Abstract: Drive envelope determination is described. In an example, a vehicle can capture sensor data while traversing an environment and can provide the sensor data to computing system(s). The sensor data can indicate agent(s) in the environment and the computing system(s) can determine, based on the sensor data, a planned path through the environment relative to the agent(s). The computing system(s) can also determine lateral distance(s) to the agent(s) from the planned path. In an example, the computing system(s) can determine modified distance(s) based at least in part on the lateral distance(s) and information about the agents. The computing system can determine a drive envelope based on the modified distance(s) and can determine a trajectory in the drive envelope.

Abstract: Techniques for controlling a vehicle on and off a route structure in an environment are discussed herein. A vehicle computing system controls the vehicle along a route based on a route-based reference system. The vehicle computing system may determine to operate off the route, such as to operate in reverse, park, etc. The vehicle computing system may modify vehicle operations to an inertial-based reference system to navigate to a location off the route. The vehicle computing system may determine a vehicle trajectory to the location off the route based on a reference trajectory between a location on the route and the location off the route and a corridor associated therewith. The vehicle computing system may transition between the route-based reference system and the inertial-based reference system, based on a determination to operate on or off the route.

Abstract: Techniques for determining predictions on a top-down representation of an environment based on object movement are discussed herein. Sensors of a first vehicle (such as an autonomous vehicle) may capture sensor data of an environment, which may include object(s) separate from the first vehicle (e.g., a vehicle, a pedestrian, a bicycle). A multi-channel image representing a top-down view of the object(s) and the environment may be generated based in part on the sensor data. Environmental data (object extents, velocities, lane positions, crosswalks, etc.) may also be encoded in the image. Multiple images may be generated representing the environment over time and input into a prediction system configured to output a trajectory template (e.g., general intent for future movement) and a predicted trajectory (e.g., more accurate predicted movement) associated with each object. The prediction system may include a machine learned model configured to output the trajectory template(s) and the predicted trajector(ies).

Abstract: Techniques for determining predictions on a top-down representation of an environment based on vehicle action(s) are discussed herein. Sensors of a first vehicle (such as an autonomous vehicle) can capture sensor data of an environment, which may include object(s) separate from the first vehicle (e.g., a vehicle or a pedestrian). A multi-channel image representing a top-down view of the object(s) and the environment can be generated based on the sensor data, map data, and/or action data. Environmental data (object extents, velocities, lane positions, crosswalks, etc.) can be encoded in the image. Action data can represent a target lane, trajectory, etc. of the first vehicle. Multiple images can be generated representing the environment over time and input into a prediction system configured to output prediction probabilities associated with possible locations of the object(s) in the future, which may be based on the actions of the autonomous vehicle.

Abstract: Techniques are discussed for generating and optimizing a trajectory using closed-form numerical integration in route-relative coordinates. A decision planner component of an autonomous vehicle, for example, can receive or generate a reference trajectory, which may correspond to an ideal route for an autonomous vehicle to traverse through an environment, such as a center of a road segment. Lateral dynamics (e.g., steering angles, curvature values of trajectory segments) and longitudinal dynamics (e.g., velocity and acceleration) can be represented in a single algorithm such that optimizing the reference trajectory (e.g., based on loss functions or costs) can substantially simultaneously optimize the lateral dynamics and longitudinal dynamics in a single convergence operation. In some cases, the trajectory can be used to control the autonomous vehicle to traverse an environment.

Abstract: In autonomous driving, it is often useful to plan trajectories in a curvilinear coordinate frame with respect to some reference trajectory, like a path produced by a hi-level route planner. This disclosure includes techniques for developing efficient approximate path coordinate motion primitives appropriate for fast planning in autonomous driving scenarios. These primitives are approximate in that particular quantities, like the path length, acceleration, and track offset trajectory, are known with some degree of certainty, and values that depend on the curvature of the reference path can be bound. Such approximate motion primitives can be used to control the autonomous vehicle to follow the trajectory in an environment.

Abstract: Drive envelope determination is described. In an example, a vehicle can capture sensor data while traversing an environment and can provide the sensor data to computing system(s). The sensor data can indicate agent(s) in the environment and the computing system(s) can determine, based on the sensor data, a planned path through the environment relative to the agent(s). The computing system(s) can also determine lateral distance(s) to the agent(s) from the planned path. In an example, the computing system(s) can determine modified distance(s) based at least in part on the lateral distance(s) and information about the agents. The computing system can determine a drive envelope based on the modified distance(s) and can determine a trajectory in the drive envelope.

Abstract: Model-based control of dynamical systems typically requires accurate domain-specific knowledge and specifications system components. Generally, steering actuator dynamics can be difficult to model due to, for example, an integrated power steering control module, proprietary black box controls, etc. Further, it is difficult to capture the complex interplay of non-linear interactions, such as power steering, tire forces, etc. with sufficient accuracy. To overcome this limitation, a recurring neural network can be employed to model the steering dynamics of an autonomous vehicle. The resulting model can be used to generate feedforward steering commands for embedded control. Such a neural network model can be automatically generated with less domain-specific knowledge, can predict steering dynamics more accurately, and perform comparably to a high-fidelity first principle model when used for controlling the steering system of a self-driving vehicle.

Abstract: A vehicle computing system may implement techniques to determine an action for a vehicle to take based on a cost associated therewith. The cost may be based in part on the effect of the action on an object (e.g., another vehicle, bicyclist, pedestrian, etc.) operating in the environment. The vehicle computing system may detect the object based on sensor data and determine an object trajectory based on a predicted reaction of the object to the vehicle performing the action. The vehicle computing system may determine costs associated with safety, comfort, progress, and/or operating rules for each action the vehicle could take based on the action and/or the predicted object trajectory. In some examples, the lowest cost action may be selected for the vehicle to perform.

Abstract: An autonomous vehicle guidance system that generates a path for controlling an autonomous vehicle based at least in part on a data structure generated based at least in part on sensor data that may indicate occupied space in an environment surrounding an autonomous vehicle. The guidance system may receive a grid and generate a grid associated with the grid and the data structure. The guidance system may additionally or alternatively sub-sample the grid (latterly and/or longitudinally) dynamically based at least in part on characteristics determined from the data structure. The guidance system may identify a path based at least in part on a set of precomputed motion primitives, costs associated therewith, and/or a heuristic cost plot that indicates a cheapest cost to move from one pose to another.

Abstract: Techniques for determining predictions on a top-down representation of an environment based on vehicle action(s) are discussed herein. Sensors of a first vehicle (such as an autonomous vehicle) can capture sensor data of an environment, which may include object(s) separate from the first vehicle (e.g., a vehicle or a pedestrian). A multi-channel image representing a top-down view of the object(s) and the environment can be generated based on the sensor data, map data, and/or action data. Environmental data (object extents, velocities, lane positions, crosswalks, etc.) can be encoded in the image. Action data can represent a target lane, trajectory, etc. of the first vehicle. Multiple images can be generated representing the environment over time and input into a prediction system configured to output prediction probabilities associated with possible locations of the object(s) in the future, which may be based on the actions of the autonomous vehicle.

Abstract: Techniques for determining a trajectory for an autonomous vehicle are described herein. In general, determining a route can include utilizing a search algorithm such as Monte Carlo Tree Search (MCTS) to search for possible trajectories, while using temporal logic formulas, such as Linear Temporal Logic (LTL), to validate or reject the possible trajectories. Trajectories can be selected based on various costs and constraints optimized for performance. Determining a trajectory can include determining a current state of the autonomous vehicle, which can include determining static and dynamic symbols in an environment. A context of an environment can be populated with the symbols, features, predicates, and LTL formula. Rabin automata can be based on the LTL formula, and the automata can be used to evaluate various candidate trajectories. Nodes of the MCTS can be generated and actions can be explored based on machine learning implemented as, for example, a deep neural network.

Abstract: A trajectory for an autonomous vehicle (AV) can be generated using curvature segments. A decision planner component can receive a reference trajectory for the AV to follow in an environment. A number of subdivisions (frames) of the reference trajectory may be associated with a curvature value and a tangent vector. Starting with an initial position of the AV, a candidate trajectory can be determined by continuously intersecting a segment with an origin at the initial position of the AV and a reference line associated with a particular frame. The reference line can be substantially perpendicular to the tangent vector of the particular frame. A location of the intersection between the segment and the reference line can be based on a curvature value of the segment. Optimizing a candidate trajectory can include varying curvature values associated with various segments and determining costs of the various candidate trajectories.

Abstract: Techniques are discussed for generating and optimizing a trajectory using closed-form numerical integration in route-relative coordinates. A decision planner component of an autonomous vehicle, for example, can receive or generate a reference trajectory, which may correspond to an ideal route for an autonomous vehicle to traverse through an environment, such as a center of a road segment. Lateral dynamics (e.g., steering angles, curvature values of trajectory segments) and longitudinal dynamics (e.g., velocity and acceleration) can be represented in a single algorithm such that optimizing the reference trajectory (e.g., based on loss functions or costs) can substantially simultaneously optimize the lateral dynamics and longitudinal dynamics in a single convergence operation. In some cases, the trajectory can be used to control the autonomous vehicle to traverse an environment.

Abstract: Model-based control of dynamical systems typically requires accurate domain-specific knowledge and specifications system components. Generally, steering actuator dynamics can be difficult to model due to, for example, an integrated power steering control module, proprietary black box controls, etc. Further, it is difficult to capture the complex interplay of non-linear interactions, such as power steering, tire forces, etc. with sufficient accuracy. To overcome this limitation, a recurring neural network can be employed to model the steering dynamics of an autonomous vehicle. The resulting model can be used to generate feedforward steering commands for embedded control. Such a neural network model can be automatically generated with less domain-specific knowledge, can predict steering dynamics more accurately, and perform comparably to a high-fidelity first principle model when used for controlling the steering system of a self-driving vehicle.

Abstract: Command determination for controlling a vehicle, such as an autonomous vehicle, is described. In an example, individual requests for controlling the vehicle relative to each of multiple objects or conditions in an environment are received (substantially simultaneously) and based on the request type and/or additional information associated with a request, command controllers can determine control commands (e.g., different accelerations, steering angles, steering rates, and the like) associated with each of the one or more requests. The command controllers may have different controller gains (which may be based on functions of distance, distance ratios, time to estimated collisions, etc.) for determining the controls and a control command may be determined based on the all such determined controls.

Abstract: Drive envelope determination is described. In an example, a vehicle can capture sensor data while traversing an environment and can provide the sensor data to computing system(s). The sensor data can indicate agent(s) in the environment and the computing system(s) can determine, based on the sensor data, a planned path through the environment relative to the agent(s). The computing system(s) can also determine lateral distance(s) to the agent(s) from the planned path. In an example, the computing system(s) can determine modified distance(s) based at least in part on the lateral distance(s) and information about the agents. The computing system can determine a drive envelope based on the modified distance(s) and can determine a trajectory in the drive envelope.

Abstract: A method for autonomously operating a driverless vehicle along a path between a first geographic location and a destination may include receiving communication signals from the driverless vehicle. The communication signals may include sensor data from the driverless vehicle and data indicating occurrence of an event associated with the path. The communication signals may also include data indicating that a confidence level associated with the path is less than a threshold confidence level due to the event. The method may also include determining, via a teleoperations system, a level of guidance to provide the driverless vehicle based on data associated with the communication signals, and transmitting teleoperations signals to the driverless vehicle. The teleoperations signals may include guidance to operate the driverless vehicle according to the determined level of guidance, so that a vehicle controller maneuvers the driverless vehicle to avoid, travel around, or pass through the event.

Abstract: Techniques for determining a trajectory for an autonomous vehicle are described herein. In general, determining a route can include utilizing a search algorithm such as Monte Carlo Tree Search (MCTS) to search for possible trajectories, while using temporal logic formulas, such as Linear Temporal Logic (LTL), to validate or reject the possible trajectories. Trajectories can be selected based on various costs and constraints optimized for performance. Determining a trajectory can include determining a current state of the autonomous vehicle, which can include determining static and dynamic symbols in an environment. A context of an environment can be populated with the symbols, features, predicates, and LTL formula. Rabin automata can be based on the LTL formula, and the automata can be used to evaluate various candidate trajectories. Nodes of the MCTS can be generated and actions can be explored based on machine learning implemented as, for example, a deep neural network.

Abstract: A trajectory for an autonomous vehicle (AV) can be generated using curvature segments. A decision planner component can receive a reference trajectory for the AV to follow in an environment. A number of subdivisions (frames) of the reference trajectory may be associated with a curvature value and a tangent vector. Starting with an initial position of the AV, a candidate trajectory can be determined by continuously intersecting a segment with an origin at the initial position of the AV and a reference line associated with a particular frame. The reference line can be substantially perpendicular to the tangent vector of the particular frame. A location of the intersection between the segment and the reference line can be based on a curvature value of the segment. Optimizing a candidate trajectory can include varying curvature values associated with various segments and determining costs of the various candidate trajectories.