A computer, including a processor and a memory, the memory including instructions to be executed by the processor to generate a first color image of a road environment, determine one or more value decompositions of one or more of the red, green, and blue channels of the first color image, obtain one or more modified singular value decompositions by modifying respective ones of the singular value decompositions by a non-linear equation and reconstruct a second color image based on the modified one or more singular value decompositions. The instructions can include further instructions to train a deep neural network based on the second color image and operate a vehicle based on the deep neural network.

Examples of the disclosure relate to example devices and methods for generating a map of a vehicle’s passenger compartment and controlling features of the vehicle based on the map. An example vehicle with a passenger compartment mapping system includes one or more sensors to capture sensor data, which includes depth data for a passenger compartment of the vehicle. The system also includes one or more processors to generate a passenger compartment map from the sensor data, and control one or more subsystems of the vehicle based on the passenger compartment map.

A method of providing a reinforcement learning, RL, agent for decision-making to be used in controlling an autonomous vehicle. The method includes: a plurality of training sessions, in which the RL agent interacts with a first environment including the autonomous vehicle, each training session having a different initial value and yielding a state-action value function Q.sub.k(s, a) dependent on state and action; an uncertainty evaluation on the basis of a variability measure for the plurality of state-action value functions evaluated for one or more state-action pairs corresponding to possible decisions by the trained RL agent; additional training, in which the RL agent interacts with a second environment including the autonomous vehicle, wherein the second environment differs from the first environment by an increased exposure to a subset of state-action pairs for which the variability measure indicates a relatively higher uncertainty.

A state estimation method includes a first step of preparing a plurality of first estimation models estimating a state of an object to be monitored based on a state variable measured with respect to the state of the object to be monitored, a second step of preparing a second estimation model estimating, from among the plurality of first estimation models, which one of the first estimation models estimates the state of the object to be monitored with the highest accuracy based on the state variable measured with respect to the object to be monitored or a state variable measured with respect to a utilization device that runs by using the object to be monitored, and a third step of outputting a result of estimation of the state of the object to be monitored of the one first estimation model estimated by the second estimation model.

The present disclosure provides a method and a server for platooning. The method provides: obtaining, based on a platooning request message, a first vehicle type and first kinematic information of a vehicle to join a platoon, a second vehicle type and second kinematic information of a current tail of the platoon, first sensor operating status information of the vehicle to join the platoon, and second sensor operating status information of the current tail vehicle; performing vehicle kinematic determination to obtain a kinematic determination result; performing sensor determination to obtain a sensor determination result; transmitting a confirmation request message to a vehicle of the platoon when the determination results are both successful; and controlling, upon receiving a request approval message, the vehicle to join the platoon to establish a V2V communication connection with each vehicle in the platoon. The method can achieve platoon without any road side unit.

A vehicle includes an actuator, a drivetrain configured to receive mechanical power from the actuator, an accelerator pedal position sensor configured to output a driver-demanded torque, and a controller in electric communication with the sensor and the actuator. The controller is programmed to receive the driver-demanded torque and output a shaped torque command to mitigate driveline disturbances caused by backlash and shaft compliance.

In a self-driving autonomous vehicle, a controller architecture includes multiple processors within the same box. Each processor monitors the others and takes appropriate safe action when needed. Some processors may run dormant or low priority redundant functions that become active when another processor is detected to have failed. The processors are independently powered and independently execute redundant algorithms from sensor data processing to actuation commands using different hardware capabilities (GPUs, processing cores, different input signals, etc.). Intentional hardware and software diversity improves fault tolerance. The resulting fault-tolerant/fail-operational system meets ISO26262 ASIL D specifications based on a single electronic controller unit platform that can be used for self-driving vehicles.

Techniques for using a set of variables to estimate a vehicle velocity of a vehicle are discussed herein. A system may determine an estimated velocity of the vehicle using a minimization based on an initial estimated velocity, steering angle data and wheel speed data. The system may then control an operation of the vehicle based at least in part on the estimated velocity.

A method for synchronizing a clock-free computational node in a vehicle such that the clock-free computational node is able to perform an action at a pre-defined time, including generating a mapping between at least one current counter value received from the clock-free computational node and at least one current time information received from a clock unit. Moreover, the method includes determining a pre-defined counter value being associated with the pre-defined time based on the mapping. Furthermore, the method includes providing the pre-defined counter value to the clock-free computational node. Additionally, a method for performing an action at a pre-defined time using a clock-free computational node is presented. The disclosure is directed to a corresponding first computational node including the clock unit and a second computational node being clock-free and a computational system for a vehicle.

A control arithmetic device comprises a mixed state equation generation unit to generate a plurality of vehicle state equations each including one or more first state variables that are acquisition targets of one or more internal sensors installed in a vehicle, and generate a first mixed state equation by weighting each of the vehicle state equations using a first weighting function, a vehicle state acquisition unit to acquire a current value of each first state variable by the one or more internal sensors, a target route generation unit to generate a target route of the vehicle based on peripheral information acquired by one or more external sensors installed in the vehicle, and a target value arithmetic unit to calculate a target control value for the vehicle to travel along the target route based on the first mixed state equation and the current value of each first state variable, and output the target control value to a control unit that controls the vehicle.