Abstract: This disclosure provides systems, methods, and devices that support machine learning techniques for vehicle guidance. In one aspect, a method is provided that includes receiving an image frame from an image sensor, determining a projected image frame to correct a vertical axis of the image frame, and generating a top-down representation of the environment based on the projected image frame. The projected image frame may be aligned with the top down representation's vertical axis and may be projected using a cylindrical family projection. The top-down representation may be produced by assembling depth and context matrices from vertically aligned pixels in the projected image frame.

Abstract: In some aspects, an apparatus may receive, from a lidar sensor, a bounding geometry in accordance with one or more data points of a point cloud representing two or more scans of a sensed object. The bounding geometry may have a bounding volume associated with the sensed object. The apparatus may extend the bounding volume of the bounding geometry to enclose at least a subset of the data points in the point cloud. The apparatus may estimate a data point displacement in accordance with two or more subsets of the data points of the point cloud. The apparatus may estimate a velocity of the sensed object in accordance with the data point displacement. Numerous other aspects are described.

Abstract: Systems and techniques are described herein for object detection. For example, a computing device can extract, by an encoder of the computing device, a plurality of features from one or more images of an environment of the computing device. The computing device can determine, based on the plurality of features, a first detection of one or more objects and three-dimensional (3D) coordinates for the one or more objects. The computing device can back-project the 3D coordinates of the one or more objects onto the one or more images. The computing device can determine one or more regions of at least one first image of the one or more images based on the back-projection of the 3D coordinates of the one or more objects. computing device determine, based on the one or more regions of the at least one first image, a second detection of the one or more objects.

Abstract: Techniques and systems are provided for shape estimation. For instance, a process can include: detecting first features of an object in a first frame of an environment, the environment including the object; determining a first set of three-dimensional (3D) points for the first frame based on the detected first features and first distance information obtained for the object; detecting second features of the object in a second frame of the environment; determining a second set of 3D points for the second frame based on the detected second features and second distance information obtained for the object; combining the first set of 3D points and the second set of 3D points to generate a combined set of 3D points; and estimating a shape of the object based on the combined set of 3D points.

Abstract: Image interpolation techniques for multi-sensor training of feature detection models are disclosed. The techniques can include obtaining a detection-and-ranging (DAR) frame captured by a DAR sensor, obtaining a first image frame captured at a first time by a camera and a second image frame captured at a second time by the camera, wherein at a capture time of the DAR frame, a field of view (FOV) of the DAR sensor overlaps an FOV of the camera, and wherein the capture time of the DAR frame is between the first time and the second time, interpolating based on the first image frame and the second image frame to create an interpolated image frame, wherein a nominal capture time of the interpolated image frame corresponds to the capture time of the DAR frame, and training a feature detection model using the DAR frame and the interpolated image frame.

Abstract: This disclosure provides systems, methods, and devices that support machine learning techniques for vehicle guidance. In one aspect, a method is provided that includes receiving an image frame from an image sensor, determining a projected image frame to correct a vertical axis of the image frame, and generating a top-down representation of the environment based on the projected image frame. The projected image frame may be aligned with the top down representation's vertical axis and may be projected using a cylindrical family projection. The top-down representation may be produced by assembling depth and context matrices from vertically aligned pixels in the projected image frame.

Abstract: A method is for providing an object prediction representation. In an embodiment, the method includes simulating movement of a host vehicle and an object in a longitudinal-lateral coordinate system based on their initial positions at the starting time and a movement of the host vehicle and determined movement of the object. Points in time when at least one reference point of the host vehicle and an object reference point of a plurality of object reference points have the same longitudinal position are detected. Further, for each detected point in time, an associated lateral position of the object reference point at the detected point in time is detected.

Abstract: The present invention relates to a method for determining an evasive path for a host vehicle (10), the method i.a. comprising: establishing (S1) a predicted traffic environment (12) of the host vehicle in a time-lateral position domain, the predicted traffic environment comprising an object prediction representation (16) of a traffic object; determining (S2) a start node (24) for the host vehicle; defining (S3) an end node (26) for the host vehicle; placing (S4) several boundary nodes (28) relative to the object prediction representation; setting (S5) node connections (30, 32, 34a-b, 36a-b, 38, 40, 42) between the start node, the boundary nodes, and the end node; and traversing (S6) the nodes using a graph-search algorithm in order to find a path from the start node to the end time with a lowest cost.

Abstract: A collision mitigation system for a vehicle may include a forward sensing system and at least one controller in communication with the forward sensing system. The forward sensing system may be configured to detect objects in front of the vehicle. The at least one controller may be configured to determine whether the vehicle can be steered into a path along side of the vehicle to avoid a collision with detected objects in front of the vehicle and to issue a command to brake the vehicle if the vehicle cannot be steered into the path along side of the vehicle.

Abstract: The present invention relates to a method for determining an evasive path for a host vehicle (10), the method i.a. comprising: establishing (S1) a predicted traffic environment (12) of the host vehicle in a time-lateral position domain, the predicted traffic environment comprising an object prediction representation (16) of a traffic object; determining (S2) a start node (24) for the host vehicle; defining (S3) an end node (26) for the host vehicle; placing (S4) several boundary nodes (28) relative to the object prediction representation; setting (S5) node connections (30, 32, 34a-b, 36a-b, 38, 40, 42) between the start node, the boundary nodes, and the end node; and traversing (S6) the nodes using a graph-search algorithm in order to find a path from the start node to the end time with a lowest cost.

Abstract: A method is for providing an object prediction representation. In an embodiment, the method includes simulating movement of a host vehicle and an object in a longitudinal-lateral coordinate system based on their initial positions at the starting time and a movement of the host vehicle and determined movement of the object. Points in time when at least one reference point of the host vehicle and an object reference point of a plurality of object reference points have the same longitudinal position are detected. Further, for each detected point in time, an associated lateral position of the object reference point at the detected point in time is detected.

Abstract: A method and an arrangement are provided for determining a trajectory for a host vehicle H in order to as smoothly as possible avoid or mitigate a collision. The arrangement may include a processor and at least one of a sensor system or a communication system. The method may include identifying positions of one or more external objects in relation to the host vehicle H within a predefined distance, generating a plurality of trajectories that are valid for enabling the host vehicle H to pass any desired number of external objects, removing any trajectories intersecting with any one of the external objects, estimating lateral position, lateral velocity, lateral acceleration and the lateral jerk that will act on the host vehicle H driving along any one of the trajectories, and selecting the trajectory for which the lateral jerk acting on the host vehicle H is minimized.

Abstract: A method and an arrangement are provided for determining a trajectory for a host vehicle H in order to as smoothly as possible avoid or mitigate a collision. The arrangement may include a processor and at least one of a sensor system or a communication system. The method may include identifying positions of one or more external objects in relation to the host vehicle H within a predefined distance, generating a plurality of trajectories that are valid for enabling the host vehicle H to pass any desired number of external objects, removing any trajectories intersecting with any one of the external objects, estimating lateral position, lateral velocity, lateral acceleration and the lateral jerk that will act on the host vehicle H driving along any one of the trajectories, and selecting the trajectory for which the lateral jerk acting on the host vehicle H is minimized.

Abstract: A system for collision course prediction includes a host vehicle sensor system detecting information relating to a target object including a position and a velocity relative to the host vehicle; a time-to-collision estimator control block calculating an estimated time-to-collision based on a longitudinal distance, longitudinal velocity and longitudinal acceleration of the target object relative to the host vehicle; a lateral distance estimator control block, which estimates the lateral distance between the centers of the host vehicle and target object at the estimated time-to-collision; and a collision course condition determination unit determining, at a determination instant prior to the estimated time-to-collision, a probability that the host vehicle will collide with the target object dependent at least in part upon whether the lateral distance is within a first interval, the first interval based on at least a lateral width of the host vehicle and a lateral width of the target object.

Abstract: A method for determining a safety zone of a foreign object for a forward collision risk reduction system. The method may include determining the type of the foreign object; setting the safety zone to a predetermined value corresponding to the type of the foreign object; determining the traffic behavior of the foreign object; determining if the traffic behavior corresponds to a species of a predetermined group of traffic behaviors; and if the traffic behavior corresponds to a species of a predetermined group of traffic behaviors, modifying the safety zone in accordance with a predetermined species function corresponding to the species.

Abstract: A method and system for assessing vehicle paths in a road environment including a vehicle and external objects includes: determining a position and a dynamic state of the host vehicle and of each of the external objects; generating a plurality of paths that may be followed by the vehicle as it moves toward the external objects; and evaluating a cost associated with each of the paths, the cost based on a lateral acceleration experienced by the vehicle. Each path is made up of path segments, including: an initial path segment corresponding to the vehicle continuing along a selected route without consideration of the external objects, and a maneuver pair for each external object, each maneuver pair including a path segment passing to the left of the external object and a path segment passing to the right of the external object.

Abstract: A method for collision avoidance for a host vehicle includes the following steps; receiving input data relating to a set of objects external to the host vehicle, wherein an object position (r,?), and an object velocity ({dot over (r)}) are associated with each object by a sensor system arranged on the host vehicle, then estimating future trajectories of each external object, while considering influence by the future trajectories of the other external objects.

Abstract: A method for determining a safety zone of a foreign object for a forward collision risk reduction system. The method may include determining the type of the foreign object; setting the safety zone to a predetermined value corresponding to the type of the foreign object; determining the traffic behavior of the foreign object; determining if the traffic behavior corresponds to a species of a predetermined group of traffic behaviors; and if the traffic behavior corresponds to a species of a predetermined group of traffic behaviors, modifying the safety zone in accordance with a predetermined species function corresponding to the species.

Abstract: A collision mitigation system for a vehicle may include a forward sensing system and at least one controller in communication with the forward sensing system. The forward sensing system may be configured to detect objects in front of the vehicle. The at least one controller may be configured to determine whether the vehicle can be steered into a path along side of the vehicle to avoid a collision with detected objects in front of the vehicle and to issue a command to brake the vehicle if the vehicle cannot be steered into the path along side of the vehicle.

Abstract: A method for collision avoidance for a host vehicle includes the following steps; receiving input data relating to a set of objects external to the host vehicle, wherein an object position (r,?), and an object velocity ({dot over (r)}) are associated with each object by a sensor system arranged on the host vehicle, then estimating future trajectories of each external object, while considering influence by the future trajectories of the other external objects.