Systems and methods are provided for implementing hybrid prediction. Hybrid prediction integrates two deep learning based trajectory prediction approaches: grid-based approaches and graph-based approaches. Hybrid prediction techniques can achieve enhanced performance by combining the grid and graph approaches in a manner that incorporates appropriate inductive biases for different elements of a high-dimensional space. A hybrid prediction framework processor can generate trajectory predictions relating to movement of agents in a surrounding environment based on a prediction model generating using hybrid prediction. Trajectory predictions output from the hybrid prediction framework processor can be used to control an autonomous vehicle. For example, the autonomous vehicle can perform safety-aware and autonomous operations to avoid oncoming objects, based on the trajectory predictions.

In a system trigger mode and a driver trigger mode, when a start condition for an automated lane change function by autonomous travel control is satisfied, lane change information as to whether to accept execution of the automated lane change function is presented; when an acceptance input of accepting the execution is detected, a determination is made as to whether the lane change is possible; and the automated lane change function is executed when it is determined possible. In the driver trigger mode, when a predetermined lane change instruction operation is performed, a determination is made as to whether the lane change is possible, and when a determination is made that lane change is possible, the automated lane change function is executed. The time for the lane change determination by the driver trigger mode is set shorter than the time for the lane change determination by the system trigger mode.

The travel control device, when determining that the traffic congestion is occurred in the other lane, detects, behind another vehicle in the other lane, an approach space; determines whether the approach space meets a predetermined condition; when determining that the approach space meets the predetermined condition, sets the target posture of the own vehicle at the target position behind the other vehicle in the other lane based on the shape of the approach space; and generates a target traveling trajectory from a current position of the own vehicle to the target position. The travel control device controls the motion of the own vehicle so that the own vehicle tracks the target traveling trajectory.

A method of maintaining vehicle formation includes receiving a desired cross track offset distance and a desired along track offset distance between a lead vehicle and a follower vehicle; receiving a current position, a current yaw rate, and a current speed of the lead vehicle; determining a current turn radius of the lead vehicle based on the current yaw rate and the current speed of the lead vehicle; determining a projected turn radius of the follower vehicle based on the current turn radius of the lead vehicle, the desired cross track offset distance, and the desired along track offset distance; determining a commanded curvature and a next speed of the follower vehicle based on a current position of the follower vehicle and the projected turn radius of the follower vehicle; and outputting the next speed and the commanded curvature to a control system of the follower vehicle.

A method of maintaining vehicle formation includes receiving a desired formation distance between a lead vehicle and a follower vehicle; receiving a pre-planned path for the follower vehicle; and defining a dynamic zone around a current position of the lead vehicle. The dynamic zone has a boundary characterized by a first radius from the current position of the lead vehicle. The first radius can be substantially equal to the desired formation distance. The method further includes determining a next speed of the follower vehicle based on a current position of the follower vehicle with respect to the boundary of the dynamic zone; determining a commanded curvature of the follower vehicle based on the current position of the follower vehicle with respect to the pre-planned path; and outputting the next speed and the commanded curvature to a control system of the follower vehicle for navigation of the follower vehicle.

The present disclosure relates to a vehicle and associated method capable of effectively responding to a cut-in of a nearby vehicle in various road conditions. The method includes obtaining driving situation information; drawing an integrated lane by selectively applying a lanelink, a lainside, and a point level path (PLP) based on the obtained driving situation information; determining a cut-in target based on the integrated lane and a predicted path of each of at least one nearby vehicle; calculating a control point to be followed for driving control of an ego vehicle based on an intersection of a predicted path of the cut-in target and the integrated lane; generating a speed profile and a driving path based on the calculated control point; and performing driving control based on a parameter corresponding to the speed profile and the driving path.

The technology relates to planning trajectories for self-driving vehicles in order to transport passengers or cargo from a pickup location to a destination. Trajectory planning includes generating a speed plan for an upcoming portion of the trip in view of one or more constraints. The constraints may be due to proximity to an adjacent vehicle or other road user, and can include projected overlaps between the vehicle and other objects in the vehicle's nearby environment. Certain constraints may be binary or otherwise discontinuous in nature, in which a condition either exists at a given point in time or it does not. Noise in sensor data or prediction models may trigger such binary conditions, which in turn may cause the vehicle to alter the speed plan. Aspects of the technology employ permeable speed constraints that enable the vehicle to avoid problems associated with discontinuous constraints.

System and methods are provided that allow users of shared vehicles to benefit from an enhanced user experience that seamlessly unlocks and/or provides access to features for autonomous vehicles by proactively computing an interest index based on detected contextual behavioral patterns of the pedestrians such as the trajectory a candidate passenger is walking given a locational context.

The method for dynamically updating an environmental representation of an autonomous agent can include: receiving a set of inputs S210; generating an environmental representation S220; and updating the environmental representation S230. Additionally or alternatively, the method S200 can include providing the environmental representation to a planning module S240 and/or any other suitable processes. The method S200 functions to generate and/or dynamically update an environmental representation to facilitate control of an autonomous agent.

A system trains a generator unit and a discriminator unit simultaneously. The generator unit is configured to determine a future trajectory of at least one other road user in the environment of a vehicle considering an observed trajectory of the at least one other road user. The discriminator unit is configured to determine whether the determined future trajectory of the other road user is an actual future trajectory of the other road user. The system is configured to train the generator unit and the discriminator unit simultaneously with gradient descent.