A vehicle trajectory prediction device is provided. The vehicle trajectory prediction device includes a transceiver, at least one processor, and at least one memory operatively connected with the at least one processor to store at least one instruction causing the at least one processor to perform operations. The operations receive first trajectory data for an ego-vehicle and second trajectory data for at least one neighbor-vehicle, obtain a first feature vector from a first extractor and obtain a second feature vector from a second extractor, obtain an interdependency feature vector between the ego-vehicle and the at least one neighbor-vehicle from a third extractor having mapping data generated by mapping the second feature vector to the second trajectory data as input data, and generate predicted trajectory data of the ego-vehicle from a trajectory generator having the first feature vector and the interdependency feature vector as input data.

Systems and methods for operating a robotic system. The methods comprise: inferring, by a computing device, a first heading distribution for the object from a 3D point cloud; obtaining, by the computing device, a second heading distribution from a vector map; obtaining, by the computing device, a posterior distribution of a heading using the first and second heading distributions; defining, by the computing device, a cuboid on a 3D graph using the posterior distribution; and using the cuboid to facilitate driving-related operations of a robotic system.

Provided are methods for graph exploration for rulebook trajectory generation. Some methods described include generating a next set of alternative trajectories for the vehicle from a next pose, the next set of alternative trajectories representing operation of the vehicle from the next pose, wherein the next pose is located at an end of an identified trajectory. Next trajectories are iteratively identified from corresponding next sets of alternative trajectories, wherein a next trajectory violates a lowest behavioral rule of the hierarchical plurality of rules, the lowest behavioral rule having a priority less than a priority of behavioral rules associated with other trajectories in a corresponding next set of alternative trajectories until a goal pose or timeout is reached to generate a graph. Systems and computer program products are also provided.

Provided are methods for goal determination using an eye tracker device and LiDAR point cloud data, which can include receiving first data characterizing a three-dimensional coordinate associated with a first location. The data obtained via a sensor affixed to a vehicle. Some methods described also include receiving second data characterizing LiDAR point cloud data obtained from a LiDAR device affixed to the vehicle. The LiDAR point cloud data including the three-dimensional coordinate associated with the first location. A visual indication of the first location can be provided on a user interface of the vehicle. The visual indication can be generated based on the first data and the second data. The vehicle can be operated to navigate to the first location responsive to a user input selecting the visual indication. Systems and computer program products are also provided.

Provided are methods for vehicle state estimation based on sensor data, which can include receiving the sensor data generated by one or more sensors, calculating a cornering stiffness value associated with the vehicle, predicting a lateral velocity value associated with the vehicle based on the cornering stiffness value, and outputting a set of vehicle state variables indicative of a current state of the vehicle at least by inputting the lateral velocity value into a recursive filter. Some methods described also include updating the cornering stiffness value based on the set of vehicle state variables, updating the lateral velocity value based on the updated cornering stiffness value, and updating the set of vehicle state variables based on the updated lateral velocity value. Systems and computer program products are also provided.

A system for communicating fleet-specific features to one or more personal electronic devices includes one or more controllers in wireless communication with a centralized computer system including one or more databases for storing the fleet-specific features. The one or more controllers execute instructions to undergo a passive wireless interaction with the one or more personal electronic devices. The passive wireless interaction involves determining the one or more personal electronic devices are located within a predefined proximity around a vehicle without human interaction. In response to undergoing the passive wireless interaction, the one or more controllers transmit one or more fleet-specific features to the one or more personal electronic devices, wherein the one or more fleet-specific features are shared by a fleet of vehicles, where the fleet of vehicles include an immediate vehicle and a group of vehicles that share one or more common attributes.

Techniques to predict object behavior in an environment are discussed herein. For example, such techniques may include inputting data into a model and receiving an output from the model representing a discretized representation. The discretized representation may be associated with a probability of an object reaching a location in the environment at a future time. A vehicle computing system may determine a trajectory and a weight associated with the trajectory using the discretized representation and the probability. A vehicle, such as an autonomous vehicle, can be controlled to traverse an environment based on the trajectory and the weight output by the vehicle computing system.

Methods and systems for deploying autonomous vehicles to form a barricade in a coordinated response to an imminent threat are described. In one embodiment, a method for deploying autonomous vehicles to form a barricade is described. The method includes determining at least one location for a barricade and determining a plurality of autonomous vehicles that are available to form the barricade. The method also includes sending instructions to the plurality of autonomous vehicles to form the barricade at the at least one location. In response to the instructions, the plurality of autonomous vehicles are configured to move to the at least one location and form the barricade.

This disclosure is directed to calibrating sensors mounted on an autonomous vehicle. A dense depth map can be generated in a two-dimensional camera space using point cloud data generated by one of the sensors. Image data from another of the sensors can be compared to the dense depth map in the two-dimensional camera space. Differences determined by the comparison can indicate alignment errors between the sensors. Calibration data associated with the errors can be determined and used to calibrate the sensors without the need for calibration infrastructure.

Disclosed are systems, apparatuses, methods, and computer-readable media to train a neural network model implemented into a perception stack in an autonomous vehicle (AV) for detecting objects. A method includes pretraining an uninitialized ML model to yield a first ML model; training the first ML model with a first testing dataset for a first number of iterations based on a first configuration; analyzing the first ML model based on a convergence of the first ML model and a previous iteration of training; generating a report based on the analysis of the first ML; and after generating the report, training the first ML model to yield a second ML model.