A vehicle blind spot cooperative perception system and method based on roadside sensors

Abstract

This invention discloses a vehicle blind spot collaborative perception system and method based on roadside sensors, belonging to the field of intelligent vehicle technology. The system, based on roadside sensors, comprises a blind spot recognition module, a risk assessment module, and a perception information fusion module. It uses roadside sensor perception information to identify and classify blind spots in various vehicle scenarios, accurately assess blind spot risks in real time, and perform spatiotemporal fusion and feedback compensation of the perception information to achieve collaborative perception.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent vehicles, and in particular relates to a vehicle blind spot collaborative perception system and method based on roadside sensors. Background Art

[0002] Traffic accidents can cause casualties and significant economic losses, and blind spots are one of the main causes. Blind spots seriously impact driving safety, and the hazards they pose cannot be ignored. Whether it's autonomous vehicles or traditional vehicles, there's a risk of accidents caused by blind spots.

[0003] The vehicle's inherent structure inevitably creates blind spots. At the same time, the vehicle's driving is also affected by the terrain, obstructions from buildings at intersections, and obstructions from other vehicles, which can also create blind spots.

[0004] Currently, research on vehicle blind spot safety is primarily based on single-vehicle intelligence solutions, but these solutions also suffer from shortcomings such as limited perception range and susceptibility to occlusion. Risk assessment of vehicle blind spots is key to reducing traffic accidents caused by blind spots. Currently, there is limited research in this area. Some studies are based on data-driven models. For example, some scholars use regional characteristics as a discrimination criterion, estimating the collision risk of potential vehicles based on the relationship between area and vehicle speed. However, this approach lacks research on the properties of blind spots themselves. Most studies lack universality, often focusing on a single scenario and limiting their use cases. Other studies use probability-driven methods, predicting the risk of blind spots based on the size of the blind spots constructed. However, these models are relatively subjective and therefore cannot accurately predict the risks that may exist in blind spots.

[0005] Chinese patent publication number CN115052266A, titled "A Blockchain-Based Vehicle Blind Spot Detection Method and Application System in a Vehicle Network Environment," primarily provides a blockchain-based vehicle blind spot detection method and application system. The method first obtains basic information about the target vehicle and other traffic participants within the target vehicle's blind spot. Then, through a smart contract, the system automatically executes the use of vehicle network technology to send the basic information of other traffic participants within the target vehicle's blind spot to the target vehicle. While blockchain-based smart contracts enable effective information exchange between the two parties, the invention lacks the necessary spatiotemporal integration of the basic information of other traffic participants. Consequently, the target vehicle cannot directly use this information for effective response operations. Its single function fails to effectively achieve the purpose of vehicle-road collaborative perception.

[0006] Chinese patent publication number CN114155713A, titled "A Method and System for Predicting Blind Spots at Intersections Based on Roadside Crowd Intelligence Computing," utilizes a crowd intelligence computing system to collect vehicle information, then uses this information to predict the trajectories of pedestrians and vehicles, using traffic lights to warn pedestrians and vehicles at intersections to slow down. However, this invention lacks universal applicability due to its lack of research into the properties of blind spots and its focus on a single intersection scenario. Summary of the Invention

[0007] The purpose of the present invention is to overcome the shortcomings and deficiencies of the existing technology and provide a vehicle blind spot collaborative perception system and method based on roadside sensors. The purpose is to use the perception information of roadside sensors to identify and divide the blind spots of vehicles in various scenarios, and accurately assess the blind spot risks in real time, and perform spatiotemporal fusion and feedback compensation of the perception information to achieve the purpose of collaborative perception.

[0008] On the one hand, the present invention provides a vehicle blind spot cooperative perception system based on roadside sensors, including a blind spot recognition module, a risk assessment module and a perception information fusion module;

[0009] The blind spot recognition module includes a roadside perception module and a blind spot type recognition module. The roadside perception module is used to obtain roadside perception data and extract traffic participant information based on the roadside perception data. The blind spot type recognition module determines whether there is a blind spot for the main vehicle based on the roadside perception data. If there is a blind spot for the main vehicle, it identifies the type of the blind spot and obtains the location and area of ​​the blind spot.

[0010] The risk assessment module is used to determine traffic participants that are considered essential risk. If a blind spot exists for the host vehicle, the module predicts the trajectories of traffic participants within the blind spot based on the identified location and area of ​​the blind spot, and assesses the collision risk between the host vehicle and other traffic participants within the blind spot. Based on the collision risk, the module determines the traffic participants that are considered essential risk.

[0011] The perception information fusion module includes a perception information feedback compensation module and a perception fusion module;

[0012] The perception information feedback compensation module is used to perform spatiotemporal fusion on the information of traffic participants with necessary risks to obtain spatiotemporal fusion information; the spatiotemporal fusion information is transmitted back to the host vehicle through the roadside perception module;

[0013] The perception fusion module is used to fuse spatiotemporal fusion information and compensate for the information of traffic participants that the host vehicle has not recognized in its blind spot.

[0014] Optionally, the roadside perception module includes a roadside sensor and an edge server; the roadside sensor is used to detect all roadside perception data within the detection range; the edge server is equipped with a roadside sensor detection and calculation module, which is used to obtain roadside perception data from the roadside sensor, identify traffic participant information, and calculate traffic participant information to obtain traffic participant information.

[0015] Optionally, the traffic participant information includes information such as the location, speed, heading angle, category and location coordinates of the traffic parameter person.

[0016] Optionally, the roadside sensor recognizes global information from a bird's-eye view and obtains roadside perception data within the detection range.

[0017] Optionally, the types of the host vehicle's blind spots include vehicle obstruction blind spots and building obstruction blind spots.

[0018] On the other hand, the present invention provides a vehicle blind spot cooperative perception method based on roadside sensors, the specific steps are as follows:

[0019] Step 1: Acquire roadside sensing data and obtain traffic participant information based on the roadside sensing data;

[0020] Step 2: Determine the main vehicle from all vehicles in the traffic participant information;

[0021] Step 3: Determine whether there is a blind spot for the host vehicle based on the roadside sensing data; if there is a blind spot for the host vehicle, identify the type of the blind spot and obtain the location and boundary of the blind spot, thereby obtaining the area of ​​the blind spot;

[0022] Step 4: Determine whether there are other traffic participants in the blind spot of the host vehicle;

[0023] If there are other traffic participants, go to step 5;

[0024] Step 5: Based on the identified location and area of ​​the blind spot of the host vehicle, predict the movement trajectory of traffic participants within the area of ​​the blind spot of the host vehicle;

[0025] Step 6: Based on the predicted trajectories of traffic participants within the blind spot of the host vehicle, the collision risk between the host vehicle and other traffic participants within the blind spot of the host vehicle is evaluated to obtain a risk assessment result.

[0026] Step 7: Obtain necessary risk traffic participants based on risk assessment results;

[0027] If it is necessary to risk traffic participants, proceed to step 8;

[0028] Step 8: Perform spatiotemporal fusion on the information of the traffic participants with necessary risks to obtain spatiotemporal fusion information; and transmit the spatiotemporal fusion information back to the host vehicle;

[0029] Step 9: The main vehicle integrates the spatiotemporal fusion information.

[0030] Optionally, the types of blind spots of the main vehicle include vehicle obstruction blind spots and building obstruction blind spots; based on the heading angle and position coordinate information of the traffic participant, the corresponding obstruction blind spot boundary line is obtained.

[0031] Optionally, if the blind spot of the host vehicle is a building-blocked blind spot, the specific steps for obtaining the building blind spot boundary are as follows:

[0032] Obtain the pre-calibrated building corner points, and use the connecting line between the main vehicle center coordinates and the building corner points to divide the vehicle's sight area and blind spot boundary line. The expression of the building blind spot boundary line is as follows:

[0033]

[0034]

[0035] Among them, x0, y0 are the coordinates of the center point of the main vehicle, and x1, y1, x2, y2 are the coordinates of the two corner points of the building closest to the main vehicle.

[0036] Optionally, if the blind spot type of the host vehicle is a vehicle obstruction blind spot, the specific steps for obtaining the boundary line of the vehicle obstruction blind spot are as follows:

[0037] The blind spot of vehicles within the preset radius of the main vehicle is modeled and divided. The coordinates of the four rectangular boxes of the obstructing vehicles are expressed as follows:

[0038]

[0039]

[0040]

[0041]

[0042] Among them, x0, y0 are the coordinates of the center point of the main vehicle, x1, y1 are the coordinates of the center point of the occluding vehicle, l, w are the length and width of the occluding vehicle respectively, and yaw is the heading angle of the occluding vehicle;

[0043] The coordinates of the center point of the main vehicle and the four endpoints of the obstructing vehicle are connected as lines. The two connecting lines with the largest angle between the main vehicle and the four endpoints are taken as the boundary lines of the vehicle blind spot. The expression is:

[0044]

[0045]

[0046] Optionally, the method further includes step 10, traversing all other vehicles in the traffic participant information, selecting another vehicle as the main vehicle, and repeating the above steps until all vehicles are traversed.

[0047] The present invention has at least the following beneficial effects:

[0048] (1) This invention expands the perception range of autonomous vehicles through vehicle-road cooperative technology, enabling them to obtain more complete environmental information, including road conditions, vehicles, pedestrians, or other obstacles. This enables the vehicle to detect and avoid potential collisions in advance and better plan its driving path, thereby providing the vehicle with more accurate perception of the environment and improving traffic efficiency and safety.

[0049] (2) The present invention introduces roadside sensor equipment and edge computing equipment into vehicle blind spot hazard warning, solving the problem that the vehicle's own sensors cannot accurately predict the vehicle's blind spot risks in real time. Through the roadside sensor equipment, various static and dynamic factors in the traffic environment, such as buildings, vehicles and various traffic participants, are intelligently identified. Then, based on this information, the blind spots are scientifically divided and the risks are assessed, and then each vehicle is served, realizing real-time multi-vehicle risk blind spot information analysis and information feedback compensation to achieve the purpose of collaborative perception.

[0050] (3) The present invention can be used not only for self-driving cars, but also for traditional cars. The blind spot warning function can be used to remind traditional vehicles of dangers from invisible areas.

[0051] In the present invention, the above-mentioned technical solutions can be combined with each other to achieve more preferred combination solutions. Other features and advantages of the present invention will be described in the subsequent description, and some advantages will become apparent from the description or be understood through practice of the present invention. The objectives and other advantages of the present invention can be achieved and obtained through the contents particularly pointed out in the description and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The accompanying drawings are only for the purpose of illustrating particular embodiments and are not to be considered limiting of the present invention. Like reference symbols denote like parts throughout the drawings.

[0053] Figure 1 Schematic diagram of the vehicle blind spot cooperative perception system based on roadside sensors of the present invention.

[0054] Figure 2 This is a schematic diagram of the classification of the blind spots of the host vehicle according to the present invention.

[0055] Figure 3 This is a flow chart of the vehicle blind spot collaborative perception method of the present invention. DETAILED DESCRIPTION

[0056] The preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings, wherein the accompanying drawings constitute a part of the present invention and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.

[0057] A specific embodiment of the present invention, as Figure 1 As shown, a vehicle blind spot cooperative perception system based on roadside sensors is disclosed, including a blind spot recognition module, a risk assessment module and a perception information fusion module;

[0058] The blind spot recognition module includes a roadside perception module and a blind spot type recognition module. The roadside perception module is used to obtain roadside perception data and extract traffic participant information based on the roadside perception data. The blind spot type recognition module determines whether there is a blind spot for the host vehicle based on the roadside perception data. If there is a blind spot for the host vehicle, the module identifies the type of the blind spot for the host vehicle and obtains the location and area of ​​the blind spot. If there is no blind spot for the host vehicle, the module terminates the determination.

[0059] The risk assessment module is used to determine traffic participants that are considered to be at a necessary risk. If a blind spot exists for the host vehicle, the module predicts the trajectories of traffic participants within the blind spot of the host vehicle based on the identified location and area of ​​the blind spot, and assesses the collision risk between the host vehicle and other traffic participants within the blind spot of the host vehicle. Based on the collision risk, the module determines traffic participants that are considered to be at a necessary risk. If no traffic participants are considered to be at a necessary risk, the module concludes the assessment.

[0060] It can be understood that other traffic participants are all traffic participants except the host vehicle.

[0061] The perception information fusion module includes a perception information feedback compensation module and a perception fusion module;

[0062] The perception information feedback compensation module is used to perform spatiotemporal fusion of information on traffic participants with necessary risks to obtain spatiotemporal fusion information. The spatiotemporal fusion information is transmitted back to the host vehicle through the roadside perception module, thereby obtaining information on traffic participants that may collide with it.

[0063] The perception fusion module is used to fuse spatiotemporal fusion information and compensate for the information of traffic participants that the main vehicle has not identified in the main vehicle's blind spot, thereby realizing the exchange and sharing of traffic environment data information, thereby providing the main vehicle with more accurate perception environment information.

[0064] Optionally, the roadside perception module includes roadside sensors and edge servers; the roadside sensors are used to detect all roadside perception data within range; the edge servers are equipped with a roadside sensor detection and calculation module, which is used to obtain roadside perception data from the roadside sensors, identify traffic participant information, and calculate traffic participant information. Traffic participant information includes information such as the location, speed, heading angle, category, and location coordinates of traffic participants.

[0065] Preferably, the edge server is equipped with a point cloud detection and calculation module; the roadside sensor identifies global information from a bird's-eye view and obtains all roadside perception data within the detection range; traffic participants include motor vehicles, non-motor vehicles and pedestrians within the visible range.

[0066] Optionally, the blind spot type identification module first determines whether the main vehicle's blind spot exists and the type of blind spot that exists. Based on the type of the main vehicle's blind spot obtained and combined with the roadside perception data of the roadside sensor, different types of blind spots are modeled to divide the area range of the main vehicle's blind spot; among which, the blind spot categories include vehicle obstruction blind spots and building obstruction blind spots.

[0067] Optionally, the risk assessment module obtains predicted trajectories of each traffic participant and the host vehicle within the blind spot of the host vehicle based on traffic participant information obtained by roadside sensors, and assesses a collision risk assessment for the host vehicle's driving safety based on the intersection of the predicted trajectories of each traffic participant and the predicted trajectory of the host vehicle. Furthermore, a collision threshold for the collision risk assessment is set. If the collision risk assessment is higher than the collision threshold, the corresponding traffic participant is considered to be at necessary risk.

[0068] Optionally, spatiotemporal fusion includes coordinate system conversion and timestamp matching; information about traffic participants who are necessary risks in the blind spot of the main vehicle is transmitted back to the main vehicle through the communication node of the edge server.

[0069] Optionally, the roadside sensor is a radar;

[0070] Optionally, the intersection size is the number of intersection points between the predicted trajectory of each traffic participant and the predicted trajectory of the host vehicle.

[0071] For example, Figure 2 As shown in the figure, the blind spot type recognition model classifies the blind spot of the main vehicle into two types of occlusion: building occlusion blind spot Figure 2 (a) and the blind spot blocked by vehicles as shown in Figure 2 As shown in (b), two different blind areas are modeled based on these two types, and the occlusion blind area is extracted and divided.

[0072] Another specific embodiment of the present invention is as follows Figure 3As shown, a vehicle blind spot cooperative perception method based on roadside sensors is disclosed, using the aforementioned vehicle blind spot cooperative perception system, and the specific steps are as follows:

[0073] Step 1: Acquire roadside sensing data and obtain traffic participant information based on the roadside sensing data;

[0074] Step 2: Determine the main vehicle from all vehicles in the traffic participant information;

[0075] Step 3: Determine whether there is a blind spot for the host vehicle based on the roadside sensing data; if there is a blind spot for the host vehicle, identify the type of the blind spot and obtain the location and boundary of the blind spot, thereby obtaining the area of ​​the blind spot;

[0076] Among them, the types of blind spots of the main vehicle include vehicle obstruction blind spots and building obstruction blind spots; based on the heading angle, position coordinates and other information of the traffic participants, the corresponding obstruction blind spot boundary line is obtained;

[0077] If the blind spot type of the main vehicle is a building-blocked blind spot, obtain the building blind spot boundary line. Figure 2 (a) Obtain pre-calibrated building corner points and use the connecting line between the main vehicle center coordinates and the building corner points to divide the vehicle's sight area and blind spot boundary line. The expression of the building blind spot boundary line is as follows:

[0078]

[0079]

[0080] Among them, x0, y0 are the coordinates of the center point of the main vehicle, and x1, y1, x2, y2 are the coordinates of the two corner points of the building closest to the main vehicle.

[0081] If the blind spot type of the main vehicle is a vehicle-occluded blind spot, obtain the boundary line of the vehicle-occluded blind spot. Figure 2 (b) Model and divide the occlusion blind area of ​​vehicles within the preset radius of the main vehicle. The coordinates of the four rectangular boxes of the occluding vehicles are expressed as:

[0082]

[0083]

[0084]

[0085]

[0086] Among them, x0, y0 are the coordinates of the center point of the main vehicle, x1, y1 are the coordinates of the center point of the occluding vehicle, l, w are the length and width of the occluding vehicle respectively, and yaw is the heading angle of the occluding vehicle;

[0087] The coordinates of the center point of the main vehicle and the four endpoints of the obstructing vehicle are connected as lines. The two connecting lines with the largest angle between the main vehicle and the four endpoints are taken as the boundary lines of the vehicle blind spot. The expression is:

[0088]

[0089]

[0090] Preferably, since the short-range occlusion blind spot has a greater impact on the main vehicle and the long-range occlusion blind spot can be ignored, vehicles outside a radius of 10 meters from the main vehicle and the rear vehicles are filtered out, so the impact distance range of the blind spot occluded by other vehicles is set to 10 meters.

[0091] Step 4: Determine whether there are other traffic participants in the blind spot of the host vehicle;

[0092] If there are other traffic participants, go to step 5;

[0093] If there are no other traffic participants, go to step 10;

[0094] Step 5: Based on the identified location and area of ​​the blind spot of the host vehicle, predict the movement trajectory of traffic participants within the area of ​​the blind spot of the host vehicle;

[0095] Step 6: Based on the predicted trajectories of traffic participants within the blind spot of the host vehicle, the collision risk between the host vehicle and other traffic participants within the blind spot of the host vehicle is evaluated to obtain a risk assessment result.

[0096] Step 7: Obtain necessary risk traffic participants based on risk assessment results;

[0097] If it is necessary to risk traffic participants, proceed to step 8;

[0098] The predicted trajectory of each traffic participant and the host vehicle within the blind spot of the host vehicle is obtained, and a collision risk assessment for the driving safety of the host vehicle is evaluated based on the intersection of the predicted trajectory of each traffic participant and the predicted trajectory of the host vehicle. Furthermore, a collision threshold for the collision risk assessment is set. If the collision risk assessment is higher than the collision threshold, the corresponding traffic participant is considered a necessary risk, i.e., a necessary risk traffic participant.

[0099] If there is no need to risk traffic participants, go to step 10;

[0100] Step 8: Perform spatiotemporal fusion on the information of the traffic participants with necessary risks to obtain spatiotemporal fusion information; and transmit the spatiotemporal fusion information back to the host vehicle;

[0101] Step 9: The main vehicle integrates the spatiotemporal fusion information;

[0102] Step 10: traverse all other vehicles in the traffic participant information, select another vehicle as the main vehicle, and repeat the above steps until all vehicles are traversed.

[0103] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by any technician familiar with this technical field within the technical scope disclosed by the present invention should be covered by the scope of protection of the present invention.

Claims

1. A vehicle blind spot cooperative perception system based on roadside sensors, including a blind spot recognition module, a risk assessment module and a perception information fusion module, characterized in that: The blind spot recognition module includes a roadside perception module and a blind spot type recognition module; the roadside perception module is used to obtain roadside perception data and extract traffic participant information based on the roadside perception data; The blind spot type recognition module determines whether there is a blind spot of the main vehicle based on the roadside sensing data; if there is a blind spot of the main vehicle, it identifies the type of the blind spot of the main vehicle and obtains the location and area range of the blind spot of the main vehicle; The main vehicle is determined from all vehicles in the traffic participant information; the types of the main vehicle's blind spots include vehicle obstruction blind spots and building obstruction blind spots; The risk assessment module is used to determine traffic participants that are considered essential risks. If a blind spot exists for the host vehicle, the module predicts the trajectories of traffic participants within the blind spot based on the identified location and area of ​​the blind spot, and assesses the collision risk between the host vehicle and other traffic participants within the blind spot. Based on the collision risk, traffic participants that are considered essential risks are identified. If the collision risk assessment exceeds a collision threshold, the corresponding traffic participant is considered essential risk. The perception information fusion module includes a perception information feedback compensation module and a perception fusion module; The perception information feedback compensation module is used to perform spatiotemporal fusion on the information of traffic participants with necessary risks to obtain spatiotemporal fusion information; the spatiotemporal fusion information is transmitted back to the host vehicle through the roadside perception module; The perception fusion module is used to fuse spatiotemporal fusion information and compensate for the information of traffic participants that the host vehicle has not recognized in its blind spot.

2. The vehicle blind spot cooperative perception system according to claim 1, characterized in that: The roadside perception module includes roadside sensors and edge servers; the roadside sensors are used to detect all roadside perception data within the detection range; The edge server is equipped with a roadside sensor detection and calculation module, which is used to obtain roadside perception data from roadside sensors, identify and calculate traffic participant information to obtain traffic participant information.

3. The vehicle blind spot cooperative perception system according to claim 2, characterized in that: Traffic participant information includes traffic parameters such as location, speed, heading angle, category and location coordinates.

4. The vehicle blind spot cooperative perception system according to claim 2, characterized in that: Roadside sensors identify global information from a bird's-eye view and obtain roadside perception data within the detection range.

5. A vehicle blind spot cooperative perception method based on roadside sensors, characterized in that: The specific steps are as follows: Step 1: Acquire roadside sensing data and obtain traffic participant information based on the roadside sensing data; Step 2: Determine the main vehicle from all vehicles in the traffic participant information; Step 3: Determine whether there is a blind spot for the host vehicle based on the roadside sensing data; if there is a blind spot for the host vehicle, identify the type of the blind spot and obtain the location and boundary of the blind spot, thereby obtaining the area of ​​the blind spot; Among them, the types of blind spots of the main vehicle include vehicle obstruction blind spots and building obstruction blind spots; Step 4: Determine whether there are other traffic participants in the blind spot of the host vehicle; If there are other traffic participants, go to step 5; Step 5: Based on the identified location and area of ​​the blind spot of the host vehicle, predict the movement trajectory of traffic participants within the area of ​​the blind spot of the host vehicle; Step 6: Based on the predicted trajectories of traffic participants within the blind spot of the host vehicle, the collision risk between the host vehicle and other traffic participants within the blind spot of the host vehicle is evaluated to obtain a risk assessment result. Step 7: Obtain necessary risk traffic participants based on the risk assessment results. If the collision risk assessment is higher than the collision threshold, the corresponding traffic participant is considered necessary risk. If it is a necessary risk traffic participant, proceed to step 8. Step 8: Perform spatiotemporal fusion on the information of the traffic participants with necessary risks to obtain spatiotemporal fusion information; and transmit the spatiotemporal fusion information back to the host vehicle; Step 9: The main vehicle integrates the spatiotemporal fusion information; Step 10: Traverse all other vehicles in the traffic participant information, select another vehicle as the main vehicle, and repeat the above steps 3-9 until all vehicles are traversed.

6. The vehicle blind spot cooperative perception method according to claim 5, characterized in that: Based on the heading angle and position coordinate information of the traffic participant, the corresponding blind area boundary line is obtained.

7. The vehicle blind spot cooperative perception method according to claim 6, characterized in that: If the blind spot of the main vehicle is a building-occluded blind spot, obtain the boundary line of the building blind spot. The specific steps are as follows: Obtain the pre-calibrated building corner points, and use the connecting line between the main vehicle center coordinates and the building corner points to divide the vehicle's sight area and blind spot boundary line. The expression of the building blind spot boundary line is as follows: Among them, x0, y0 are the coordinates of the center point of the main vehicle, and x1, y1, x2, y2 are the coordinates of the two corner points of the building closest to the main vehicle.

8. The vehicle blind spot cooperative perception method according to claim 6, characterized in that: If the blind spot type of the host vehicle is a vehicle obstruction blind spot, obtain the boundary line of the vehicle obstruction blind spot. The specific steps are as follows: The blind spot of vehicles within the preset radius of the main vehicle is modeled and divided. The coordinates of the four rectangular boxes of the obstructing vehicles are expressed as follows: Among them, x0, y0 are the coordinates of the center point of the main vehicle, x1, y1 are the coordinates of the center point of the occluding vehicle, l, w are the length and width of the occluding vehicle respectively, and yaw is the heading angle of the occluding vehicle; The coordinates of the center point of the main vehicle and the four endpoints of the obstructing vehicle are connected as lines. The two connecting lines with the largest angle between the main vehicle and the four endpoints are taken as the boundary lines of the vehicle blind spot. The expression is:

Images