Abstract: Techniques are discussed herein for determining optimal driving trajectories for autonomous vehicles in complex multi-agent driving environments. A baseline trajectory may be perturbed and parameterized into a vector of vehicle states associated with different segments (or portions) of the trajectory. Such a vector may be modified to ensure the resultant perturbed trajectory is kino-dynamically feasible. The vectorized perturbed trajectory may be input, including a representation of the current driving environment and additional agents, into a prediction model trained to output a predicted future driving scene. The predicted future driving scene, including predicted future states for the vehicle and predicted trajectories for the additional agents in the environment, may be evaluated to determine costs associated with each perturbed trajectory.

Abstract: Trajectory generation for controlling motion or other behavior of an autonomous vehicle may include alternately determining a candidate action and predicting a future state based on that candidate action. The technique may include determining a cost associated with the candidate action that may include an estimation of a transition cost from a current or former state to a next state of the vehicle. This cost estimate may be a lower bound cost or an upper bound cost and the tree search may alternately apply the lower bound cost or upper bound cost exclusively or according to a ratio or changing ratio. The prediction of the future state may be based at least in part on a machine-learned model's classification of a dynamic object as being a reactive object or a passive object, which may change how the dynamic object is modeled for the prediction.

Abstract: Techniques for accurately predicting and avoiding collisions with objects detected in an environment of a vehicle are discussed herein. A vehicle computing device can implement a model to output data indicating costs for potential intersection points between the object and the vehicle in the future. The model may employ a control policy and a time-step integrator to determine whether an object may intersect with the vehicle, in which case the techniques may include predicting vehicle actions by the vehicle computing device to control the vehicle.

Abstract: Techniques for accurately predicting and avoiding collisions with objects detected in an environment of a vehicle are discussed herein. A vehicle computing device can implement a model to output data indicating costs for potential intersection points between the object and the vehicle in the future. The model may employ a control policy and a time-step integrator to determine whether an object may intersect with the vehicle, in which case the techniques may include predicting vehicle actions by the vehicle computing device to control the vehicle.

Abstract: Trajectory generation for controlling motion or other behavior of an autonomous vehicle may include alternately determining a candidate action and predicting a future state based on that candidate action. The technique may include determining a cost associated with the candidate action that may include an estimation of a transition cost from a current or former state to a next state of the vehicle. This cost estimate may be a lower bound cost or an upper bound cost and the tree search may alternately apply the lower bound cost or upper bound cost exclusively or according to a ratio or changing ratio. The prediction of the future state may be based at least in part on a machine-learned model's classification of a dynamic object as being a reactive object or a passive object, which may change how the dynamic object is modeled for the prediction.

Abstract: Techniques for accurately predicting and avoiding collisions with objects detected in an environment of a vehicle are discussed herein. A vehicle computing device can implement a model to output data indicating costs for potential intersection points between the object and the vehicle in the future. The model may employ a control policy and a time-step integrator to determine whether an object may intersect with the vehicle, in which case the techniques may include predicting vehicle actions by the vehicle computing device to control the vehicle.

Abstract: Techniques are discussed herein for executing log-based driving simulations to evaluate the performance and functionalities of vehicle control systems. A simulation system may execute a log-based driving simulation including playback agents whose behavior is based on the log data captured by a vehicle operating in an environment. The simulation system may determine interactions associated with the playback agents, and may convert the playback agents to smart agents during the driving simulation. During a driving simulation, playback agents that have been converted to smart agents may interact with additional playback agents, causing a cascading effect of additional conversions. Converting playback agents to smart agents may include initiating a planning component to control the smart agent, which may be based on determinations of a destination and/or driving attributes based on the playback agent.

Abstract: Techniques are discussed herein for executing log-based driving simulations to evaluate the performance and functionalities of vehicle control systems. A simulation system may execute a log-based driving simulation including playback agents whose behavior is based on the log data captured by a vehicle operating in an environment. The simulation system may determine interactions associated with the playback agents, and may convert the playback agents to smart agents during the driving simulation. During a driving simulation, playback agents that have been converted to smart agents may interact with additional playback agents, causing a cascading effect of additional conversions. Converting playback agents to smart agents may include initiating a planning component to control the smart agent, which may be based on determinations of a destination and/or driving attributes based on the playback agent.

Abstract: A trained classifier to be used with a navigation algorithm for use with mobile robots to compute safe and efficient trajectories. An offline learning process is used to train a classifier for the navigation algorithm (or motion planner), and the classifier functions, after training is complete, to accurately detect intentions of humans within a space shared with the robot to block the robot from traveling along its current trajectory. At runtime, the trained classifier can be used with regression based on past trajectories of humans (or other tracked, mobile entities) to predict where the humans will move in the future and whether the humans are likely to be blockers. The planning algorithm or motion planner generates trajectories based on predictions of human behavior that allow the robot to navigate amongst crowds of people more safely and efficiently.

Abstract: A trained classifier to be used with a navigation algorithm for use with mobile robots to compute safe and efficient trajectories. An offline learning process is used to train a classifier for the navigation algorithm (or motion planner), and the classifier functions, after training is complete, to accurately detect intentions of humans within a space shared with the robot to block the robot from traveling along its current trajectory. At runtime, the trained classifier can be used with regression based on past trajectories of humans (or other tracked, mobile entities) to predict where the humans will move in the future and whether the humans are likely to be blockers. The planning algorithm or motion planner generates trajectories based on predictions of human behavior that allow the robot to navigate amongst crowds of people more safely and efficiently.