M,-N≤j≤N}; 5.2 Binarization of situation field: Set the safety threshold S0, and use the safety threshold to binarize the grid map of the chain vehicle situation field and the environmental target situation field: 5.3 Collision Warning: Design of Collision Warning Function F alert The calculation method is: convert the local grid map coordinates of the chained vehicle itself and the environmental target situation field to the global grid map coordinate system. veh,global With Grid obs,global Calculate the intersection of the two map grid coordinates and add them together; Collision warning function F alert : If F alert ≥1, the intersection of the two maps is considered not to be an empty set, the vehicles may collide, and a collision warning message is output. Otherwise, F alert =0, the intersection of the two grid maps is an empty set and there is no risk of collision.

Images