Abstract: An autonomous vehicle capable of trajectory prediction may include a first sensor, a second sensor, a processor, a trajectory planner, a low-level controller, and vehicle actuators. The first sensor may be of a first sensor type and may detect an obstacle and a goal. The second sensor may be of a second sensor type and may detect the obstacle and the goal. The processor may perform matching on the obstacle detected by the first sensor and the obstacle detected by the second sensor, model an existence probability of the obstacle based on the matching, and track the obstacle based on the existence probability and a constant velocity model. The trajectory planner may generate a trajectory for the autonomous vehicle based on the tracked obstacle, the goal, and a non-linear model predictive control (NMPC). The low-level controller may implement the trajectory for the autonomous vehicle by driving vehicle actuators.

Abstract: A system and method for providing a comprehensive trajectory planner for a person-following vehicle that includes receiving image data and LiDAR data associated with a surrounding environment of a vehicle. The system and method also include analyzing the image data and detecting the person to be followed that is within an image and analyzing the LiDAR data and detecting an obstacle that is located within a predetermined distance from the vehicle. The system and method further include executing a trajectory planning algorithm based on fused data associated with the detected person and the detected obstacle.

Abstract: According to one aspect, systems and techniques for lane selection may include receiving a current state of an ego vehicle and a traffic participant vehicle, and a goal position, projecting the ego vehicle and the traffic participant vehicle onto a graph network, where nodes of the graph network may be indicative of discretized space within an operating environment, determining a current node for the ego vehicle within the graph network, and determining a subsequent node for the ego vehicle based on identifying adjacent nodes which may be adjacent to the current node, calculating travel times associated with each of the adjacent nodes, calculating step costs associated with each of the adjacent nodes, calculating heuristic costs associated with each of the adjacent nodes, and predicting a position of the traffic participant vehicle.

Abstract: Systems and methods for curiosity development in an agent located in an uncertain environment are provided. In one embodiment, the system includes a goal state module, a curiosity module, and a planning module. The goal module is configured to calculate a goal state of a goal associated with the environment. The curiosity module is configured to determine an uncertainty value for the environment and calculate a curiosity reward based on the uncertainty value. The planning module is configured to update a motion plan based on the goal state and the curiosity reward.

Abstract: A system and method for implementing reward based strategies for promoting exploration that include receiving data associated with an agent environment of an ego agent and a target agent and receiving data associated with a dynamic operation of the ego agent and the target agent within the agent environment. The system and method also include implementing a reward function that is associated with exploration of at least one agent state within the agent environment. The system and method further include training a neural network with a novel unexplored agent state.

Abstract: According to one aspect, cooperative multi-goal, multi-agent, multi-stage (CM3) reinforcement learning may include training a first agent using a first policy gradient and a first critic using a first loss function to learn goals in a single-agent environment using a Markov decision process, training a number of agents based on the first policy gradient and a second policy gradient and a second critic based on the first loss function and a second loss function to learn cooperation between the agents in a multi-agent environment using a Markov game to instantiate a second agent neural network, each of the agents instantiated with the first agent neural network in a pre-trained fashion, and generating a CM3 network policy based on the first agent neural network and the second agent neural network. The CM3 network policy may be implemented in a CM3 based autonomous vehicle to facilitate autonomous driving.

Abstract: A system and method for multi-agent reinforcement learning in a multi-agent environment that include receiving data associated with the multi-agent environment in which an ego agent and a target agent are traveling and learning a single agent policy that is based on the data associated with the multi-agent environment and that accounts for operation of at least one of: the ego agent and the target agent individually. The system and method also include learning a multi-agent policy that accounts for operation of the ego agent and the target agent with respect to one another within the multi-agent environment. The system and method further include controlling at least one of: the ego agent and the target agent to operate within the multi-agent environment based on the multi-agent policy.

Abstract: According to one aspect, uncertainty prediction based deep learning may include receiving, using a memory, a trained neural network policy ? trained based on a first dataset in a first environment, implementing, via a controller, the trained neural network policy ? in a second environment by receiving an input and generating an output y, calculating an uncertainty array U[T] for a time window T, wherein the uncertainty array is indicative of a level of uncertainty associated with an output sample distribution of the output across the time window T based on a temporal divergence, an entropy H, a variational ratio VR, and a standard deviation SD of the output y, and executing, via the controller and one or more systems, an action based on the uncertainty array U[T], such as discontinuing use of the trained neural network policy ?.

Abstract: According to one aspect, systems and techniques for lane selection may include receiving a current state of an ego vehicle and a traffic participant vehicle, and a goal position, projecting the ego vehicle and the traffic participant vehicle onto a graph network, where nodes of the graph network may be indicative of discretized space within an operating environment, determining a current node for the ego vehicle within the graph network, and determining a subsequent node for the ego vehicle based on identifying adjacent nodes which may be adjacent to the current node, calculating travel times associated with each of the adjacent nodes, calculating step costs associated with each of the adjacent nodes, calculating heuristic costs associated with each of the adjacent nodes, and predicting a position of the traffic participant vehicle.

Abstract: Autonomous vehicle interactive decision making may include identifying two or more traffic participants and gaps between the traffic participants, selecting a gap and identifying a traffic participant based on a coarse probability of a successful merge between the autonomous vehicle and a corresponding traffic participant, generating an intention prediction associated with the identified traffic participant based on vehicle dynamics of the identified traffic participant, predicted behavior of the identified traffic participant in the absence of the autonomous vehicle, and predicted behavior of the identified traffic participant in the presence of the autonomous vehicle making a maneuver creating an interaction between the identified traffic participant and the autonomous vehicle, generating an intention prediction associated with the autonomous vehicle, calculating an updated probability of a successful interaction between the identified traffic participant and the autonomous vehicle based on the intention pr

Abstract: A system and method for providing autonomous vehicular navigation within a crowded environment that include receiving data associated with an environment in which an ego vehicle and a target vehicle are traveling. The system and method also include determining an action space based on the data associated with the environment. The system and method additionally include executing a stochastic game associated with navigation of the ego vehicle and the target vehicle within the action space. The system and method further include controlling at least one of the ego vehicle and the target vehicle to navigate in the crowded environment based on execution of the stochastic game.

Abstract: Systems and methods for curiosity development in an agent located in an uncertain environment are provided. In one embodiment, the system includes a goal state module, a curiosity module, and a planning module. The goal module is configured to calculate a goal state of a goal associated with the environment. The curiosity module is configured to determine an uncertainty value for the environment and calculate a curiosity reward based on the uncertainty value. The planning module is configured to update a motion plan based on the goal state and the curiosity reward.

Abstract: According to one aspect, a system for reinforcement learning with iterative reasoning may include a memory for storing computer readable code and a processor operatively coupled to the memory, the processor configured to receive a level-0 policy and a desired reasoning level n. The processor may repeat for k=1 . . . n times, the following: populate a training environment with a level-(k?1) first agent, populate the training environment with a level-(k?1) second agent, and train a level-k agent based on the level-(k?1) first agent and the level-(k?1) second agent to derive a level-k policy.

Abstract: A system and method for providing accurate trajectory following for automated vehicles in dynamic environments that include receiving image data and LiDAR data associated with a dynamic environment of a vehicle. The system and method also include processing a planned trajectory of the vehicle that is based on an analysis of the image data and LiDAR data. The system and method further include communicating control signals associated with following the planned trajectory to autonomously control the vehicle to follow the planned trajectory to navigate within the dynamic environment to reach a goal.

Abstract: A system and method for providing a comprehensive trajectory planner for a person-following vehicle that includes receiving image data and LiDAR data associated with a surrounding environment of a vehicle. The system and method also include analyzing the image data and detecting the person to be followed that is within an image and analyzing the LiDAR data and detecting an obstacle that is located within a predetermined distance from the vehicle. The system and method further include executing a trajectory planning algorithm based on fused data associated with the detected person and the detected obstacle.

Abstract: An autonomous vehicle capable of trajectory prediction may include a first sensor, a second sensor, a processor, a trajectory planner, a low-level controller, and vehicle actuators. The first sensor may be of a first sensor type and may detect an obstacle and a goal. The second sensor may be of a second sensor type and may detect the obstacle and the goal. The processor may perform matching on the obstacle detected by the first sensor and the obstacle detected by the second sensor, model an existence probability of the obstacle based on the matching, and track the obstacle based on the existence probability and a constant velocity model. The trajectory planner may generate a trajectory for the autonomous vehicle based on the tracked obstacle, the goal, and a non-linear model predictive control (NMPC). The low-level controller may implement the trajectory for the autonomous vehicle by driving vehicle actuators.

Abstract: A system and method for implementing reward based strategies for promoting exploration that include receiving data associated with an agent environment of an ego agent and a target agent and receiving data associated with a dynamic operation of the ego agent and the target agent within the agent environment. The system and method also include implementing a reward function that is associated with exploration of at least one agent state within the agent environment. The system and method further include training a neural network with a novel unexplored agent state.

Abstract: Autonomous vehicle interactive decision making may include identifying two or more traffic participants and gaps between the traffic participants, selecting a gap and identifying a traffic participant based on a coarse probability of a successful merge between the autonomous vehicle and a corresponding traffic participant, generating an intention prediction associated with the identified traffic participant based on vehicle dynamics of the identified traffic participant, predicted behavior of the identified traffic participant in the absence of the autonomous vehicle, and predicted behavior of the identified traffic participant in the presence of the autonomous vehicle making a maneuver creating an interaction between the identified traffic participant and the autonomous vehicle, generating an intention prediction associated with the autonomous vehicle, calculating an updated probability of a successful interaction between the identified traffic participant and the autonomous vehicle based on the intention pr

Abstract: According to one aspect, cooperative multi-goal, multi-agent, multi-stage (CM3) reinforcement learning may include training a first agent using a first policy gradient and a first critic using a first loss function to learn goals in a single-agent environment using a Markov decision process, training a number of agents based on the first policy gradient and a second policy gradient and a second critic based on the first loss function and a second loss function to learn cooperation between the agents in a multi-agent environment using a Markov game to instantiate a second agent neural network, each of the agents instantiated with the first agent neural network in a pre-trained fashion, and generating a CM3 network policy based on the first agent neural network and the second agent neural network. The CM3 network policy may be implemented in a CM3 based autonomous vehicle to facilitate autonomous driving.

Abstract: A system and method for providing autonomous vehicular navigation within a crowded environment that include receiving data associated with an environment in which an ego vehicle and a target vehicle are traveling. The system and method also include determining an action space based on the data associated with the environment. The system and method additionally include executing a stochastic game associated with navigation of the ego vehicle and the target vehicle within the action space. The system and method further include controlling at least one of the ego vehicle and the target vehicle to navigate in the crowded environment based on execution of the stochastic game.