Abstract: Systems and methods for predicting drivable paths relative to road segments are disclosed. In one implementation, a system includes a processor programmed to access topographical information associated with a road segment; generate a topographical representation of the road segment based on the topographical information; input the topographical representation of the road segment to a trained model, wherein the trained model includes a graph neural network and is configured to predict at least one drivable path relative to the road segment based on the topographical representation of the road segment; receive, from the trained model, information identifying the drivable path; and store the information identifying the drivable path in a map.

Abstract: Systems and methods are provided for vehicle navigation. In one implementation, at least one processor may be programmed to receive, from a camera, a captured image representative of features in an environment of the vehicle. The processor may generate a warped image based on the received captured image, which may simulate a view of the features in the environment of the vehicle from a simulated viewpoint elevated relative to an actual position of the camera. The processor may further identify a road feature represented in the warped image, which may be transformed in one or more respects relative to a representation of the road feature in the captured image. The processor may then determine a navigational action for the vehicle based on the identified feature represented in the warped image and cause at least one actuator system of the vehicle to implement the determined navigational action.

Abstract: A system for automatically mapping a road segment may include: at least one processor programmed to: receive, from at least one camera mounted on a vehicle, a plurality of images acquired as the vehicle traversed the road segment; convert each of the plurality of images to a corresponding top view image to provide a plurality of top view images; aggregate the plurality of top view images to provide an aggregated top view image of the road segment; analyze the aggregated top view image to identify at least one road feature associated with the road segment; automatically annotate the at least one road feature relative to the aggregated top view image; and output to at least one memory the aggregated top view image including the annotated at least one road feature.

Abstract: Systems and methods are provided for vehicle navigation. In one implementation, at least one processor may be programmed to receive, from a camera, a captured image representative of features in an environment of the vehicle. The processor may generate a warped image based on the received captured image, which may simulate a view of the features in the environment of the vehicle from a simulated viewpoint elevated relative to an actual position of the camera. The processor may further identify a road feature represented in the warped image, which may be transformed in one or more respects relative to a representation of the road feature in the captured image. The processor may then determine a navigational action for the vehicle based on the identified feature represented in the warped image and cause at least one actuator system of the vehicle to implement the determined navigational action.

Abstract: Systems and methods are provided for vehicle navigation. In one implementation, at least one processor may receive, from a camera of the host vehicle, at least one image captured from an environment of the host vehicle and identify at least one object represented in the at least one image. The at least one object may be identified based on an output of a neural network configured to process the at least one image using a first plurality of kernels and a second plurality of kernels and at least one of the second plurality of kernels may be symmetric with respect to at least one of the first plurality of kernels, The processor may further determine a navigational action for the host vehicle based on the identified at least one object represented in the at least one image and cause the host vehicle to implement the determined navigational action.

Abstract: Systems and methods are disclosed for mapping lanes for use in vehicle navigation. In one implementation, at least one processing device may be programmed to receive navigational information from a first vehicle and a second vehicle that have navigated along a road segment including a lane split feature; receive at least one image associated with the road segment; determine, from the first navigational information, a first actual trajectory of the first vehicle and a second actual trajectory of the second vehicle; determine a divergence between the first actual trajectory and the second actual trajectory; determine, based on analysis of the at least one image, that the divergence between the first actual trajectory and the second actual trajectory is indicative of the lane split feature; and update a vehicle road navigation model to include a first target trajectory and a second target trajectory that branches from the first target trajectory after the lane split feature.

Abstract: A system for autonomously navigating a vehicle along a road segment may include at least one processor. The at least one processor may be programmed to receive from at least one sensor information relating to one or more aspects of the road segment. The processor may also be programmed to determine a local feature of the road segment based on the received information. Further the processor may be programmed to compare the local feature to a predetermined signature feature for the road segment. The processor may be programmed to determine a current location of the vehicle along a predetermined road model trajectory associated with the road segment based on the comparison of the received information and the predetermined signature feature. The processor may also be programmed to determine an autonomous steering action for the vehicle based on a direction of the predetermined road model trajectory at the determined location.

Abstract: Systems and methods use cameras to provide autonomous navigation features. In one implementation, top-down refinement in lane marking navigation is provided. The system may include one or more memories storing instructions and one or more processors configured to execute the instructions to cause the system to receive from one or more cameras one or more images of a roadway in a vicinity of a vehicle, the roadway comprising a lane marking comprising a dashed line, update a model of the lane marking based on odometry of the one or more cameras relative to the roadway, refine the updated model of the lane marking based on an appearance of dashes derived from the received one or more images and a spacing between dashes derived from the received one or more images, and cause one or more navigational responses in the vehicle based on the refinement of the updated model.

Abstract: Systems and methods are provided for vehicle navigation. In one implementation, at least one processor may be programmed to receive, from a camera, a captured image representative of features in an environment of the vehicle. The processor may generate a warped image based on the received captured image, which may simulate a view of the features in the environment of the vehicle from a simulated viewpoint elevated relative to an actual position of the camera. The processor may further identify a road feature represented in the warped image, which may be transformed in one or more respects relative to a representation of the road feature in the captured image. The processor may then determine a navigational action for the vehicle based on the identified feature represented in the warped image and cause at least one actuator system of the vehicle to implement the determined navigational action.

Abstract: Systems and methods use cameras to provide autonomous navigation features. In one implementation, top-down refinement in lane marking navigation is provided. The system may include one or more memories storing instructions and one or more processors configured to execute the instructions to cause the system to receive from one or more cameras one or more images of a roadway in a vicinity of a vehicle, the roadway comprising a lane marking comprising a dashed line, update a model of the lane marking based on odometry of the one or more cameras relative to the roadway, refine the updated model of the lane marking based on an appearance of dashes derived from the received one or more images and a spacing between dashes derived from the received one or more images, and cause one or more navigational responses in the vehicle based on the refinement of the updated model.

Abstract: Systems and methods use cameras to provide autonomous navigation features. In one implementation, top-down refinement in lane marking navigation is provided. The system may include one or more memories storing instructions and one or more processors configured to execute the instructions to cause the system to receive from one or more cameras one or more images of a roadway in a vicinity of a vehicle, the roadway comprising a lane marking comprising a dashed line, update a model of the lane marking based on odometry of the one or more cameras relative to the roadway, refine the updated model of the lane marking based on an appearance of dashes derived from the received one or more images and a spacing between dashes derived from the received one or more images, and cause one or more navigational responses in the vehicle based on the refinement of the updated model.

Abstract: A system for autonomously navigating a vehicle along a road segment may include at least one processor. The at least one processor may be programmed to receive from at least one sensor information relating to one or more aspects of the road segment. The processor may also be programmed to determine a local feature of the road segment based on the received information. Further the processor may be programmed to compare the local feature to a predetermined signature feature for the road segment. The processor may be programmed to determine a current location of the vehicle along a predetermined road model trajectory associated with the road segment based on the comparison of the received information and the predetermined signature feature. The processor may also be programmed to determine an autonomous steering action for the vehicle based on a direction of the predetermined road model trajectory at the determined location.

Abstract: Systems and methods use cameras to provide autonomous navigation features. In one implementation, top-down refinement in lane marking navigation is provided. The system may include one or more memories storing instructions and one or more processors configured to execute the instructions to cause the system to receive from one or more cameras one or more images of a roadway in a vicinity of a vehicle, the roadway comprising a lane marking comprising a dashed line, update a model of the lane marking based on odometry of the one or more cameras relative to the roadway, refine the updated model of the lane marking based on an appearance of dashes derived from the received one or more images and a spacing between dashes derived from the received one or more images, and cause one or more navigational responses in the vehicle based on the refinement of the updated model.