Communication System for Determining Vehicle Situations and Intentions Based on Cooperative Infrastructure-Aware Messages

Abstract

A communication system for determining the situation and intent of a particular remote vehicle located in the surrounding environment of a host vehicle includes one or more controllers for receiving sensed perception data including the sensed perception data. The one or more controllers execute instructions to determine a plurality of vehicle parameters associated with the particular remote vehicle. The one or more controllers execute instructions to associate the particular remote vehicle with a particular driving lane of a road based on map data. The one or more controllers determine possible maneuvers, possible exit lanes, and speed limits for the particular remote vehicle for the particular driving lane based on the map data, and determine the situation and intent of the particular remote vehicle based on the plurality of vehicle parameters, the possible maneuvers, the possible exit lanes of the particular remote vehicle, and the speed limits associated with the particular remote vehicle.

Description

Technical Field

[0001] The present disclosure relates to a communication system and method for determining a situation and an intention of a vehicle based on cooperative infrastructure sensed messages. Background Art

[0002] Cooperative sensor sharing involves wirelessly transmitting data collected by various sensors to neighboring host users or vehicles. Thus, a host vehicle can receive information about sensed objects from multiple neighboring users. In cooperative sensor sharing, remote vehicles and road infrastructure share data related to the sensed objects with the host vehicle. For example, an infrastructure camera such as a red light or a speed camera can capture data related to a remote vehicle, which is then transmitted to the host vehicle.

[0003] Vehicle-to-everything (V2X) is an umbrella term for vehicle connectivity communications and includes both vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2X) applications involving broadcasting messages from one entity to a host vehicle. However, if a particular vehicle is not equipped with V2X technology, based on cooperative sensor sharing from sources such as infrastructure cameras, the host vehicle only receives data related to the position, speed, position geometry, and forward direction of the particular vehicle. That is, in other words, the host vehicle does not receive information related to the situation of the particular vehicle (which refers to the short history of the vehicle's path) and the intention of the particular vehicle (which refers to the short prediction of the vehicle's intended path).

[0004] Therefore, while current vehicle connectivity communications achieve their intended purposes, there is still a need in the art for a method of determining the situation and intention of a vehicle when the situation and intention are not available. Summary of the Invention

[0005] A communication system for determining the situation and intention of a specific remote vehicle located in the surrounding environment of a host vehicle is disclosed in several aspects. The communication system includes one or more controllers for receiving sensed perception data related to the specific remote vehicle. The one or more controllers execute instructions to determine a plurality of vehicle parameters related to the specific remote vehicle based on the sensed perception data. The one or more controllers associate the specific remote vehicle with a specific driving lane of a road based on map data, where the map data indicates information related to the driving lane of the road along which the specific remote vehicle is traveling. The one or more controllers determine possible maneuvers, possible exit lanes, and speed limits of the specific remote vehicle for the specific driving lane based on the map data. Finally, the one or more controllers determine the situation and intention of the specific remote vehicle based on the plurality of vehicle parameters, possible maneuvers, possible exit lanes of the specific remote vehicle, and speed limits related to the specific remote vehicle.

[0006] In one aspect, a plurality of coordinate pairs based on a world coordinate system are converted into image frame coordinates for noise modeling based on a homography matrix, where the coordinate pairs represent a monitored area of the surrounding environment of the host vehicle.

[0007] In another aspect, the one or more controllers execute instructions for a Kalman filter to determine a plurality of error recovery vehicle parameters related to the specific remote vehicle based on the noise associated with the conversion noise, which is associated with converting the coordinate pairs based on the world coordinate system into image frame coordinates.

[0008] In yet another aspect, the one or more controllers execute instructions to divide an image representing the monitored area of the surrounding environment into a plurality of pixel bins.

[0009] In one aspect, the one or more controllers determine how many coordinate pairs based on the world coordinate system map to each pixel bin of the image and determine a distance covariance map and a speed covariance map based on each pixel bin that is part of the image.

[0010] In another aspect, the one or more controllers execute instructions to reproduce image data representing the specific remote vehicle and execute an object detection algorithm to detect the specific remote vehicle in the image data, where the detected specific remote vehicle is the detected object pixels. The one or more controllers match the detected object pixels with the speed covariance map and the distance covariance map.

[0011] In yet another aspect, one or more controllers execute instructions to determine noise associated with a bounding box based on multiple still images of a particular remote vehicle, and to determine pixel bins affected by the noise associated with the bounding box. The one or more controllers calculate an average velocity covariance matrix and an average distance covariance matrix for each affected pixel bin, and match pixels belonging to the detected object with a velocity covariance map and a distance covariance map. Finally, the one or more controllers send the world coordinates of the detected object and the matched velocity covariance and the matched distance covariance to a Kalman filter-based state tracking module.

[0012] In yet another aspect, one or more controllers execute instructions to determine when a particular remote vehicle is in a small lane. In response to determining that the particular remote vehicle is in a small lane, the one or more controllers set the situation to be equal to the distance traveled by the particular remote vehicle in the small lane plus the distance traveled in an adjacent lane. In response to determining that the particular remote vehicle is not in a small lane, the one or more controllers set the situation to be equal to the length of the current travel lane.

[0013] In one aspect, one or more controllers execute instructions to determine the type of travel allowed in the current travel lane of a particular remote vehicle, where the type of travel includes only straight through and allowing turns. In response to determining that the type of travel allowed in the current travel lane is only straight through, the one or more controllers set the intention to a connecting exit lane having a length represented as an intended distance.

[0014] In another aspect, one or more controllers execute instructions to determine the type of travel allowed in the current travel lane of a particular remote vehicle, where the type of travel includes only straight through and allowing turns. In response to determining that the current travel lane of the particular remote vehicle allows turns, the one or more controllers set multiple values for the intention, where each value corresponds to the length of a potential connecting exit lane.

[0015] In yet another aspect, one or more controllers execute instructions to determine a confidence level indicating the probability that an intention is accurate.

[0016] In yet another aspect, multiple vehicle parameters indicate the position, velocity, position geometry, and forward direction of a particular remote vehicle.

[0017] In one aspect, a method for determining the situation and intent of a specific remote vehicle located in the surrounding environment of a host vehicle is disclosed. The method includes receiving, by one or more controllers, sensed perception data related to the specific remote vehicle. The method includes determining, by one or more controllers, a plurality of vehicle parameters related to the specific remote vehicle based on cooperative infrastructure sensing messages. The method further includes associating the specific remote vehicle with a specific driving lane of a road based on map data, wherein the map data indicates information related to the driving lanes of the road along which the specific remote vehicle is traveling. The method further includes determining, based on the map data, possible maneuver actions, possible exit lanes, and speed limits for the specific remote vehicle for the specific driving lane. Finally, the method includes determining the situation and intent of the specific remote vehicle based on the plurality of vehicle parameters, possible maneuver actions, possible exit lanes of the specific remote vehicle, and speed limits related to the specific remote vehicle.

[0018] In another aspect, the method includes converting a plurality of coordinate pairs based on a world coordinate system into image frame coordinates for noise modeling based on a homography matrix, wherein the coordinate pairs represent a monitored area of the surrounding environment of the host vehicle.

[0019] In yet another aspect, the method includes determining, by a Kalman filter, a plurality of error recovery vehicle parameters related to the specific remote vehicle based on noise associated with the conversion noise, the conversion noise being associated with converting coordinate pairs based on the world coordinate system into image frame coordinates.

[0020] In yet another aspect, the method includes dividing an image representing the monitored area of the surrounding environment into a plurality of pixel bins, determining how many coordinate pairs based on the world coordinate system map to each pixel bin of the image, and determining a distance covariance map and a speed covariance map based on each pixel bin that is part of the image.

[0021] In one aspect, the method includes reproducing image data representing the specific remote vehicle, performing an object detection algorithm to detect the specific remote vehicle in the image data, wherein the detected specific remote vehicle is the detected object pixels, and matching the detected object pixels with the speed covariance map and the distance covariance map.

[0022] In another aspect, the method includes determining noise associated with a bounding box based on a plurality of still images of the specific remote vehicle, determining pixel bins affected by the noise associated with the bounding box, calculating an average speed covariance matrix and an average distance covariance matrix for each affected pixel bin, matching pixels belonging to the detected object with the speed covariance map and the distance covariance map, and sending the world coordinates of the detected object and the matched speed covariance and matched distance covariance to a Kalman filter-based state tracking module.

[0023] In yet another aspect, the method includes determining when a particular remote vehicle is in a small lane. In response to determining that the particular remote vehicle is in the small lane, the method includes setting the situation to be equal to the distance traveled by the particular remote vehicle in the small lane plus the distance traveled in an adjacent lane. In response to determining that the particular remote vehicle is not in the small lane, the method includes setting the situation to be equal to the length of the current driving lane.

[0024] In another aspect, the method includes determining the type of driving permitted in the current driving lane of a particular remote vehicle, where the type of driving includes only going straight through and permitted turns. In response to determining that the type of driving permitted in the current driving lane is only going straight through, the method includes setting the intention to be a connecting exit lane having a length represented as an intention distance.

[0025] In yet another aspect, the method includes determining the type of driving permitted in the current driving lane of a particular remote vehicle, where the type of driving includes only going straight through and permitted turns. In response to determining that turning is permitted in the current driving lane of the particular remote vehicle, the method includes setting multiple values for the intention, where each value corresponds to the length of a potential connecting exit lane.

[0026] Based on the description provided herein, additional application areas will become apparent. It should be understood that the description and specific examples are for illustrative purposes only and are not intended to limit the scope of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The drawings described herein are for illustrative purposes only and are not intended to limit the scope of the present disclosure in any way.

[0028] Figure 1 is a schematic diagram of a host vehicle including a communication system for determining the situation and intention of a particular remote vehicle according to an exemplary embodiment;

[0029] Figure 2 is a schematic diagram showing an exemplary environment in which Figure 1 the host vehicle shown in receives a cooperative infrastructure sensing message related to a particular remote vehicle;

[0030] Figure 3 is a block diagram of a controller that is part of the communication system shown in Figure 1 ;

[0031] Figure 4 is an exemplary illustration of image data reproducing a representation of a particular remote vehicle that is part of the surrounding environment shown in Figure 2 ;

[0032] Figure 5Ais a process flow diagram showing a method for determining the context of a specific remote vehicle according to an exemplary embodiment;

[0033] Figure 5B is a flowchart showing a method for determining the intention of a specific remote vehicle according to an exemplary embodiment; and

[0034] Figure 6 is a flowchart showing a method for determining the context and intention of a specific remote vehicle according to an exemplary embodiment. Detailed Description

[0035] The following description is merely exemplary in nature and is not intended to limit the present disclosure, application, or uses.

[0036] Reference Figure 1 , an exemplary host vehicle 10 is shown. The vehicle 10 is part of a communication system 12 that includes a controller 20 in electronic communication with a plurality of sensors 22 and one or more antennas 24. In an example as Figure 1 shown, the plurality of sensors 22 includes one or more radar sensors 30, one or more cameras 32, an Inertial Measurement Unit (IMU) 34, a Global Positioning System (GPS) 36, and a Light Detection and Ranging (LiDAR) 38. However, it should be understood that additional sensors may also be used. The communication system 12 also includes one or more remote objects 40 in the surrounding environment 26 of the host vehicle 10, as shown in Figure 2 . Referring to Figure 1 and Figure 2 both, in one embodiment, the remote objects 40 include, but are not limited to, one or more remote vehicles 42 and remote infrastructure 44 based on cooperative perception and communication systems. For example, in an embodiment as Figure 2 shown, the remote infrastructure 44 includes infrastructure cameras (such as red light cameras) and a processor and communication module (not shown). The controller 20 of the communication system 12 receives a cooperative infrastructure sensing message 46 related to a specific remote vehicle 42 based on Vehicle-to-Infrastructure (V2X). However, it should be understood that the controller 20 of the communication system 12 may alternatively receive the cooperative infrastructure sensing message 46 based on a cellular signal. In an embodiment, the cooperative infrastructure sensing message 46 related to a specific remote vehicle 42 may be determined by another vehicle (not shown) in the surrounding environment 26 rather than the remote infrastructure 44.

[0037] In an embodiment, the particular remote vehicle 42 does not include vehicle-to-vehicle (V2V) communication capabilities. Accordingly, the cooperative infrastructure sensed message 46 sent to the controller 20 of the vehicle 10 indicates only information related to the location and dynamics of the particular remote vehicle 42, rather than the context and intent of the particular remote vehicle 42. The context of the particular remote vehicle 42 indicates a driving history, and the intent predicts the expected path of the particular remote vehicle 42. As explained below, the disclosed communication system 12 determines the context and intent of the particular remote vehicle 42 based on the location and dynamics indicated by the cooperative infrastructure sensed message 46.

[0038] Figure 3 is a block diagram of the communication system 12. In the embodiment shown as Figure 2 in the controller 20 includes a tracking and detection module 50, a coordinate transformation module 52, an original position module 54, a noise modeling module 56, an object detection module 57, a positioning and map matching module 58, a context module 60, and a confidence and intent module 62. However, it should be understood that the communication system 12 can be a distributed computing system that determines the context and intent of the particular remote vehicle 42 on one or more controllers of the remote infrastructure 44 shown in Figure 2 The tracking and detection module 50 of the controller 20 receives the cooperative infrastructure sensed message 46 related to the particular remote vehicle 42 ( Figure 2 ), and the cooperative infrastructure sensed message 46 includes the sensed perception data from a perception device such as an infrastructure camera. The tracking and detection module 50 determines a plurality of vehicle parameters 68 related to the particular remote vehicle 42 based on the cooperative infrastructure sensed message 46. In an embodiment, in addition to the image data 64 collected by the remote infrastructure 44 (i.e., Figure 2 the red light camera seen in), the plurality of vehicle parameters 68 also indicate the location, detection time, size, identifier, speed, position geometry, and forward direction of the particular remote vehicle 42. However, it should be understood that other information can also be included. In one embodiment, the image data 64 is collected by a single camera, and thus the depth perception is limited. However, it should be understood that the image data 64 can also be collected from multiple cameras. The plurality of vehicle parameters 68 are then sent to the coordinate transformation module 52. The coordinate transformation module 52 converts the position represented in camera or image frame coordinates into global coordinates. The location, speed, position geometry, and forward direction are then sent to the original position module 54. The original position module 54 determines the position information of the particular remote vehicle 42, as well as the speed, acceleration, and forward direction parameters.

[0039] As explained below, the noise modeling module 56 determines the noise associated with converting coordinates from a world coordinate system (also referred to as a GPS coordinate system) to image frame coordinates. The noise modeling module 56 receives a plurality of parameters 68 and the detected pixel coordinates x, y associated with a particular remote vehicle 42, and a plurality of world coordinate pairs X, Y representing the monitored area of the surrounding environment 26 ( Figure 2 ). Specifically, the world coordinate pairs X, Y indicate the latitude and longitude of a particular point located along a stretch of road and are based on the world coordinate system. The noise modeling module 56 converts the plurality of coordinate pairs X, Y represented based on the world (e.g., GPS) coordinate system to image frame coordinates based on a homography matrix. The homography matrix is a mapping between two planes, namely, the image plane and the world coordinate plane (i.e., GPS coordinates). In an embodiment, the homography matrix is pre-computed and stored in the memory of the controller 20.

[0040] Once the image frame coordinates have been determined, the noise modeling module 56 then performs homography noise modeling by determining the noise associated with converting the world coordinate pairs X, Y to image frame coordinates. Specifically, the noise modeling module 56 then divides the image representing the monitored area of the surrounding environment 26 ( Figure 2 ) into a plurality of pixel bins. For example, the image can be divided into M×N pixel bins (such as 2×2 or 4×4). The noise modeling module 56 then determines how many world coordinates map to each pixel bin of the image. The noise modeling module 56 then determines a distance covariance map and a velocity covariance map for each pixel bin that is part of the image. The noise modeling module 56 uses the distances between the world coordinates mapped to each pixel bin to determine the distance covariance map for each pixel bin of the image. The noise modeling module 56 determines the velocity covariance map for each pixel bin by dividing the distance between the world coordinates mapped to a particular pixel bin by the inter-frame time.

[0041] Figure 4 is an exemplary illustration of the image data 64 transmitted by the cooperative infrastructure sensing message 46. The image data 64 reproduces a representation of a particular remote vehicle 42 that is part of the surrounding environment 26 ( Figure 2 ) of the host vehicle 10. Referring to Figure 3 and Figure 4 , the object detection module 57 executes an object detection algorithm to monitor the image data 64 as part of a plurality of parameters 68 to detect a particular remote vehicle 42. For example, in one non-limiting embodiment, the object detection module 57 can execute a You Only Look Once (YOLO) algorithm to detect a particular remote vehicle 42 that is part of the image data 64. Figure 4Shows a particular remote vehicle 42 detected by the object detection module 57 as the detected object pixels 70. The noise modeling module 56 then matches the detected object pixels 70 with the velocity covariance map and the distance covariance map.

[0042] Continuing to refer to Figure 3 and Figure 4 the noise modeling module 56 receives multiple still images of the particular remote vehicle 42 as input and determines the noise associated with the bounding box 72. The bounding box 72 is a rectangle that defines the detected object, which is the particular remote vehicle 42 located within the still image. The noise associated with the bounding box 72 is determined based on any available method (e.g., object detection algorithms). In an embodiment, the noise associated with the bounding box 72 is represented as a covariance matrix. The noise modeling module 56 then determines the pixel bins affected by the noise associated with the bounding box 72. The noise modeling module 56 then calculates the average velocity covariance matrix and the average distance covariance matrix for each affected pixel bin. When an object is detected, the noise modeling module 56 matches the pixels belonging to the detected object with the velocity covariance map and the distance covariance map, and the world coordinates of the detected object and the matched velocity covariance map and the matched distance covariance map are sent to the Kalman filter-based state tracking module. The Kalman filter-based state tracking module then determines the noise associated with converting the world coordinate pair X, Y into image frame coordinates. The Kalman filter then determines a plurality of error recovery vehicle parameters 78 associated with the particular remote vehicle 42 based on the noise associated with converting the world coordinate pair X, Y into image frame coordinates. In an embodiment, the plurality of error recovery vehicle parameters 78 indicate the position, velocity, and heading direction of the particular remote vehicle 42.

[0043] Returning to Figure 3 the positioning and map matching module 58 receives the plurality of error recovery vehicle parameters 78 associated with the particular remote vehicle 42 and the map data 80 from the road geometry database 82 and determines possible maneuver actions, possible exit lanes, and speed limits for the particular remote vehicle 42 based on the input. Specifically, the map data 80 most relevant to the driving direction of the particular remote vehicle 42 can be selected. The most relevant map data is based on the driving direction of the particular remote vehicle 42. For example, if the particular remote vehicle 42 is traveling from north to south, the map data 80 related to the road geometry of traveling from north to south can be used. The map data 80 indicates information related to the driving lanes of the road along which the particular remote vehicle 42 is traveling. For example, the map data 80 can indicate the number of lanes and the lane types related to the road along which the particular remote vehicle 42 is traveling. For example, the map data 80 can indicate that the road includes three lanes, where the lane types include the left lane, the middle lane, and the right lane.

[0044] The map data 80 also indicates the attributes of each lane included in the road. These attributes indicate the allowed maneuvering actions and connecting lanes. Maneuvering actions refer to the allowed driving directions, such as the allowed turns, dedicated straight-through lanes, possible connecting lanes, the starting points of turning slip lanes, and speed limits. In this example, the left lane can be a dedicated left-turn lane, the middle lane is a straight-through lane, and the right lane is a dedicated right-turn lane. A connecting lane is a lane along which a vehicle can travel after performing a maneuvering action. The positioning and map matching module 58 associates a particular remote vehicle 42 with a particular driving lane of the road based on the map data 80. The positioning and map matching module 58 then determines the possible maneuvering actions, the possible exit lanes for the particular remote vehicle 42, and the speed limit for the particular remote vehicle 42 for the particular driving lane based on the map data 80.

[0045] The positioning and map matching module 58 sends to the situation module 60 a plurality of error recovery vehicle parameters 78, the possible maneuvering actions, the possible exit lanes for the particular remote vehicle 42, and the speed limit associated with the particular remote vehicle 42. The situation module 60 then determines the situation 84 of the particular remote vehicle 42 based on the plurality of error recovery vehicle parameters 78, the possible maneuvering actions, the possible exit lanes for the particular remote vehicle 42, and the speed limit associated with the particular remote vehicle 42. The situation 84 represents the driving history of the particular remote vehicle 62 and, in an embodiment, is represented as a driving history distance.

[0046] Figure 5A is a process flow diagram showing a method 200 for determining the driving history distance or situation 84 of a particular remote vehicle 42. Specifically referring to Figure 3 and FIG. 5, the method 200 can start from decision block 202. In decision block 202, the situation module 60 checks the memory of the controller 20 to determine whether the particular remote vehicle 42 is detected for the first time and whether its associated information is saved in the memory. If the answer is yes, the method 200 proceeds to block 204, and the previous driving history distance is used as the situation 84 of the particular remote vehicle 42. The method 200 can then terminate. However, if the answer is no, the method 200 can proceed to decision block 206.

[0047] In decision block 206, the situation module 60 determines whether a particular remote vehicle 42 is in a small lane. If the particular remote vehicle 42 is in the small lane, method 200 can proceed to block 208. In block 208, in response to determining that the particular remote vehicle 42 is in the small lane, it is assumed that the particular remote vehicle 42 has changed lanes from an adjacent lane into the small lane. Method 200 can then proceed to block 210. In block 210, the situation module 60 determines that situation 84 is equal to the distance traveled by the particular remote vehicle 42 in the small lane plus the distance traveled in the adjacent lane. Method 200 can then terminate.

[0048] In the event that the situation module 60 determines that the particular remote vehicle 42 is not in the small lane, method 200 can then proceed to block 212. In block 212, the situation module 60 determines that the particular remote vehicle 42 is already in the current driving lane. Method 200 can then proceed to block 214. In block 214, the situation module 60 determines that situation 84 is equal to the length of the current driving lane. Method 200 can then terminate. Thus, the situation module 60 determines when the particular remote vehicle 42 is in the small lane, and in response to determining that the particular remote vehicle 42 is in the small lane, sets situation 84 to be equal to the distance traveled by the particular remote vehicle 42 in the small lane plus the distance traveled in the adjacent lane. However, in response to determining that the particular remote vehicle 42 is not in the small lane, the situation module 60 sets situation 84 to be equal to the length of the current driving lane. It should be understood that situation 84 can be restricted to a predetermined threshold, which can be determined based on the situation module 60.

[0049] Return reference Figure 3 , the confidence and intent module 62 receives multiple error recovery vehicle parameters 78, possible maneuvers, possible exit lanes of the particular remote vehicle 42, and speed limits associated with the particular remote vehicle 42 from the positioning and map matching module 58, and determines the confidence level 86 and intent 88 of the particular remote vehicle 42. The confidence level 86 indicates the probability that the intent 88 calculated by the confidence and intent module 62 is accurate. In an embodiment, the confidence level 86 is measured as a percentage, which can be mapped to high, medium, and low levels. As Figure 3 seen, the confidence and intent module 62 receives the cached location information 92 from the vehicle database 90, where the cached location information stores data related to the previous calculation of the confidence level.

[0050] Figure 5B is a process flow diagram showing method 300 for determining the confidence level 86 and intent 88 of a particular remote vehicle 42. Reference Figure 3 and Figure 5B, Method 300 may start at decision block 302. In decision block 302, the confidence and intent module 62 determines the type of driving allowed in the current driving lane of the specific remote vehicle 42, where the type of driving includes only going straight through and allowed turns. In response to determining that the type of driving allowed in the current driving lane is only going straight through, method 300 may proceed to block 304.

[0051] In block 304, the confidence and intent module 62 sets the intent 88 to a connecting exit lane having a length represented as an intent distance x, where x is in meters. The intent distance is the distance measured along the driving path of the specific remote vehicle 42, which is measured from the current position to the start of the predicted exit lane plus a predetermined distance in the exit lane. It should be understood that the intent distance includes a minimum length specified by calibration parameters. The confidence and intent module 62 also sets the initial confidence level of the intent 88 to high because it is obvious that since turning is not allowed, the specific remote vehicle 42 will typically continue to drive in the connecting exit lane. Method 300 may then proceed to decision block 308, which will be described below.

[0052] Returning to decision block 302, in response to determining that the current driving lane of the specific remote vehicle 42 allows turning, method 300 may proceed to block 306. In block 306, the confidence and intent module 62 sets multiple values for the intent 88, where each value corresponds to the length of a potential connecting exit lane. The length is represented as an intent distance x(i), where x is in meters and i represents the number of potential connecting exit lanes. The confidence and intent module 62 also sets the initial confidence level of the intent 88 for each potential exit lane based on vehicle dynamics and any traffic lights. For example, in an embodiment, the initial confidence level is a function of speed, acceleration, and traffic lights. In a particular case, the traffic lights may affect the confidence level. For example, if the left-turn lane currently has a red light but the straight-through lane has a green light, and if the specific remote vehicle 42 decelerates when approaching the two traffic lights, then the specific remote vehicle 42 is likely to plan a left turn. However, if the specific remote vehicle 42 does not decelerate, then the specific remote vehicle 42 is likely to plan to go straight. Method 300 may then proceed to decision block 308.

[0053] In decision block 308, the initially determined confidence level, determined in block 304 or block 306, is then compared with the cached location information 92 from vehicle database 90, where the cached location information indicates a previously calculated confidence level. The confidence and intent module 62 compares the initially determined confidence level with the previously calculated confidence level. In response to determining that the initially determined confidence level is greater than or equal to the previously calculated confidence level, method 300 proceeds to block 310. However, in response to determining that the initially determined confidence level is less than or equal to the previously calculated confidence level, method 300 proceeds to block 312.

[0054] In block 310, the confidence and intent module 62 increases the initially determined confidence level by a predetermined value and then sets the confidence level 86 to the initially determined confidence level. Method 200 can then terminate.

[0055] In block 312, the confidence and intent module 62 sets the confidence level 86 to the initially determined confidence value. Method 200 can then terminate.

[0056] Figure 6 is a flowchart showing method 400 for determining Figure 2 the context 84 and intent 88 of a particular remote vehicle 42 shown in Figure 1 , Figure 2 , Figure 3 and Figure 6 . Referring generally to

[0057] In block 402, the tracking and detection module 50 of controller 20 receives a cooperative infrastructure sensing message 46 that includes sensed perception data from a perception device such as an infrastructure camera. Method 400 can then proceed to block 404. Figure 2 In block 404, the tracking and detection module 50 of controller 20 determines a plurality of vehicle parameters 68 associated with a particular remote vehicle 42 based on the sensed perception data from the perception device. As mentioned above, in addition to image data 64 collected by the remote infrastructure 44 (i.e.,

[0058] the red light camera seen in Figure 3 ), the plurality of vehicle parameters 68 also indicate the location, location geometry, detection time, and identifier of the particular remote vehicle 42. Method 400 can then proceed to block 406.

[0059] In block 408, the Kalman filter determines a plurality of error recovery vehicle parameters 78 associated with a particular remote vehicle 42 based on the noise associated with converting world coordinate pairs X, Y into image frame coordinates. It should be understood that in some embodiments, block 408 may be omitted, and alternatively the plurality of vehicle parameters 68 are not adjusted based on the noise associated with converting world coordinate pairs X, Y into image frame coordinates. Method 400 may then proceed to block 410.

[0060] In block 410, the positioning and map matching module 58 of the controller 20 associates the particular remote vehicle 42 with a particular driving lane of a road based on map data 80 from the road geometry database 82. Method 400 may then proceed to block 412.

[0061] In block 412, the positioning and map matching module 58 of the controller 20 determines possible maneuvers, possible exit lanes for the particular remote vehicle 42, and speed limits for the particular remote vehicle 42 for a particular driving lane based on map data 80 from the road geometry database 82. Method 400 may then proceed to block 414.

[0062] In block 414, the situation module 60 of the controller 20 determines the situation 84 of the particular remote vehicle 42 based on the plurality of error recovery vehicle parameters 78, possible maneuvers, possible exit lanes for the particular remote vehicle 42, and speed limits associated with the particular remote vehicle 42. Method 200 may then proceed to block 416.

[0063] In block 416, the confidence and intent module 62 of the controller 20 determines the confidence 86 and intent 88 of the particular remote vehicle 42 based on the plurality of error recovery vehicle parameters 78, possible maneuvers, possible exit lanes for the particular remote vehicle 42, and speed limits associated with the particular remote vehicle 42. Method 400 may then terminate.

[0064] Generally referring to the drawings, the disclosed communication system provides various technical effects and benefits. Specifically, the communication system provides a method for determining the situation and intent of a remote vehicle in the absence of available situation and intent based on currently available information.

[0065] A controller can refer to an electronic circuit, combinational logic circuit, field programmable gate array (FPGA), a processor (shared, dedicated, or group) that executes code, or some or all combination of the foregoing (such as in a system-on-chip), or a portion thereof. Additionally, the controller can be microprocessor-based, such as a computer having at least one processor, memory (RAM and / or ROM), and associated input and output buses. The processor can operate under the control of an operating system resident in the memory. The operating system can manage computer resources such that computer program code, implemented as one or more computer software applications (such as an application resident in the memory), can have instructions executed by the processor. In an alternative embodiment, the processor can execute the application directly, in which case the operating system can be omitted.

[0066] The description of the present disclosure is merely exemplary in nature, and variations that do not depart from the gist of the present disclosure are intended to be within the scope of the present disclosure. Such variations should not be regarded as a departure from the spirit and scope of the present disclosure.

Claims

1. A communication system for determining the situation and intention of a specific remote vehicle located in the surrounding environment of a host vehicle, the communication system comprises: One or more controllers for receiving the sensed perception data, the sensed perception data including the sensed perception data related to the specific remote vehicle, wherein the one or more controllers execute instructions to: Determine a plurality of vehicle parameters related to the specific remote vehicle based on the sensed perception data; Associate the specific remote vehicle with a specific driving lane of a road based on map data, wherein the map data indicates information related to the driving lanes of the road along which the specific remote vehicle is traveling; Determine possible maneuvers, possible exit lanes, and speed limits of the specific remote vehicle for the specific driving lane based on the map data; Determine the situation and intention of the specific remote vehicle based on the plurality of vehicle parameters, the possible maneuvers, the possible exit lanes of the specific remote vehicle, and the speed limits related to the specific remote vehicle; Convert a plurality of coordinate pairs based on a world coordinate system into image frame coordinates for noise modeling based on a homography matrix, wherein the coordinate pairs represent a monitored area of the surrounding environment of the host vehicle; and Determine a plurality of error recovery vehicle parameters related to the specific remote vehicle by a Kalman filter based on the noise associated with converting the coordinate pairs based on the world coordinate system into the image frame coordinates.

2. The communication system according to claim 1, wherein, The one or more controllers execute instructions to: Divide an image representing the monitored area of the surrounding environment into a plurality of pixel bins; Determine how many coordinate pairs based on the world coordinate system map to each pixel bin of the image; and Determine a distance covariance map and a speed covariance map based on each pixel bin that is part of the image.

3. The communication system according to claim 2, wherein, The one or more controllers execute instructions to: Reproduce image data representing the specific remote vehicle; Execute an object detection algorithm to detect the specific remote vehicle in the image data, wherein the detected specific remote vehicle is the detected object pixels; and Match the detected object pixels with the speed covariance map and the distance covariance map.

4. The communication system according to claim 3, wherein, The one or more controllers execute instructions to: Determine the noise associated with a bounding box based on a plurality of still images of the specific remote vehicle; Determine the pixel bins affected by the noise associated with the bounding box; Calculate an average speed covariance matrix and an average distance covariance matrix for each affected pixel bin; Match the pixels belonging to the detected object with the speed covariance map and the distance covariance map; and Send the world coordinates of the detected object, the matched speed covariance, and the matched distance covariance to a state tracking module based on a Kalman filter.

5. The communication system according to claim 1, wherein, The one or more controllers execute instructions to: Determine when the particular remote vehicle is in a small lane; In response to determining that the particular remote vehicle is in the small lane, set the situation to be equal to the distance traveled by the particular remote vehicle in the small lane plus the distance traveled in an adjacent lane; And In response to determining that the particular remote vehicle is not in the small lane, set the situation to be equal to the length of the current driving lane.

6. The communication system according to claim 1, Wherein, The one or more controllers execute instructions to: Determine the type of driving permitted in the current driving lane of the particular remote vehicle, wherein the type of driving includes only going straight through and permitted turns; and In response to determining that the type of driving permitted in the current driving lane is only going straight through, set the intention to be a connecting exit lane having a length represented as an intended distance.

7. The communication system according to claim 1, Wherein, The one or more controllers execute instructions to: Determine the type of driving permitted in the current driving lane of the particular remote vehicle, wherein the type of driving includes only going straight through and permitted turns; and In response to determining that the current driving lane of the particular remote vehicle permits turning, set multiple values for the intention, wherein each value corresponds to the length of a potential connecting exit lane.

8. The communication system according to claim 1, Wherein, The one or more controllers execute instructions to: Determine a confidence level indicating the probability that the intention is accurate.

Images