A Curriculum Learning stage

Recent memory used in PPOAgent

Wraps a CARLA Environment, collecting input observations and output actions that can be later used for pre-training or imitation learning purpo

This class is the data access layer for fetching data from the carla server instance for GlobalRoutePlanner

Class for Aggressive agents.

BehaviorAgent implements an agent that navigates scenes to reach a given target destination, by computing the shortest possible path to it.

Context manager to synchronize output from different sensors. Synchronous mode is enabled as long as we are inside this context with

Class for Cautious agents.

A constant learning rate schedule that wraps a constant float learning rate value

This class provides a very high level route plan. Instantiate the class by passing a reference to A GlobalRoutePlannerDAO object.

LocalPlanner implements the basic behavior of following a trajectory of waypoints that is generated on-the-fly. The low-level motion of t

Class for Normal agents.

Creates generic objects with fields given as named arguments.

PIDLateralController implements lateral control using a PID.

PIDLongitudinalController implements longitudinal control using a PID.

Wraps a list of [(carla.Waypoint, RoadOption)] returned by a RoutePlanner.

Provides a high-level route plan.

A wrapper for built-in tf.keras' learning rate schedules

VehiclePIDController is the combination of two PID controllers (lateral and longitudinal) to perform the low level control a vehicle from

Base class to define agents in CARLA

Agent abstract class

AGENT_STATE represents the possible states of a roaming agents

Base extendable environment for the CARLA driving simulator

CARLA benchmark, as described in the paper: "End-to-end Driving via Conditional Imitation Learning" - https://arxiv.org/pdf/1710.02410

Available events (callbacks) related to CARLAEnvironment

The CARLAgent network

Makes an already instantiated CARLAEnvironment be playable with a keyboard

Wraps a CARLA Environment in order to record input observations

Wraps some carla.Color instances.

Interface for learning rate schedule wrappers as dynamic-parameters

A testing-only environment with the same state- and action-space of a CARLA Environment

Compute mean, variance, and standard deviation incrementally.

One camera (front) CARLA Environment

One-camera CARLA Environment with discrete action-space

RoadOption represents the possible topological configurations when moving from a segment of lane to other.

Base class for wrapping sensors.

Kind of tasks that the benchmark supports

Three Camera (front, lateral left and right) CARLA Environment

Three-camera CARLA Environment with discrete action-space