Techniques for determining unified futures of objects in an environment are discussed herein. Techniques may include determining a first feature associated with an object in an environment and a second feature associated with the environment and based on a position of the object in the environment, updating a graph neural network (GNN) to encode the first feature and second feature into a graph node representing the object and encode relative positions of additional objects in the environment into one or more edges attached to the node. The GNN may be decoded to determine a predicted position of the object at a subsequent timestep. Further, a predicted trajectory of the object may be determined using predicted positions of the object at various timesteps.

A method, apparatus and computer program product are provided for automatically identify wrong-way driven vehicles in real-time on a roadway. Methods may include: receiving an image from at least one image sensor captured at a location; processing the image using a computer vision model to detect features within the image; identifying, using the computer vision model, a wrong-way driven vehicle in the image; incrementing a confidence value for detection of the wrong-way driven vehicle; and generating an alert indicating the presence of the wrong-way driven vehicle at the location in response to the confidence value satisfying a predetermined value.

A vehicle configured to alert pedestrians. The vehicle includes a driver-assist system including a memory device, a processor, and at least one camera. The memory device includes instructions which, when executed by the processor, cause the processor to detect, utilizing the at least one camera, at least one pedestrian near the vehicle, determine a proximity of the at least one pedestrian to the vehicle, compare the proximity to a threshold proximity, and automatically emit an audible alert from the vehicle in response to the proximity being less than the threshold proximity.

A problem of the present invention is to provide an object detection device etc. that can accurately detect an object regardless of a view angle position of and distance to the object. An object detection device of the present invention has: a stereo distance detection portion 105 that detects a distance to an object; a position detection portion 106 that detects a position of the object; a pose detection portion 111 that detects a pose of the object; a vehicle information input portion that inputs state information about a host vehicle and a different vehicle; a position prediction portion 109 that predicts a position of the different vehicle based on the state information about the host vehicle and the different vehicle; a pose prediction portion 110 that predicts a pose of the different vehicle based on the state information about the host vehicle and the different vehicle; and a determination portion 112 that determines a distance to, a position of, and a pose of the different vehicle in response to the information detected or predicted by the distance detection portion, the position detection portion, the pose detection portion, the position prediction portion, and the pose prediction portion.

Lane level localization techniques for a vehicle utilize a plurality of perception sensor systems each configured to perceive a position of the vehicle relative to its environment, a map system configured to maintain map data that includes lane lines, and a controller configured to detect a position of the vehicle and a first set of lane lines using the plurality of perception sensors, detect a second set of lane lines using the position of the vehicle and the map data, obtain an aligned set of lane lines based on the first and second sets of lane lines, and use the aligned set of lane lines for an autonomous driving feature of the vehicle.

A driving assistance method is provided that assists driving of a vehicle. The driving assistance method stores a traveling route traveled by the vehicle according to a driving operation of a user in a training traveling mode. All of the traveling route, from a traveling start position to a target parking position in the traveling route, is superimposed on one image displayed on a display. A superimposed image superimposing all of the traveling route from the traveling start position to the target parking position in the traveling route is displayed.

A system and method for navigating a vehicle automatically from a current location to a destination location without a human operator is disclosed. The method includes identifying a vehicle location using global positioning system (GPS) data regarding the vehicle. Also included is identifying that the vehicle location is near or at a parking location. Then, using mapping data defined for the parking location. The mapping data at least in part is used to find a path at the parking location to avoid a collision of the vehicle with at least one physical structure when the vehicle is automatically moved at the parking location. The method includes instructing the electronics of the vehicle to proceed with controlling the vehicle to automatically move from the current location to the destination location at the parking location. The electronics use as input at least part of the mapping data and sensor data collected from around the vehicle by at least two vehicle sensors. The path is configured to be updatable dynamically based on changes in the destination location or changes along the path. The destination location is a parking spot for the vehicle at the parking location.

A map-information obstacle-tracking system and a method thereof are provided. The system is installed in a vehicle. The method includes: using a vehicular dynamic positioning module to acquire a position of a vehicle, and using a map-information module to acquire map-information routes of an area neighboring the position of the vehicle; comparing position of the vehicle with the map-information routes to find out candidate routes in the moving direction of the vehicle; determining one of the candidate routes where said obstacle appears, and predicting a moving trajectory of the obstacle; estimating and outputting a position of the obstacle. The present invention is characterized in using map-information and able to acquire the curvature and slope of the front curved lane. Therefore, the present invention can improve the precision of the obstacle position and stabilizes the accuracy of detecting an obstacle in a curved lane.

Systems and methods use cameras to provide autonomous navigation features. In one implementation, a method for navigating a user vehicle may include acquiring, using at least one image capture device, a plurality of images of an area in a vicinity of the user vehicle; determining from the plurality of images a first lane constraint on a first side of the user vehicle and a second lane constraint on a second side of the user vehicle opposite to the first side of the user vehicle; enabling the user vehicle to pass a target vehicle if the target vehicle is determined to be in a lane different from the lane in which the user vehicle is traveling; and causing the user vehicle to abort the pass before completion of the pass, if the target vehicle is determined to be entering the lane in which the user vehicle is traveling.

An autonomous vehicle configured for active sensing may also be configured to weigh expected information gains from active-sensing actions against risk costs associated with the active-sensing actions. An example method involves: (a) receiving information from one or more sensors of an autonomous vehicle, (b) determining a risk-cost framework that indicates risk costs across a range of degrees to which an active-sensing action can be performed, wherein the active-sensing action comprises an action that is performable by the autonomous vehicle to potentially improve the information upon which at least one of the control processes for the autonomous vehicle is based, (c) determining an information-improvement expectation framework across the range of degrees to which the active-sensing action can be performed, and (d) applying the risk-cost framework and the information-improvement expectation framework to determine a degree to which the active-sensing action should be performed.