A control device for a vehicle drive unit is configured to control, based on an operating state of a vehicle, a vehicle drive unit having one or more power sources. The control device includes a processor and a storage device. The storage device is configured to store a vehicle front-rear acceleration prediction model being a machine learning model that receives as an input a command torque and outputs predicted acceleration. The processor is configured to: execute a predicted acceleration calculation process using the vehicle front-rear acceleration prediction model; and execute a command torque calculation process to calculate the command torque that minimizes an evaluation function. The evaluation function minimizes a deviation of the predicted acceleration with respect to a target vehicle front-rear acceleration according to a target torque based on the operating state while reducing a deviation of the command torque with respect to the target torque.

A remote traveling vehicle remotely operated by a remote operator receives remote control information of the remote operator from a server via a communication network. The remote traveling vehicle acquires driving environment information of the remote traveling vehicle, and acquires meandering state information including whether a meandering state of the remote traveling vehicle is detected, based on the driving environment information. When the remote-control information includes detection of the meandering condition of the remote traveling vehicle, the remote traveling vehicle adds a limit on the upper speed limit or limit steering angle of the remote traveling vehicle.

A remote traveling vehicle remotely operated by a remote operator receives remote control information of the remote operator from a server via a communication network. The remote traveling vehicle acquires driving environment information of the remote traveling vehicle, and acquires meandering state information including whether a meandering state of the remote traveling vehicle is detected, based on the driving environment information. When the remote-control information includes detection of the meandering condition of the remote traveling vehicle, the remote traveling vehicle adds a limit on the upper speed limit or limit steering angle of the remote traveling vehicle.

A system for learning optimal driving behavior for autonomous vehicles comprises a deep neural network, a first stage training module, and a second stage training module. The deep neural network comprises a feature learning network configured to receive sensor data from a vehicle as input and output spatial temporal feature embeddings and a decision action network configured to receive the spatial temporal feature embeddings as input and output an optimal driving policy for the vehicle. The first training stage module is configured to, during a first training stage, train the feature learning network using object detection loss. The second stage training module is configured to, during a second training stage, train the decision action network using reinforcement learning.

Real-time decision-making for a vehicle using belief state determination is described. Operational environment data is received while the vehicle is traversing a vehicle transportation network, where the data includes data associated with an external object. An operational environment monitor establishes an observation that relates the object to a distinct vehicle operation scenario. A belief state model of the monitor computes a belief state for the observation directly from the operational environment data. The monitor provides the computed belief state to a decision component implementing a policy that maps a respective belief state for the object within the distinct vehicle operation scenario to a respective candidate vehicle control action. A candidate vehicle control action is received from the policy of the decision component, and a vehicle control action is selected for traversing the vehicle transportation from any available candidate vehicle control actions.

An electronic control device including a sensor fusion processing unit that integrates a plurality of pieces of sensor information having been input from a plurality of sensors. The electronic control device further including a behavior prediction processing unit that obtains a future value in which a future behavior of a target object is predicted based on joint information integrated by the sensor fusion processing unit. The electronic control device further including a comparison unit that compares a future value predicted by the behavior prediction processing unit with output information of each sensor of the sensor fusion processing unit at a predicted time.

An electronic control device including a sensor fusion processing unit that integrates a plurality of pieces of sensor information having been input from a plurality of sensors. The electronic control device further including a behavior prediction processing unit that obtains a future value in which a future behavior of a target object is predicted based on joint information integrated by the sensor fusion processing unit. The electronic control device further including a comparison unit that compares a future value predicted by the behavior prediction processing unit with output information of each sensor of the sensor fusion processing unit at a predicted time.

Devices, methods, and systems may obtain at least one point cloud, segment points in the at least one point cloud into a plurality of segments, train a neural network using known segments and a first loss function to generate a first trained neural network, train the first trained neural network using outlier segments and a second loss function to generate a second trained neural network, and train an extended isolation forest by applying an extended isolation algorithm to features of the known segments and features of the outlier segments to generate an anomaly score for each segment.

Devices, methods, and systems may obtain at least one point cloud, segment points in the at least one point cloud into a plurality of segments, train a neural network using known segments and a first loss function to generate a first trained neural network, train the first trained neural network using outlier segments and a second loss function to generate a second trained neural network, and train an extended isolation forest by applying an extended isolation algorithm to features of the known segments and features of the outlier segments to generate an anomaly score for each segment.

Operation data from one or more vehicle subsystems are input to a vehicle dynamics model. Predicted operation data of the one or more vehicle subsystems are output from the vehicle dynamics model. The operation data and the predicted operation data are input to an optimization program that is programmed to output control directives for the one or more vehicle subsystems. One or more vehicle subsystems are operated according to the output control directives.