A vehicle blind area pedestrian perception and early warning method and system based on V2X

Abstract

The application discloses a kind of vehicle blind area pedestrian perception and early warning method and system based on V2X, and the motion state information of pedestrian in blind area is obtained by the environment perception system and V2X communication module of the blocking car and road test equipment of main car, the probability of occurrence of this kind of traffic accident is reduced in combination with environment perception system and V2X module.Encryption and decryption are carried out on information using classic asymmetric cryptographic algorithm-RSA algorithm, to prevent malicious network attack, guarantee normal communication and information security between main car and blocking car.Finally, the TTC of longitudinal and transverse is calculated respectively to judge collision risk, when judging as possible collision, pedestrian protection system starts indicator system, driver warning system and active brake system according to vehicle and pedestrian collision time TTC and the time threshold value that man and car do not collide, to realize hierarchical early warning.The application can effectively predict the relative position between pedestrian and vehicle in over-the-horizon scene, reduce the hidden danger brought by blind area, realize early warning and active collision avoidance function.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of pedestrian traffic safety, and in particular to a vehicle blind area pedestrian perception and warning method and system based on V2X. BACKGROUND

[0002] With the continuous development of the automobile industry and the increasing number of vehicles, the increasingly complex road environment has led to an increase in the rate of traffic accidents. As one of the most vulnerable groups among traffic participants, pedestrians account for a large proportion of deaths. A "ghost probe" is a colloquial term for a scenario in which a non-motor vehicle or pedestrian suddenly emerges from the roadside when the line of sight is blocked by a vehicle or obstacle in front. Because the driver does not avoid in time, neither has the time and space to react and avoid, often resulting in tragic accidents of vehicle injuries and deaths. A significant portion of traffic accidents in the country are caused by pedestrians or non-motor vehicles emerging from the driver's blind area, so there is an urgent need for a preventive measure to determine the likelihood of a collision and to alert the driver to reduce or even prevent such accidents from occurring.

[0003] Currently, there are many studies on preventing "ghost probes", which have certain effects, but the limitations are also very obvious. For example, relying on the front camera of the vehicle to monitor whether a moving object appears in front of the vehicle, this prevention method is often passive, and the moving object needs to enter the camera's field of view before taking appropriate emergency measures. At high speeds, it is difficult to avoid accidents. In addition, a "ghost probe" warning system based on brake light recognition identifies the brake lights of other vehicles in front of the host vehicle through the host vehicle's camera to determine whether a pedestrian is emerging from in front of the host vehicle. However, for vehicles temporarily parked on the roadside, the host vehicle cannot identify the brake lights and cannot determine whether a pedestrian is emerging from in front of the vehicle.

[0004] Therefore, active collision avoidance technology for pedestrian protection has become a hot topic in the research of automobile active safety technology, and it is urgent to solve the hidden dangers caused by blind areas. SUMMARY

[0005] In order to solve the problems in the prior art, the present application provides a vehicle blind area pedestrian perception and warning method and system based on V2X, which can receive pedestrian information in the blind area of the host vehicle through the vehicle-to-vehicle communication module when the pedestrian is blocked, solve the problem of blocked pedestrian perception, and avoid accidents.

[0006] The technical solutions adopted by the present application are as follows:

[0007] A vehicle blind area pedestrian perception and warning system based on V2X, comprising:

[0008] A blocked vehicle environment perception system is arranged on the blocking vehicle, and the blocked vehicle environment perception system perceives pedestrian information in the blind area of the host vehicle;

[0009] The shelter car information encryption and decryption system is arranged on the shelter car, receives and encrypts the pedestrian information input by the shelter car environment perception system;

[0010] The main car information encryption and decryption system is arranged on the main car, receives and decrypts the pedestrian information input by the shelter car information encryption and decryption system; the main car information encryption and decryption system and the shelter car information encryption and decryption system exchange information through V2X communication;

[0011] The main car environment perception system is arranged on the main car, and perceives the pedestrian information in front of the main car;

[0012] The main car information processing system is arranged on the main car, receives the pedestrian information input by the main car environment perception system and the pedestrian information decrypted by the main car information encryption and decryption system; the main car information processing system judges the intention of the pedestrian to cross the road based on the pedestrian information, and judges the collision according to the longitudinal and lateral TTC; if it is judged that the collision will occur, the main car information processing system controls the pedestrian protection system to work; if it is judged that the collision will not occur, the main car information processing system only needs to monitor the real-time collected pedestrian information;

[0013] The pedestrian protection system is arranged on the main car, and includes an indicator light system, a driver warning system and an active brake system; the pedestrian protection system starts the indicator light system, the driver warning system and the active brake system in turn according to the time to collision TTC of the vehicle and the pedestrian and the time threshold value at which the pedestrian and the vehicle do not collide.

[0014] Further, three time threshold values of the time at which the pedestrian does not collide are arranged respectively, and are represented as t1>t2>t3; if t2<TTC<t1, the indicator light system is started; if t3<TTC<t2, the driver warning system is started; if TTC≤t3, the active brake system is started.

[0015] Further, the main car environment perception system includes a camera, a millimeter wave radar, a laser radar and a speed encoder arranged on the main car, and the camera, the millimeter wave radar, the laser radar and the speed encoder are signal connected with the vehicle-mounted microprocessor of the main car; the vehicle-mounted microprocessor processes and extracts features according to the image of the surrounding environment of the main car collected by the camera, identifies whether the obstacle appearing in front of the main car from the blind area of the main car is a pedestrian, and tracks the trajectory of the pedestrian in real time.

[0016] Further, the occlusion vehicle environment perception system comprises a camera, a millimeter wave radar, a laser radar and a speed encoder arranged on the occlusion vehicle, and each unit in the occlusion vehicle environment perception system is in signal connection with the vehicle-mounted microprocessor of the occlusion vehicle; the vehicle-mounted microprocessor processes and extracts features according to the image collected by the camera, identifies whether the obstacle in front of the occlusion vehicle is a pedestrian, and tracks the trajectory of the pedestrian in real time.

[0017] Further, the main vehicle information encryption and decryption system and the occlusion vehicle information encryption and decryption system both internally embed RSA algorithm, and the transmitted information is encrypted or decrypted by using the RSA algorithm.

[0018] Further, the road test device further comprises a road test device environment perception system and a road test device information encryption and decryption system.

[0019] The road test device environment perception system comprises a camera, a millimeter wave radar, a laser radar, a speed encoder and a road test device processor arranged on the roadside, and is used for acquiring pedestrian information in the blind area of the main vehicle.

[0020] The road test device information encryption and decryption system receives and encrypts the pedestrian information input by the road test device environment perception system, and the road test device information encryption and decryption system and the main vehicle information encryption and decryption system adopt V2X communication to interact information.

[0021] Further, the vehicle-mounted microprocessor or the road test device processor internally tracks the trajectory of the pedestrian in real time by using Deepsort algorithm.

[0022] A vehicle blind area pedestrian perception and early warning method based on V2X comprises the following steps:

[0023] Step 1: The main vehicle information processing system receives the pedestrian information in the blind area of the main vehicle acquired by the environment perception system of the occlusion vehicle or the road test device environment perception system; the main vehicle information processing system receives and processes the pedestrian information to obtain the intention of the pedestrian to cross the road;

[0024] Step 2: For the pedestrian with the intention to cross the road, the risk of collision between the pedestrian and the main vehicle in the future is further judged;

[0025] Step 3: For the pedestrian with the risk of collision, the pedestrian protection system is started, and a three-level early warning mechanism is started in turn; according to the real-time calculation of the TTC of the pedestrian and the main vehicle,

[0026] First-level early warning mechanism: when the pedestrian protection system determines that the main vehicle and the pedestrian will collide, and the calculated t2 < TTC < t1, the indicator light system is started at this time to remind the driver that there is a pedestrian in front, so as to prepare the driver for braking;

[0027] Second level warning mechanism: when the indicator system is started, and the calculated t3 < TTC < t2, the driver warning system is started, and a warning signal is sent to the driver;

[0028] Third level warning mechanism: when the driver warning system is started, and the calculated TTC < t3, the active brake system is started, the safety belt of the host vehicle is tightened, and emergency braking measures are taken.

[0029] Further, the method for judging the risk of collision between the pedestrian and the host vehicle in the future in step 2 is:

[0030] The collision time TTC between the vehicle and the pedestrian is divided into longitudinal TTC L and lateral TTC T , and a longitudinal danger state judgment model TTC L is established, and the motion relationship equation is:

[0031]

[0032] In the formula, v px is the longitudinal speed of the pedestrian relative to the host vehicle; a px is the longitudinal acceleration of the pedestrian relative to the host vehicle; D is the relative distance between the center of the host vehicle and the center of the pedestrian; L = L1 + L2; L1 is the distance from the center of the host vehicle to the front bumper of the host vehicle; L2 is the distance from the center of the pedestrian to the body part that may collide with the host vehicle;

[0033] When Δ = v px 2 - 2a px (D-L) ≥ 0, the host vehicle and the pedestrian will collide in the longitudinal direction, and the TTC L at this time is:

[0034] When the pedestrian and the host vehicle are uniformly moving in the longitudinal direction, and the longitudinal speed of the pedestrian is less than the longitudinal speed of the host vehicle, that is, a px = 0 and v px < 0:

[0035]

[0036] When the longitudinal acceleration of the pedestrian relative to the host vehicle is less than 0, and the longitudinal speed of the pedestrian is less than the longitudinal speed of the host vehicle, that is, a px < 0 and v px < 0:

[0037]

[0038] The calculation formula of the lateral TTC T is as follows:

[0039] When the estimated collision point is at the rightmost side of the front bumper of the host vehicle, the lateral TTC at this time T is TTC T1 :

[0040]

[0041] When the estimated collision point is at the leftmost side of the front bumper of the host vehicle, the lateral TTC at this time T is TTC T2 :

[0042]

[0043] where S is the lateral relative distance between the pedestrian and the host vehicle; W0 is the width of the host vehicle; W is the width of the pedestrian; d is the minimum safety distance required to be maintained between the host vehicle and the pedestrian. p s

[0044] When TTC L ≤ TTC T1 and TTC L ≥ TTC T2 , i.e., the pedestrian and the host vehicle do not reach the collision point at the same time, the pedestrian and the host vehicle will not collide, which is a safe state, and the monitoring of the pedestrian in the blind area of the host vehicle continues.

[0045] When TTC T1 ≤ TTC L ≤ TTC T2 , i.e., the longitudinal collision time is within the range of the lateral collision time, the pedestrian and the host vehicle will collide.

[0046] Further, the method for determining the pedestrian crossing intention is as follows:

[0047] Step 1.1, coordinate conversion is performed on the pedestrian information to obtain the position of the pedestrian;

[0048] Step 1.2, the longitudinal speed v px and the lateral speed v py of the pedestrian are obtained based on the pedestrian information;

[0049] Step 1.3, an LSTM encoder and a decoder are used to predict the motion trajectory of the pedestrian;

[0050] Step 1.4, the road is divided into a safe area and a dangerous area, and based on the predicted motion trajectory of the pedestrian each time, if the last trajectory point of the predicted future t-second pedestrian trajectory falls within the safe area, it is determined that the pedestrian has no intention to cross the road; if the last trajectory point of the predicted future t-second pedestrian trajectory falls within the dangerous area, it is considered that the pedestrian has an intention to cross the road.

[0051] The present application has the following advantages:​​

[0052] (1) The vehicle blind area pedestrian perception method and early warning system based on V2X, the main feature is to apply V2X over-the-horizon perception to solve the problem of pedestrian collision avoidance, which can receive the pedestrian information in the blind area of the host vehicle through the vehicle communication module in the case of blocked pedestrians, solve the problem of blocked pedestrians unable to perceive, and ensure the safety of traffic participants.

[0053] (2) The vehicle blind area pedestrian perception method and early warning system based on V2X, the main feature is to calculate the longitudinal TTC and lateral TTC of the host vehicle and the pedestrian, to determine whether there is a danger of collision. And based on the TTC P when the warning is started W and the TTC B when the active brake is established, a pedestrian collision grading warning strategy is established, thereby reducing the safety hazards caused by "ghost probes", making the blind area no longer "blind". The grading warning strategy is beneficial to provide additional reaction time for the driver, to brake in advance, to avoid emergency braking to a certain extent, and to greatly improve the comfort of driving.

[0054] (3) The vehicle blind area pedestrian perception method and early warning system based on V2X, the main feature is to use RSA algorithm to encrypt information. V2X realizes efficient sharing of information and improves traffic safety, but due to the characteristics of a large number of users in the traffic environment and the open V2X network environment, attackers can easily interfere with the normal communication between vehicles through listening, tampering and replay attacks. Therefore, the RSA public key encryption algorithm is used to encrypt the information transmitted by the host vehicle and the blocking vehicle, improving the efficiency and safety of information transmission.

[0055] (4) In summary, the use of V2X technology can solve the problems existing in the prior art, and its over-the-horizon perception capability makes it possible to solve the "ghost probe" problem. In the case of blocked pedestrians, the vehicle communication module can receive pedestrian information in the blind area of the ego vehicle, solve the problem of blocked pedestrians unable to perceive, and avoid accidents. BRIEF DESCRIPTION OF DRAWINGS

[0056] Figure 1 is a schematic diagram of the method of the embodiment of the present application;

[0057] Figure 2 is a schematic diagram of the pedestrian trajectory and area division of the embodiment of the present application;

[0058] Figure 3 is a schematic diagram of the position relationship between the host vehicle, the blocking vehicle and the pedestrian of the embodiment of the present application;

[0059] Figure 4A method schematic diagram of the RSA algorithm of the embodiment of the present application is shown in the figure.

[0060] Figure 5 A main vehicle and pedestrian lateral distance information diagram of the embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical scheme and advantages of the present application more clear, the present application is further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0062] As shown in the figure, a vehicle blind area pedestrian perception and early warning system based on V2X designed by the present application includes a main vehicle and a shielding vehicle; wherein the main vehicle includes a main vehicle environment perception system, a main vehicle information encryption and decryption system, a main vehicle information processing system, and a pedestrian protection system. Figure 1 The shielding vehicle includes a shielding vehicle environment perception system and a shielding vehicle information encryption and decryption system.

[0063] The specific structure of each system is as follows:

[0064] The main vehicle environment perception system specifically includes a camera, a millimeter wave radar, a laser radar and a speed encoder arranged on the main vehicle, and the above-mentioned camera, millimeter wave radar, laser radar and speed encoder are signal connected between the vehicle-mounted microprocessor of the main vehicle, for perceiving the pedestrian information in front of the main vehicle, the pedestrian information perceived by the main vehicle includes pedestrian trajectory and motion state information (including pedestrian position, pedestrian speed). More specifically, the camera is used to collect the image of the environment around the main vehicle, and the collected image is processed and feature extracted by the vehicle-mounted microprocessor, to identify whether the obstacle appearing in front of the main vehicle from the blind area of the main vehicle is a pedestrian, and to track the trajectory of the pedestrian in real time, and to filter and process the pedestrian trajectory by using a Kalman filter to remove noise points. The position information of the pedestrian is obtained by the vehicle-mounted millimeter wave radar and laser radar, and the obtained information is input into the vehicle-mounted microprocessor; the speed information of the pedestrian is collected by the speed encoder, and the collected speed information is input into the vehicle-mounted microprocessor.

[0065]

[0066] ​The sheltered vehicle environment perception system specifically comprises a camera, a millimeter wave radar, a laser radar and a speed encoder arranged on the sheltered vehicle, and each unit in the sheltered vehicle environment perception system is in signal connection with the vehicle-mounted microprocessor of the sheltered vehicle, for perceiving pedestrian information in the blind area of the host vehicle. The pedestrian information perceived by the sheltered vehicle includes pedestrian trajectory and motion state information (including pedestrian position and pedestrian speed). More specifically, the camera is used to collect images of the environment around the sheltered vehicle, and the collected images are processed and feature extracted by the vehicle-mounted microprocessor, to identify whether the obstacle in front of the sheltered vehicle is a pedestrian, and to track the trajectory of the pedestrian in real time, and to filter the pedestrian trajectory by using a Kalman filter to remove noise. The position information of the pedestrian is obtained by the millimeter wave radar and the laser radar, and the obtained information is input into the vehicle-mounted microprocessor. The speed information of the pedestrian is collected by the speed encoder, and the collected speed information is input into the vehicle-mounted microprocessor.

[0067] The host vehicle information encryption and decryption system and the sheltered vehicle information encryption and decryption system both internally have RSA algorithm, which is used to encrypt or decrypt the transmitted information. V2X communication is adopted between the host vehicle information encryption and decryption system and the sheltered vehicle information encryption and decryption system, and V2X includes V2V and V2I. The information interaction between the host vehicle and the sheltered vehicle adopts V2V technology, and the information interaction between the host vehicle and the road test equipment adopts V2I technology.

[0068] More specifically, after the host vehicle enters the communication range, the sheltered vehicle encrypts the pedestrian information perceived by its environment perception system by using RSA algorithm, and sends the encrypted pedestrian information to the host vehicle information encryption and decryption system through the V2X communication module. After the host vehicle information encryption and decryption system decrypts the encrypted information sent by the sheltered vehicle by using RSA algorithm, the decrypted pedestrian information is input into the host vehicle information processing system.

[0069] In addition, after the host vehicle identifies the pedestrian by using its environment perception system, the host vehicle information encryption and decryption system encrypts the pedestrian information obtained by the host vehicle environment perception system, and sends the encrypted pedestrian information to the sheltered vehicle information encryption and decryption system through the V2X communication module. The sheltered vehicle information encryption and decryption system decrypts the encrypted information sent by the host vehicle by using RSA algorithm, to confirm whether the pedestrian information identified by the two is consistent, thereby realizing the information interaction between the two.

[0070] The host vehicle information processing system receives the pedestrian information collected by the host vehicle and the decrypted pedestrian information sent by the sheltered vehicle, judges the intention of the pedestrian to cross the road based on the pedestrian information in the host vehicle information processing system, and judges the collision based on longitudinal and lateral TTC. If it is judged that a collision will occur, the host vehicle information processing system controls the pedestrian protection system to work. If it is judged that a collision will not occur, the host vehicle information processing system only needs to monitor the real-time collected pedestrian information.

[0071] The pedestrian protection system comprises an indicator light system, a driver warning system and an active brake system; the indicator light system comprises an indicator light arranged in the cab, the driver warning system is a voice broadcast unit arranged in the cab, and the active brake system is an active brake system, a main vehicle safety belt unit and the like.

[0072] In the host vehicle information processing system, by comparing the pedestrian information obtained by the host vehicle environment perception system with the pedestrian information obtained by the occluded vehicle environment perception system received by the host vehicle information processing system, it can be determined whether the pedestrian information recognized by the two is consistent.

[0073] In the present embodiment, in addition to the V2X communication module between the host vehicle and the occluded vehicle, the application can also include a pedestrian mobile intelligent terminal, such as a smart phone, a smart watch and the like, for recording the pedestrian historical trajectory and position information, and sending it to the host vehicle through the V2P communication system. Since the V2P technology and the V2V and V2I technologies have similarities, the specific implementation is not specifically introduced. In actual scenarios, as long as at least one of the V2V, V2I and V2P technologies can be implemented, the effect of the application can be better achieved. When multiple V2X technologies can be implemented at the same time, the accuracy of the perceived pedestrian information can be better ensured.

[0074] The working principle of the vehicle blind area pedestrian perception and warning system based on V2X designed by the application is that the occluded vehicle identifies whether there is a pedestrian in front of the host vehicle through the camera, millimeter wave radar and laser radar thereon. If the occluded vehicle identifies that there is a pedestrian in the blind area of the host vehicle, the occluded vehicle tracks the trajectory of the pedestrian in real time through the camera and sensor, and informs the host vehicle of the pedestrian information obtained by the occluded vehicle through information transmission, so that the host vehicle knows the position of the pedestrian. When the pedestrian walks out of the blind area of the host vehicle, the sensor of the host vehicle is also started, and the environment perception systems of the two work at the same time, and the pedestrian information perceived by the two is compared in real time to ensure accurate tracking of the position of the pedestrian.

[0075] In the embodiment, if the sheltering vehicle does not identify the pedestrian due to the failure of the environmental perception system or other reasons, the pedestrian suddenly appears in front of the host vehicle after walking out of the blind area, which will cause the host vehicle driver to be unable to react; to solve the problem, the system designed by the application can also increase the road test equipment to perceive the pedestrian information in the blind area of the host vehicle. The road test equipment includes a road test equipment environmental perception system and a road test equipment information encryption and decryption system; more specifically, the road test equipment environmental perception system specifically includes: a camera, a millimeter wave radar, a laser radar, a speed encoder, a road test equipment processor arranged on the roadside; wherein the monocular CCD camera collects the image of the environment around the sheltering vehicle and inputs the road test equipment processor, and the road test equipment processor processes and extracts features from the collected image, identifies whether the obstacle in front of the sheltering vehicle is a pedestrian, and tracks the trajectory of the pedestrian in real time, and uses a Kalman filter to filter and process the pedestrian trajectory to remove noise points. The position information of the pedestrian is obtained by the millimeter wave radar and the laser radar, and the obtained information is input into the road test equipment processor; the speed information of the pedestrian is collected by the speed encoder, and the collected speed information is input into the road test equipment processor.

[0076] The road test equipment information encryption and decryption system is built-in RSA algorithm, which uses RSA algorithm to encrypt or decrypt the transmitted information. The road test equipment information encryption and decryption system and the host vehicle information encryption and decryption system adopt V2X communication, which can specifically adopt V2I technology.

[0077] Therefore, even when the sheltering vehicle cannot identify the pedestrian, the road test equipment can still perceive the pedestrian information and transmit it to the host vehicle.

[0078] In the embodiment, the Deepsort algorithm can be used in the vehicle-mounted microprocessor or the road test equipment processor to track the trajectory of the pedestrian in real time.

[0079] In the embodiment, as shown in Figure 4 When the host vehicle and the sheltering vehicle interact information, RSA algorithm is needed to encrypt the information to ensure safety. RSA is the most influential public key encryption algorithm at present, which has been recommended by ISO as the public key data encryption standard. This type of encryption algorithm has a pair of keys, one of which is used for encryption and the other is used for decryption. One of the pair of keys can be selected as a private key (self-preservation), and the other as a public key (publicly disclosed). The content encrypted by the private key can only be decrypted by the corresponding public key, and vice versa. The content encrypted by the public key can only be decrypted by the corresponding private key, which ensures the safety and accuracy of information interaction.

[0080] Based on the above-mentioned vehicle blind area pedestrian perception and warning system based on V2X, the application also designs a vehicle blind area pedestrian perception and warning method based on V2X, which includes the following steps:

[0081] Step 1: Obtain pedestrian information in the blind spot of the main vehicle through the vehicle's environmental perception system (or the roadside equipment's environmental perception system); the main vehicle's information processing system receives the pedestrian information and processes it to determine the pedestrian's intention to cross the road. For example... Figure 2 As shown, since the road testing equipment and the environmental perception system of the obstructing vehicle serve the same function, the scenario involving only the obstructing vehicle will be explained here. When the obstructing vehicle detects a pedestrian in its blind spot, it transmits the pedestrian information to the main vehicle. The main vehicle's information processing system uses coordinate system transformation to calculate the relative positions of the pedestrian, the obstructing vehicle, and the main vehicle, facilitating the determination of information such as distances between them. In the scenario shown, due to certain reasons, such as waiting at a red light or traffic congestion, a car is stopped in the right lane (the obstructing vehicle), and a pedestrian is about to cross the road in front of it. A car is about to pass in the left lane (the main vehicle). At this time, due to the obstruction of the obstructing vehicle, the pedestrian is in the main vehicle's blind spot, and if no action is taken, a traffic accident is highly likely.

[0082] Step 1.1: Perform coordinate transformation on pedestrian information to obtain pedestrian location.

[0083] In V2X communication, vehicles share various information, such as position, speed, and heading angle. The position information received by vehicles is typically expressed in latitude and longitude, i.e., WGS-84 coordinates. This coordinate system is not convenient for calculating information such as relative distances between vehicles. Therefore, the WGS-84 coordinate system needs to be converted to the Gaussian plane coordinate system using the Gauss-Kruger projection. The conversion relationship between WGS-84 coordinates (B, L) and Gaussian plane coordinates (x, y) is as follows:

[0084]

[0085]

[0086] In the formula, X0 is the arc length from the equator to the parallel meridian at latitude B; l is the difference between the longitude L of the projection point and the longitude of the meridian of the longitude zone where the point is located; e′ is the second eccentricity; N is the radius of curvature of the zonal circle; and for the sake of simplifying the writing of the formula, l is represented by a. 4 (5-t 2 +9μ 2 +4μ 2 )·cos 2 B, therefore it can be written as a = l 4 (5-t 2 +9μ 2 +4μ 2 )·cos 2 B; b, c, and d are similar, representing b = l respectively. 4 (61-58t 2 +t 4 cos 4B; c = 1 2 (1 - t 2 + μ 2 )cos 2 B; d = 1 4 (5 - 18t 2 + t 4 + 14μ 2 - 58μ 2 t 2 )cos 4 B; μ = e'cosB;

[0087] Since the Gaussian plane coordinate system is with the central meridian projection as the x axis and the equatorial projection as the y axis, in subsequent calculations, the Gaussian plane coordinates (x, y) need to be converted into Cartesian coordinates (X, Y) to facilitate the input of pedestrian information into the LSTM encoder:

[0088]

[0089] The information processing system assumes that the global coordinate system at this time is XOY, the occlusion vehicle coordinate system is X'O'Y', and the host vehicle coordinate system is X''O''Y''. (X1, Y1) is the coordinate of the occlusion vehicle in the global coordinate system XOY, (X2, Y2) is the coordinate of the pedestrian in the global coordinate system XOY, (X'2, Y'2) is the coordinate of the pedestrian in the vehicle coordinate system X'O'Y', and (X3, Y3) is the coordinate of the host vehicle in the global coordinate system XOY.

[0090] When the occlusion vehicle heading angle α0∈[0°, 180°], the coordinates (X2, Y2) of the pedestrian in the global coordinate system XOY can be obtained according to the coordinates (X'2, Y'2) of the pedestrian in the occlusion vehicle coordinate system X'O'Y' and the coordinates (X1, Y1) of the occlusion vehicle in the global coordinate system XOY:

[0091]

[0092] When α0∈[180°, 360°], the coordinates of the pedestrian in the global coordinate system XOY can be expressed as:

[0093]

[0094] Step 1.2, based on the pedestrian information, the longitudinal speed v px and the lateral speed v py

[0095] Then the predicted longitudinal distance D, lateral distance S and direction angle γ of the host vehicle and the pedestrian at the collision position are respectively:

[0096] D = D2cos(α h- β) + Stan(γ + α h - α0)

[0097] S = D2sin(α h - β)

[0098]

[0099] wherein, α h is the host vehicle heading angle, i.e. the angle between the vehicle driving direction and the y-axis; α0is the sheltered vehicle heading angle, i.e. the angle between the vehicle driving direction and the y-axis; v px1 is the longitudinal speed of the pedestrian in the sheltered vehicle coordinate system X'O'Y'; v py1 is the lateral speed of the pedestrian in the sheltered vehicle coordinate system X'O'Y'; D2is the relative distance between the host vehicle and the pedestrian; β is the angle between the line connecting the host vehicle and the pedestrian and the y-axis.

[0100] The pedestrian speed v p is transformed to obtain the longitudinal speed v px and the lateral speed v py of the pedestrian relative to the host vehicle coordinate system:

[0101]

[0102] Step 1.3, after the host vehicle obtains the pedestrian position and motion state information, the pedestrian motion trajectory needs to be predicted to realize real-time tracking of the pedestrian position. LSTM encoder and decoder are used to realize the input of the pedestrian historical trajectory and the output of the predicted trajectory.

[0103] The input of the LSTM encoder is:

[0104]

[0105] wherein, represents embedding the pedestrian coordinate position to obtain a feature vector as the input of the LSTM encoder; is the pedestrian position coordinate, i.e. W n is the embedding weight; is the hidden state of the pedestrian at the previous time step, such as speed information; is the weight of the LSTM unit.

[0106] In order to predict the trajectory coordinates of the pedestrian, the encoder is used as the input of the LSTM decoder, and then the prediction process is as follows:

[0107]

[0108]

[0109]

[0110] In the formula, is the estimated mean value at t+1 time; is the variance at t+1 time; is the correlation coefficient at t+1 time.

[0111] The position information of the pedestrian at the next time can be predicted, because the sheltered car, the road testing equipment and the host car perform real-time information interaction, the host car continuously processes and judges the real-time updated information,

[0112] Step 1.4, based on the predicted pedestrian motion trajectory each time, the intention of the pedestrian crossing the road is judged, and corresponding decisions are made according to the intention of the pedestrian crossing the road, the method for judging the intention of the pedestrian crossing the road is:

[0113] As shown in Figure 3 , the road is divided into a safe area and a dangerous area, and the range with a distance of d s from both sides of the road and a width of d w is set as the safe area; the dangerous area is between the two areas, and the width of the dangerous area is d L .

[0114] If the last trajectory point of the predicted future t-second pedestrian trajectory A falls within the safe area, it is judged that the pedestrian has no intention to cross the road and will not collide with the host car, and the monitoring of the pedestrian in the blind area of the host car is continued. If the last trajectory point of the predicted future t-second pedestrian trajectory B falls within the dangerous area, it can be considered that the pedestrian has the intention to cross the road.

[0115] Step 2, for the pedestrian with the intention to cross the road, the risk of collision between the pedestrian and the host car in the future needs to be further judged.

[0116] TTC is used to represent the collision time of the vehicle and the pedestrian, since the trajectory of the pedestrian has uncertainty, the pedestrian has speed in the lateral direction and the longitudinal direction, in order to accurately estimate the risk of collision, the TTC is divided into longitudinal TTC T and lateral TTC L . The calculation and judgment process of TTC is as follows:

[0117] During the driving of the host car, the TTC px longitudinal danger state judgment model can be established according to the motion state information of the host car and the pedestrian information received by the host car, and the motion relationship equation is:

[0118]

[0119] In the formula, v pxD is the relative distance between the center of the host vehicle and the center of the pedestrian; L = L1 + L2; L1 is the distance from the center of the host vehicle to the front bumper of the host vehicle; L2 is the distance from the center of the pedestrian to the body part that is likely to collide with the host vehicle.

[0120] The above motion relationship equation is a quadratic equation, and when Δ = v px 2 -2a px When (D - L) > 0, the host vehicle and the pedestrian will collide in the longitudinal direction, and at this time, TTC L is:

[0121] When the pedestrian and the host vehicle are both moving at a constant speed in the longitudinal direction, and the longitudinal speed of the pedestrian is less than that of the host vehicle, i.e., a px = 0 and v px < 0:

[0122]

[0123] When the longitudinal acceleration of the pedestrian relative to the host vehicle is less than 0, and the longitudinal speed of the pedestrian is less than that of the host vehicle, i.e., a px < 0 and v px < 0:

[0124]

[0125] As shown in Figure 5 , in addition to detecting whether there is a pedestrian in the front longitudinal position of the vehicle during driving, it also needs to detect whether there is a pedestrian in the front transverse position of the vehicle. The pedestrian will have a range of collision points in the transverse direction and the host vehicle, i.e., from the left side to the right side of the front bumper of the host vehicle, which can become a collision point. The calculation formula of transverse TTC T is as follows:

[0126] When the estimated collision point is at the rightmost side of the front bumper of the host vehicle, the transverse TTC T at this time is TTC T1 :

[0127]

[0128] When the estimated collision point is at the leftmost side of the front bumper of the host vehicle, the transverse TTC T at this time is TTC T2 :

[0129]

[0130] In the formula, S is the transverse relative distance between the pedestrian and the host vehicle; W0 is the width of the host vehicle; W p is the width of the pedestrian; d s is the minimum safety distance required to be maintained by the host vehicle and the pedestrian.

[0131] When TTC L ≤ TTC T1 and TTC L ≥ TTC T2 , i.e. the pedestrian and the host vehicle do not reach the collision point at the same time, the pedestrian and the host vehicle will not collide, and this is a safe state, and the above monitoring of the pedestrian in the blind area of the host vehicle continues.

[0132] When TTC T1 ≤ TTC L ≤ TTC T2 , i.e. the longitudinal collision time is within the range of the lateral collision time, the pedestrian and the host vehicle will collide.

[0133] Step 3, for the pedestrian at risk of collision, start the pedestrian protection system, and sequentially start the three-level warning mechanism.

[0134] The pedestrian protection system includes an indicator system, a driver warning system, and an active brake system. After receiving the instruction of the information processing system, the pedestrian protection system is started, which is a forward-looking and predictable safety system for receiving the positional relationship between the pedestrian and the vehicle, starting the pedestrian protection mechanism, taking warning and braking measures to avoid or reduce the collision with the pedestrian. The three-level warning mechanism is sequentially started, and the TTC of the pedestrian and the host vehicle is calculated in real time.

[0135] First-level warning mechanism: when the pedestrian protection system determines that the host vehicle and the pedestrian will collide, and the calculated t2 < TTC < t1, the TTC at this time is recorded as TTC P , at this time, the driver needs to be reminded that there is a pedestrian in front of the vehicle, the indicator system is started, and the indicator in the instrument panel is turned on to prepare the driver for braking. t1 is the shortest time for braking at a relatively comfortable deceleration (such as less than 2.5 m / s 2 ) and not colliding with the pedestrian.

[0136] Second-level warning mechanism: when the indicator system is started, and the calculated t3 < TTC < t2, the TTC at this time is recorded as TTC W ,

[0137] The driver warning system is started, and the warning signal output end of the warning system sends a warning signal to the driver, which provides additional reaction time for the driver and is beneficial to make better judgments. t2 is the shortest time for considering the driver's reaction time and not colliding with the pedestrian.

[0138] Third-level warning mechanism: when the driver warning system is started, and the calculated TTC ≤ t3, the TTC at this time is recorded as TTC Bt3 is the shortest time required for the active brake system to take over the control of the vehicle to avoid the collision danger, and the host vehicle is determined to collide with the pedestrian, the active brake system is started, the safety belt of the host vehicle is tightened, and the emergency brake is taken.

[0139] In summary, the present application firstly detects whether there is a pedestrian in front of the shelter through the environment perception system of the shelter and road testing equipment, and records the trajectory and position information of the pedestrian in real time; at the same time, the host vehicle enters the communication range, the information encryption and decryption system of the shelter and road testing equipment encrypts the blind area pedestrian information perceived by the shelter and sends it to the host vehicle based on the V2X module; then, when the pedestrian walks out of the blind area, the host vehicle environment perception system detects that the pedestrian appears in front of the host vehicle, and the host vehicle information encryption and decryption system encrypts the information perceived by the host vehicle and sends it to the shelter and road testing equipment based on the V2X module, and decrypts the information transmitted by the shelter and road testing equipment, the host vehicle and the shelter and road testing equipment confirm whether the information of the pedestrian is correct while interacting with each other; the host vehicle information processing system is used to process the relationship between the pedestrian and the vehicle, predict the future t second trajectory of the pedestrian based on the historical trajectory, and judge whether the pedestrian has the intention to cross the road according to the trajectory point; then, when it is judged that the pedestrian has the intention to cross the road, the pedestrian protection system starts the pedestrian protection mechanism, and calculates the TTC in the horizontal and vertical directions to judge whether the collision will occur; finally, as the TTC decreases, the indicator light system, the driver warning system and the active brake system are started in turn, which are used to remind the driver that there is a pedestrian in front of the vehicle, warn the driver to take braking measures, and the host vehicle actively brakes when necessary.

[0140] The above examples are only used to illustrate the design idea and characteristics of the present application, and the purpose is to enable those skilled in the art to understand the content of the present application and implement it, and the protection scope of the present application is not limited to the above examples. Therefore, any equivalent changes or modifications made according to the principles and design ideas disclosed by the present application are within the protection scope of the present application.

Claims

1. A method for V2X-based vehicle blind area pedestrian perception and early warning, characterized in that, The V2X-based vehicle blind area pedestrian perception and warning system comprises the following steps: Step 1: The host vehicle information processing system receives the pedestrian information in the blind area of the host vehicle obtained by the environmental perception system or the road test device environmental perception system of the occlusion vehicle; the host vehicle information processing system receives and processes the pedestrian information to obtain the pedestrian crossing road intention; The method for determining the pedestrian crossing road intention is as follows: Step 1.1: Coordinate conversion is performed on the pedestrian information to obtain the pedestrian position; Step 1.

2. Obtain the longitudinal velocity v of the pedestrian based on the pedestrian information px and the lateral velocity v py ; Step 1.3: An LSTM encoder and decoder are used to predict the pedestrian motion trajectory; Step 1.4: The road is divided into a safe area and a dangerous area, and based on the predicted pedestrian motion trajectory each time, if the last trajectory point of the predicted pedestrian motion trajectory in the future t seconds falls within the safe area, it is determined that the pedestrian has no intention to cross the road; If the last trajectory point of the predicted pedestrian motion trajectory in the future t seconds falls within the dangerous area, it is considered that the pedestrian has the intention to cross the road; Step 2: For the pedestrian with the intention to cross the road, the risk of collision between the pedestrian and the host vehicle in the future is further determined; the method for determining the risk of collision between the pedestrian and the host vehicle in the future in step 2 is as follows: Vehicle-pedestrian collision time (TTC) is divided into longitudinal TTC. L and horizontal TTC T Establish TTC L The motion relationship equation for the longitudinal hazard assessment model is as follows: where v px is the longitudinal speed of the pedestrian relative to the host vehicle; a px is the longitudinal acceleration of the pedestrian relative to the host vehicle; D is the relative distance between the host vehicle center and the pedestrian center; L = LI + L2; LI is the distance from the host vehicle center to the host vehicle front bumper; L2 is the distance from the pedestrian center to the body part that can collide with the host vehicle. When Δ = v px 2 -2a px When (D-L) ≥ 0, the main vehicle and the pedestrian will collide in the longitudinal direction, and the TTC at this time is: L TTC = D - L When the pedestrian and the host vehicle are both driving at a constant speed in the longitudinal direction, and the longitudinal speed of the pedestrian is less than that of the host vehicle, i.e. a px = 0 and v px < 0: When the longitudinal acceleration of the pedestrian with respect to the host vehicle is less than 0 and the longitudinal velocity of the pedestrian is less than the longitudinal velocity of the host vehicle, i.e. a px <0 and v px <0: Lateral TTC T The formula for the calculation is as follows: When the estimated collision point is at the rightmost side of the host vehicle's front bumper, the lateral TTC at this time T is TTC T1 : When the estimated collision point is at the leftmost side of the host vehicle's front bumper, the lateral TTC at this time T is TTC T2 : where S is the lateral relative distance between the pedestrian and the host vehicle; W0is the width of the host vehicle; W p is the width of the pedestrian; d s is the minimum safety distance required to be maintained between the host vehicle and the pedestrian; When TTC L ≤ TTC T1 and TTC L ≥ TTC T2 , i.e. the pedestrian and the host vehicle do not reach the collision point at the same time, the pedestrian and the host vehicle will not collide, and this is a safe state, and the monitoring of the pedestrian information in the blind area of the host vehicle continues. When TTC T1 ≤ TTC L ≤ TTC T2 , i.e. the longitudinal collision time is within the range of the lateral collision time, the pedestrian and the host vehicle will collide. Step 3: For the pedestrian with the risk of collision, a pedestrian protection system is started, and a three-level warning mechanism is started in turn; According to the real-time calculated TTC of the pedestrian and the host vehicle, The first level warning mechanism: when the pedestrian protection system determines that the host vehicle and the pedestrian will collide, and the calculated t2 The second level warning mechanism: when the indicator light system is started, and the calculated t3 The third level warning mechanism: when the driver warning system is started, and the calculated TTC≤t3, the active brake system is started, the safety belt of the host vehicle is tightened, and emergency brake measures are taken; t1 is the shortest time for deceleration to brake and not collide with the pedestrian, t2 is the shortest time for considering the driver reaction time and not colliding with the pedestrian, and t3 is the shortest time required for the active brake system to take over the vehicle control to avoid collision danger.

2. The V2X-based vehicle blind area pedestrian perception and warning method of claim 1, wherein, The V2X-based vehicle blind area pedestrian perception and warning system comprises: An occlusion vehicle environmental perception system arranged on the occlusion vehicle, which perceives the pedestrian information in the blind area of the host vehicle; An occlusion vehicle information encryption and decryption system arranged on the occlusion vehicle, which receives and encrypts the pedestrian information input by the occlusion vehicle environmental perception system; A host vehicle information encryption and decryption system arranged on the host vehicle, which receives and decrypts the pedestrian information input by the occlusion vehicle information encryption and decryption system; the host vehicle information encryption and decryption system and the occlusion vehicle information encryption and decryption system exchange information through V2X communication; A host vehicle environmental perception system arranged on the host vehicle, which perceives the pedestrian information in front of the host vehicle; The main vehicle information processing system arranged on the main vehicle receives pedestrian information input by the main vehicle environment perception system and the pedestrian information decrypted by the main vehicle information encryption and decryption system; the main vehicle information processing system judges the intention of the pedestrian crossing the road based on the pedestrian information and makes collision judgment according to longitudinal and lateral TTC; if it is judged that a collision will occur, the main vehicle information processing system controls the pedestrian protection system to work; if it is judged that a collision will not occur, the main vehicle information processing system only needs to monitor the real-time collected pedestrian information; The pedestrian protection system arranged on the main vehicle includes an indicator light system, a driver warning system and an active brake system; the pedestrian protection system starts the indicator light system, the driver warning system and the active brake system in turn according to the time to collision TTC of the vehicle and the pedestrian and the time threshold value at which the pedestrian and the vehicle do not collide; Three time threshold values of the time at which the pedestrian does not collide are arranged respectively, which are represented as t1>t2>t3; if t2<TTC<t1, the indicator light system is started; if t3<TTC<t2, the driver warning system is started; if TTC≤t3, the active brake system is started; The road test device includes a road test device environment perception system and a road test device information encryption and decryption system; The road test device environment perception system includes a camera, a millimeter wave radar, a laser radar, a speed encoder and a road test device processor arranged on the roadside, which are used to obtain pedestrian information in the blind area of the main vehicle; The road test device information encryption and decryption system receives and encrypts the pedestrian information input by the road test device environment perception system, and the road test device information encryption and decryption system and the main vehicle information encryption and decryption system exchange information through V2X communication. 3.The V2X-based vehicle blind area pedestrian perception and warning method of claim 2, wherein, The main vehicle environment perception system includes a camera, a millimeter wave radar, a laser radar and a speed encoder arranged on the main vehicle, which are signal connected with the vehicle-mounted microprocessor of the main vehicle; the vehicle-mounted microprocessor processes and extracts features according to the image of the surrounding environment of the main vehicle collected by the camera, identifies whether the obstacle appearing in front of the main vehicle from the blind area of the main vehicle is a pedestrian, and tracks the trajectory of the pedestrian in real time.

4. The V2X-based vehicle blind area pedestrian perception and warning method of claim 2, wherein, The shelter vehicle environment perception system includes a camera, a millimeter wave radar, a laser radar and a speed encoder arranged on the shelter vehicle, and each unit in the shelter vehicle environment perception system is signal connected with the vehicle-mounted microprocessor of the shelter vehicle; the vehicle-mounted microprocessor processes and extracts features according to the image collected by the camera, identifies whether the obstacle in front of the shelter vehicle is a pedestrian, and tracks the trajectory of the pedestrian in real time.

5. The V2X-based vehicle blind area pedestrian perception and warning method of claim 2, wherein, The main vehicle information encryption and decryption system and the shelter vehicle information encryption and decryption system both internally embed RSA algorithm, which is used to encrypt or decrypt the transmitted information.

6. The V2X-based vehicle blind area pedestrian perception and warning method of claim 2, wherein, The vehicle-mounted microprocessor or the road test device processor tracks the trajectory of the pedestrian in real time through Deepsort algorithm.

Images