Sensor fusion based on an intersection scene to determine the collision potential of vehicles

Abstract

Method (400) for implementing sensor fusion to determine a collision potential for a vehicle (100), the method comprising: Identifying (410) a specific intersection (200) that the vehicle (100) is approaching, using a processor; Identify, using the processor, collision potential scenarios associated with one or more paths through the specified intersection (200), wherein each collision potential scenario defines a risk of collision between the vehicle (100) and an object in a specified area; Setting (430), using the processor, a weighting with which one or more information sources of the vehicle (100) are taken into account for each collision potential scenario, such that one or more of the one or more information sources that provide the most relevant and reliable information about the specific area corresponding to the collision potential scenario is given the highest weighting; Implement (440), using the processor, a sensor fusion based on setting the weighting of one or more information sources and performing a detection based on the sensor fusion; Providing (450), using the processor, a warning or implementing actions according to the detection, Identifying and storing the collision potential scenarios associated with the one or more paths through each intersection (200) on a map, wherein identifying the collision potential scenarios associated with the one or more paths through the particular intersection (200) includes retrieving the collision potential scenarios from a memory in real time and updating the weighting with which the one or more information sources of the vehicle (100) are considered for each upcoming collision potential scenario in real time based on real-time conditions, wherein the real-time conditions include a quality of data from the one or more information sources or ambient light conditions.

Description

[0001] The subject of the disclosure relates to sensor fusion based on an intersection scene for determining the collision potential of vehicles.

[0002] Vehicles (e.g., cars, trucks, construction equipment, agricultural machinery, automated factory systems) increasingly use sensors to gather information about the vehicle and its environment. This sensor information enables the enhancement or automation of vehicle operation. Examples of such sensors include cameras, radar (radio detection and ranging), and lidar (light detection and ranging). Sensor fusion refers to the use of information from multiple sensors. For example, sensor fusion can be used to determine the potential for a vehicle collision. However, the sensor information from not every sensor is equally relevant. Therefore, it is desirable to implement sensor fusion based on an intersection scenario to determine the collision potential of vehicles.

[0003] German patent application DE 10 2016 119 486 A1 discloses a system and a method for warning the driver of a motor vehicle of a potential collision when turning right or left at or near an intersection. The system and method provide additional analysis to limit false-positive and false-negative warnings based on specific circumstances. The method involves determining whether the carrier vehicle is likely to turn at or near the intersection and acquiring rotational speed, velocity, and position data of the carrier vehicle and any other relevant vehicles. Based on the rotational speeds, velocity, and position data of the carrier vehicle and the other vehicles, the method determines the predicted routes of the vehicles and issues a warning to the driver of the carrier vehicle if the carrier vehicle is likely to collide with one of the other vehicles based on these predicted routes.

[0004] DE 198 45 568 A1 discloses a method and a device for object detection for motor vehicles, comprising a distance sensor system formed by a plurality of distance sensors, which are arranged on the motor vehicle in such a way that they scan the environment of the motor vehicle, and an evaluation unit that determines the trajectory and speed of an object relative to the motor vehicle from the data of the distance sensor system, wherein the distance sensors can be selectively controlled by the evaluation unit and the range and / or the measurement repetition frequency and / or the resolution and / or the operating mode of the distance sensors can be changed.

[0005] DE 10 2018 115 198 A1 discloses a virtual crowdsourcing sensor generator that can be provided by or for a vehicle, for example as a service.The virtual crowdsourcing sensor generator may include, but is not limited to, a communication system configured to receive vehicle sensor data from one or more contributing vehicles, a location of one or more contributing vehicles and a target vehicle, and a processor configured to filter the received contributing vehicle sensor data based on the location of the one or more contributing vehicles and the location of the target vehicle, then aggregate the filtered vehicle sensor data into at least one data-specific dataset and one application-specific dataset, and generate a virtual sensor for the target vehicle, wherein the virtual sensor processes the filtered and aggregated sensor data of the contributing vehicle to generate output data regarding the location of the target vehicle.

[0006] In an exemplary embodiment, a method for implementing sensor fusion to determine a collision potential for a vehicle comprises identifying a specific intersection toward which the vehicle is approaching and identifying collision potential scenarios associated with one or more paths through that specific intersection. Each collision potential scenario defines a risk of collision between the vehicle and an object in a specific area. The method also includes setting a weighting system to consider one or more of the vehicle's information sources for each collision potential scenario, such that one or more of the information sources providing the most relevant and reliable information about the specific area corresponding to the collision potential scenario are given the highest weighting.Sensor fusion is implemented based on setting the weighting of one or more information sources and performing detection based on the sensor fusion; according to the detection, a warning is issued or actions are performed.

[0007] In addition to one or more of the features described here, identifying the specific intersection the vehicle is approaching includes obtaining the vehicle's location and referencing the vehicle's location on a map that identifies a multitude of intersections.

[0008] In addition to one or more of the features described herein, the procedure also includes identifying the one or more of the information sources that provide the most relevant and reliable information about the specific area corresponding to the collision potential scenario, according to real-time conditions, wherein the real-time conditions include ambient light intensity and obscured views from any of the vehicle's information sources.

[0009] In addition to one or more of the features described here, the procedure also includes considering only the collision potential scenarios that are associated with a particular one of the one or more paths through the particular intersection by identifying an intention of the driver to cross the particular one of the one or more paths, based on the fact that the vehicle is a driver-controlled vehicle.

[0010] In addition to one or more of the features described here, identifying the driver's intent includes receiving a button or indicator input from the driver or obtaining a location of the vehicle relative to route information provided to the driver.

[0011] The procedure includes identifying and storing the collision potential scenarios associated with the one or more paths through each intersection on a map, wherein identifying the collision potential scenarios associated with the one or more paths through the particular intersection involves retrieving the collision potential scenarios from memory in real time.

[0012] The procedure also includes updating the weighting, which takes into account the one or more information sources of the vehicle for each impending collision potential scenario in real time based on the real-time conditions.

[0013] Real-time conditions include the quality of data from one or more information sources or ambient light conditions.

[0014] In addition to one or more of the features described here, the information sources include sensors, and the sensors include a camera, radar system or lidar system, or the information sources include communication sources, and the communication sources include another vehicle, infrastructure or a cloud-based server.

[0015] In addition to one or more of the features described here, providing the warning includes issuing a visual, audible or haptic warning, and performing the actions includes performing automatic braking.

[0016] In another exemplary embodiment, a system for implementing sensor fusion to determine a vehicle's collision potential comprises vehicle information sources and a processor for identifying a specific intersection the vehicle is approaching and for identifying collision potential scenarios associated with one or more paths through that specific intersection. Each collision potential scenario defines a risk of collision between the vehicle and an object in a specific area. The processor also sets a weighting system to consider one or more of the vehicle's information sources for each collision potential scenario, such that one or more of the information sources providing the most relevant and reliable information about the specific area corresponding to the collision potential scenario receive the highest weighting.Sensor fusion is implemented based on the weighting of one or more information sources. Based on this sensor fusion, a detection is performed, and depending on the detection, an alert is issued or actions are taken.

[0017] In addition to one or more of the features described here, the processor identifies the specific intersection the vehicle is approaching by obtaining a location of the vehicle and referencing the vehicle's location on a map that identifies a multitude of intersections.

[0018] In addition to one or more of the features described here, the processor identifies the one or more information sources that provide the most relevant and reliable information about the specific area corresponding to the collision potential scenario, according to the real-time conditions. The real-time conditions include ambient light intensity and obstructed views of one of the vehicle's information sources.

[0019] In addition to one or more of the features described here, the processor only considers the collision potential scenarios associated with a particular one or more paths through the particular intersection by identifying an intention of the driver to cross the particular one or more paths, based on the fact that the vehicle is a driver-controlled vehicle.

[0020] In addition to one or more of the functions described here, the processor identifies the driver's intention by receiving a button or turn signal input from the driver or by obtaining a location of the vehicle relative to route information provided to the driver.

[0021] The processor identifies and stores the collision potential scenarios associated with the one or more paths through each intersection on a map, and the processor identifies the collision potential scenarios associated with the one or more paths through the specific intersection by retrieving the collision potential scenarios from memory in real time.

[0022] The processor updates the weighting with which the vehicle's one or more information sources are considered for each impending collision potential scenario in real time, based on real-time conditions.

[0023] Real-time conditions include the quality of data from one or more information sources or ambient light conditions.

[0024] In addition to one or more of the features described here, the information sources include sensors, and the sensors include a camera, radar system or lidar system, and the information sources include communication sources, and the communication sources include another vehicle, infrastructure or a cloud-based server.

[0025] In addition to one or more of the features described here, the warning is a visual, audible or haptic alert, and the actions include performing automatic braking.

[0026] Further features, advantages and details are listed only as examples in the following detailed description, which refers to the drawings in which: Fig. 1 a block diagram of a vehicle that performs sensor fusion based on an intersection scene to determine a vehicle's collision potential, according to one or more embodiments; Fig. 2 a sensor fusion based on an intersection scene for determining a vehicle collision potential, according to one or more embodiments; Fig. 3. A sensor fusion based on an intersection scene for determining a vehicle collision potential, according to one or more embodiments; and Fig. 4 a process flow of a method for performing sensor fusion based on an intersection scene to determine a vehicle collision potential, according to one or more embodiments.

[0027] As mentioned earlier, vehicle sensors can provide information that facilitates the detection of a potential collision. The information provided by vehicle sensors is beneficial in driver-controlled and semi-autonomous vehicles and essential for autonomous vehicle operation. Sensor fusion combines information from multiple sensors. For example, false alarms can be avoided by requiring that all or a majority of sensors detect an object or collision potential before such a detection is accepted. All sensors can be weighted equally according to a conventional sensor fusion algorithm. However, in a given intersection scenario, not every sensor may be equally relevant or reliable.

[0028] Implementations of the systems and methods described here involve sensor fusion based on an intersection scene to determine the collision potential of vehicles. An intersection scene refers to a specific intersection and its real-time conditions (e.g., ambient light, obstructed view). Sensor information is weighted based on the relevance and reliability of the sensor for the respective intersection scene. A given intersection scene and a given vehicle path through the intersection may include more than one collision potential scenario (e.g., road ahead, cross traffic). Each scenario may require a different weighting of the sensors involved in the sensor fusion. As described in detail, sensor fusion involves not only the fusion of information from the vehicle's sensors but also information from other sources.

[0029] According to an exemplary embodiment, Fig. 1 a block diagram of a vehicle 100 that performs sensor fusion based on an intersection scene 210 ( Fig. 2), 310 ( Fig. 3) performs to determine the collision potential of a vehicle. As mentioned earlier, intersection scene 210, 310 refers to intersection 200 ( Fig. 2, Fig. 3) and conditions (e.g., lighting, obstructed view) that affect the relevance and effectiveness of a particular sensor of the vehicle 100. The in Fig. The example vehicle 100 shown is an automobile 101 and can be driver-controlled or autonomous. The vehicle 100 contains a controller 110 that receives information from sensors such as an ambient light sensor 105, a lidar system 120, cameras 130, and a radar system 140 and performs sensor fusion. The exemplary number and arrangement of the sensors in Fig. Paragraph 1 is not intended to restrict alternative embodiments. For example, the vehicle 100 may include several radar systems 140 (e.g., one on each side of the vehicle 100 and one in the middle at the front and rear) or additionally or alternatively several lidar systems 120.

[0030] The vehicle 100 is represented with a global positioning system (GPS) 150, which provides the position of the vehicle 100 and, in conjunction with map information, can enable the controller 110 to determine upcoming intersections 200 ( Fig. 2) The controller 110 can use the GPS 150 and the map to provide the driver, for example, in a driver-controlled vehicle 100, with route information. The controller 110 can also communicate with other vehicles 160 via vehicle-to-vehicle (V2V) communication, with a cloud server 170 via vehicle-to-everything (V2X) communication, or with an infrastructure 180 via vehicle-to-infrastructure (V2I) communication. In addition to gathering information, the controller 110 can also transmit information to the driver (in the driver-controlled case) or operator (in the autonomous case) of the vehicle 100 via an infotainment system 115 or another interface. The controller 110 can issue a warning or communicate with vehicle systems to perform automatic actions based on the vehicle's collision potential assessment.

[0031] The controller 110 comprises processing circuits that may include an application-specific integrated circuit (ASIC), an electronic circuit, a processor (shared, dedicated, or grouped), memory that executes one or more software or firmware programs, a combinational logic circuit, and / or other suitable components that provide the described functionality. As described with reference to Fig. As explained in section 2, sensor fusion based on an intersection scene 210, 310 for determining the vehicle collision potential refers to a consideration of which information sources are most relevant according to the respective intersection 200 and the real-time conditions in order to weight the information as part of the sensor fusion.

[0032] Fig. Figure 2 shows a sensor fusion based on an intersection scene 210 for determining the collision potential of vehicles according to one or more embodiments. The intersection scene 210 is used to discuss aspects of an exemplary embodiment, and the in Fig. The intersection scene shown in Figure 310 is used to discuss additional aspects of the exemplary embodiment, all of which are described in Figure 310. Fig. 4 are summarized. The in Fig. The intersection scene 210 depicted refers to intersection 200 and the real-time conditions (e.g., lighting, traffic flow) at intersection 200. Vehicle 100 is traveling straight ahead, while vehicle 220 is turning in lane 205a. Vehicle 220 will cross the path of vehicle 100 when it turns left. Another vehicle, 230, is stopped illegally in lane 205b, which is adjacent to lane 205c of vehicle 100, to allow vehicle 220 to turn left. The presence of the stopped vehicle 230 creates a blind spot for vehicle 100's sensors with respect to vehicle 220.

[0033] Intersection 200 and the path of vehicle 100 through intersection 200 imply more than one possible collision scenario. One possible collision scenario involves turning lane 205a, another lane 205d. That is, a given collision potential scenario implies an area from which an object could emerge and potentially collide with vehicle 100. The sensors with a field of view covering this area are the most relevant sensors and are therefore likely to be weighted most heavily in the sensor fusion algorithm.

[0034] With regard to turning lane 205a, the most relevant sensor may be camera 130, located in the front center of vehicle 100. This sensor has a field of view that includes turning lane 205a and can therefore be given the highest weighting. Since part of turning lane 205a is obscured by the stopped vehicle 230, V2V or V2X communication that can indicate the presence of the turning vehicle 220 can be given a high weighting (e.g., the same weight as camera 130 in the front center of vehicle 100). A rule-based assignment can be used so that the controller 110 assigns weights to the various information sources according to the intersection scene 210.

[0035] Regarding lane 205d, a different weighting than the one described above for turning lane 205a can be used for sensor fusion. Alternatively, the camera 130 at the front center of the vehicle 100 can receive the highest weighting, as with turning lane 205a, due to its field of view. Additionally, the GPS 150 and the map, which together display the location of turning lane 205d relative to the vehicle 100, can be weighted equally to the camera 130 at the center right of the vehicle 100. If the ambient light sensor 105 is used, the detected ambient light intensity can influence the weighting assigned to the various sensors. For example, if the real-time ambient light intensity is below a threshold corresponding to the minimum light intensity required for the cameras 130, it would not be beneficial to assign additional weight to the camera 130 at the front center of the vehicle 100.Despite its position, camera 130 would not provide any relevant information due to the ambient light conditions. In this case, only GPS 150 and map information can be given greater weight than other sensors when detecting the upcoming lane 205d.

[0036] Fig. Figure 3 shows a sensor fusion based on an intersection scene 310 to determine the collision potential of vehicles according to one or more embodiments. The in Fig. The intersection 200 shown and other intersections 200 on a map available to vehicle 100 can be pre-assessed for collision potential scenarios for different paths through the intersections 200. If vehicle 100 is therefore the one in Fig. As the vehicle approaches the intersection 200 shown in section 3, the control unit 110 determines which information from which source should be weighted more heavily for each collision potential scenario.

[0037] Vehicle 100 can take one of several paths through intersection 200, as depicted in intersection scene 310. For example, vehicle 100 can turn right into lane 305a or lane 305b. In other examples, vehicle 100 can turn left into lane 305c or lane 305d. For each of these paths, a collision potential scenario involves lane 305a to the left of vehicle 100, assuming a keep-right rule. Therefore, a camera 130 or radar system 140 on the left front of vehicle 100 can be given the highest weight according to a rule-based approach for this collision potential scenario. Intersection scene 310 contains a bush 340 that partially obstructs the view of the left side of vehicle 100.Therefore, the rule can be modified to give greater weight to V2V or V2X information about vehicles traveling in lane 305a towards intersection 200. For a left-turning vehicle into lane 305c or 305d, another collision potential scenario concerns the lanes to the right of vehicle 100, where vehicles 320 and 325 are visible. For this collision potential scenario, a camera or radar system 140 located on the right front of vehicle 100 can be given the greatest significance.

[0038] Area 335 to the right of vehicle 100 and vehicle 330 within area 335 are not relevant (i.e., not part of a collision potential scenario) if vehicle 100 turns left or right into lane 305b. However, if vehicle 100 turns right into lane 305a, vehicle 330 or a slow-moving or stationary object within area 335 becomes part of a relevant collision potential scenario. Therefore, when vehicle 100 begins turning into lane 305a, the weighting of information from a camera 130 or radar system 140 located at the front left of vehicle 100 can be increased, and once the turning maneuver is complete, the increased weighting can be shifted to information from a camera 130 or radar system 140 located at the front center of vehicle 100.

[0039] If the vehicle 100 is driver-operated and the driver's intention (i.e., the intended path of the vehicle 100 through the intersection 200) can be determined, the controller 110 can anticipate how the weighting of the sensor fusion needs to be changed in real time. For example, the driver of the vehicle 100 can indicate the path to be taken through an intersection 200 via a button or with a turn signal. As another example, the controller 110 can be aware of an upcoming maneuver because the driver is using a route navigation system based on GPS 150 and a map. Based on the knowledge of the driver's intention, the controller 110 knows the path and thus the relevant collision potential scenarios through a particular intersection 200.The control unit 110 can adjust the weighting of information from various sensors and other sources for each potential collision scenario at intersection 200 in anticipation of encountering that scenario. Without information about the driver's intention, the control unit 110 adjusts the weighting for each potential collision scenario encountered in real time. If the vehicle 100 is an autonomous vehicle, its planned routes are known to the control unit 110. Therefore, the aforementioned intention determination is irrelevant, and only the collision scenarios along the planned route for a given intersection 200 need to be considered.

[0040] Fig. 4 is a process flow of a method 400 for performing sensor fusion based on an intersection scene to determine a vehicle collision potential according to one or more embodiments. In the discussion of Fig. 4 will continue to be applied to the Fig. 1, Fig. 2 to Fig. 3. In Block 410, the identification of vehicle paths with collision potential for each intersection 200 can be performed in advance, as previously mentioned. Thus, each collision potential scenario for each path through each intersection 200 can be stored by the controller 110 to more quickly identify collision potential scenarios for an approaching intersection 200 in real time. In Block 420, the determination of an upcoming intersection 200 can involve the use of GPS 150 and the map. Optionally, the processes in Block 420 can include the determination of the driver's intent in real time for a driver-controlled vehicle 100. The driver's intent refers to the specific path through the intersection 200 that the vehicle 100 will cross, as previously mentioned.Determining the driver's intention to take a specific route through intersection 200 in real time means that only the collision potential scenarios associated with that route are considered in real time.

[0041] In Block 430, the processes involve assigning a weight to each information source according to each collision potential scenario at intersection 200 and other available conditions. If a vehicle 100 is at a T-stop, as in Fig. As shown in Figure 3, collision potential scenarios can, for example, affect the left and right sides of vehicle 100, as with reference to Fig. 3 explained. Furthermore, other available conditions (e.g., obstacles such as the bush 340, ambient light intensity) can influence which information sources should be trusted most. A rule-based approach can be used to assign a weight to each information source based on all available information, or, according to alternative embodiments, to increase the weight of one or more information sources in block 430. The information sources could be one of the sensors (e.g., camera 130, lidar system 120, radar system 140) or a communication (e.g., V2V communication, V2X communication).

[0042] Assigning a weight to each information source (or increasing the weight of one or more information sources) in Block 430 may additionally involve assigning different weights to data (or increasing the weight of specific data) collected from a particular information source during processing. For example, as with reference to Fig. As discussed in section 2, the central front camera 130 of vehicle 100 is assigned the highest weight for the collision potential scenario in connection with turning lane 205a. Furthermore, the image data from the part of the central front camera 130's field of view that corresponds to turning lane 205a can be given a higher weight than other image data when processing the image data received from the central front camera 130.

[0043] In Block 440, the adjustment of the initial weight assigned in Block 430, based on sensor data quality and other available factors, determines whether the rule-based result needs to be adjusted based on real-time conditions. For example, if camera 130 on the left front of vehicle 100 is assigned the highest weight by controller 110 based on a rule (in Block 430), but snow obscures camera 130's field of view, the weight can be adjusted. Another example: If GPS 150 and the map were assigned the highest weight, but the map information is inaccurate due to construction or other factors, the weight can be adjusted in Block 440.

[0044] In Block 450, the processes include determining a collision potential based on sensor fusion, using the weights assigned in Blocks 430 and 440 for each collision potential scenario. The processes in Block 450 also include providing a warning or performing an action (alternatively or additionally). A warning can be issued when vehicle 100 approaches intersection 200 and, alternatively or additionally, at intersection 200. The warning can be visual, audible, haptic, or in more than one form. Alternatively or additionally to warnings, autonomous actions (e.g., automatic braking) can be performed based on the collision potential determined by sensor fusion.The increased weighting of a particular sensor or other information source in sensor fusion means that a failure to detect a collision hazard by other sensors or information sources can be overridden if it is detected by the sensor or information source with the highest weight.

Claims

[1] Method (400) for implementing sensor fusion to determine a collision potential for a vehicle (100), the method comprising: Identifying (410) a specific intersection (200) that the vehicle (100) is approaching, using a processor; Identify, using the processor, collision potential scenarios associated with one or more paths through the specified intersection (200), wherein each collision potential scenario defines a risk of collision between the vehicle (100) and an object in a specified area; Setting (430), using the processor, a weighting with which one or more information sources of the vehicle (100) are taken into account for each collision potential scenario, such that one or more of the one or more information sources that provide the most relevant and reliable information about the specific area corresponding to the collision potential scenario is given the highest weighting; Implement (440), using the processor, a sensor fusion based on setting the weighting of one or more information sources and performing a detection based on the sensor fusion; Providing (450), using the processor, a warning or implementing actions according to the detection, Identifying and storing the collision potential scenarios associated with the one or more paths through each intersection (200) on a map, wherein identifying the collision potential scenarios associated with the one or more paths through the particular intersection (200) includes retrieving the collision potential scenarios from a memory in real time and updating the weighting with which the one or more information sources of the vehicle (100) are considered for each upcoming collision potential scenario in real time based on real-time conditions, wherein the real-time conditions include a quality of data from the one or more information sources or ambient light conditions. [2] Method (400) according to claim 1, wherein identifying (410) the specific intersection (200) to which the vehicle (100) is approaching, obtaining a location of the vehicle (100) and referencing the location of the vehicle (100) on a map that identifies a plurality of intersections (200). [3] Method (400) according to claim 1, further comprising identifying one or more of the one or more information sources that provide the most relevant and reliable information about the specific area corresponding to the collision potential scenario according to real-time conditions, wherein the real-time conditions include ambient light intensity and obstructed views from any of the vehicle's (100) information sources, wherein the information sources include sensors and the sensors include a camera, a radar system or a lidar system, or the information sources include communication sources and the communication sources include another vehicle, infrastructure or a cloud-based server. [4] Method (400) according to claim 1, further comprising considering only the collision scenarios that are associated with a particular one of the one or more paths through the particular intersection (200) by identifying (420) an intention of the driver to cross the particular one of the one or more paths, based on the fact that the vehicle (100) is a driver-operated vehicle, wherein the identification (420) of the intention of the driver comprises receiving a button or indicator input from the driver or receiving a location of the vehicle (100) relative to route information provided to the driver. [5] System for implementing sensor fusion to determine a collision potential for a vehicle (100), wherein the system comprises: Vehicle information sources (100); and a processor trained to identify (410) a specific intersection (200) that the vehicle (100) is approaching in order to identify collision potential scenarios that could result in one or several paths through the specified intersection (200) are assigned, each collision potential scenario defining a risk of collision between the vehicle (100) and an object in a specified area, setting a weight (430) by which one or more of the vehicle's (100) information sources are taken into account for each collision potential scenario, such that one or more of the information sources providing the most relevant and reliable information about the specified area corresponding to the collision potential scenario is given the highest weight, implementing sensor fusion based on the setting of the weight of the one or more of the information sources (440), performing detection based on the sensor fusion and providing a warning or implementing actions according to the detection (450), wherein the processor is further configured to identify on a map the collision potential scenarios that are assigned to the one or more paths through each intersection (200) and to store, and wherein the processor is trained to identify the collision potential scenarios associated with the one or more paths through the particular intersection (200) by retrieving the collision potential scenarios in real time from a memory, and the processor is further configured to determine the weighting with which one or the other Several information sources of the vehicle (100) are taken into account for each impending collision potential scenario, updated in real time based on real-time conditions, wherein the real-time conditions include a quality of data from the one or more information sources or the ambient light conditions. [6] System according to claim 5, wherein the processor is configured to identify (410) the specific intersection (200) to which the vehicle (100) is approaching by obtaining a location of the vehicle (100) and referencing the location of the vehicle (100) on a map that identifies a plurality of intersections (200). [7] System according to claim 5, wherein the processor is further configured to identify the one or more of the one or more information sources that provide the most relevant and reliable information about the specific area corresponding to the collision potential scenario according to real-time conditions, wherein the real-time conditions include ambient light intensity and obstructed views from any of the information sources of the vehicle (100), and the information sources include sensors, and the sensors include a camera, a radar system or a lidar system, and the information sources include communication sources, and the communication sources include another vehicle, infrastructure or a cloud-based server. [8] System according to claim 5, wherein the processor is configured to consider only the collision potential scenarios associated with a particular path of the one or more paths through the particular intersection (200) by identifying an intention of the driver (420) to cross the particular path of the one or more paths, based on the fact that the vehicle (100) is a driver-operated vehicle, and wherein the processor is configured to identify the intention of the driver (420) by receiving a key or indicator input from the driver or by receiving a location of the vehicle (100) relative to route information provided to the driver.

Images