Techniques for analysis of autonomous vehicle operations are described. As an example, a method of autonomous vehicle operation includes storing sensor data from one or more sensors located on the autonomous vehicle into a storage medium, performing, based on at least some of the sensor data, a simulated execution of one or more programs associated with the operations of the autonomous vehicle, generating, based on the simulated execution of the one or more programs and as part of a simulation, one or more control signal values that control a simulated driving behavior of a simulated vehicle, and providing a visual feedback of the simulated driving behavior of the simulated vehicle on a simulated road.

Systems and methods can improve passenger safety for an Autonomous Vehicle (AV) based on the integration of sensor data captured by the AV's interior and exterior sensors. The AV can determine passenger occupancy data corresponding to where each passenger is detected within the AV by the interior sensors. The AV can determine multiple sets of one or more driving actions that the AV can perform at a future time. The AV can generate crash impact data corresponding to where each passenger is detected from one or more simulated collisions between the AV and one or more objected detected by the exterior sensors when the AV performs one or more sets of driving actions from among the multiple sets. The AV can determine ranked sets of driving actions based on the passenger occupancy data and the crash impact data.

A method includes providing a simulation environment that simulates a vehicle. The method includes providing the vehicular lane centering algorithm to the simulation environment, generating a base scenario for the vehicular lane centering algorithm for use by the simulated vehicle, and extracting, from the simulation environment, traffic lane information. The method also includes measuring performance of the vehicular lane centering algorithm during the base scenario using the extracted traffic lane information and generating a plurality of modified scenarios derived from the base scenario. Each modified scenario of the plurality of modified scenarios adjusts at least one parameter of the base scenario. The method also includes measuring performance of the vehicular lane centering algorithm during the plurality of modified scenarios using the extracted traffic lane information.

A driving control device mounted on a vehicle includes a processor. The processor is configured to create a speed profile. The processor is configured to approximate the speed profile by a predetermined approximate model and estimate a predicted amount of regenerative energy based on an approximation result. The processor is configured to set, based on the predicted amount of regenerative energy, a first region and a second region in the approximation result as a region in which the vehicle travels using an electric motor, the first region being a region from a timing of starting of the vehicle until first time has elapsed from the timing of starting of the vehicle, and the second region being a region from a timing of deceleration of the vehicle until second time has elapsed from the timing of deceleration of the vehicle.

The present application discloses a method and an apparatus for acquiring sample deviation data and an electric device, which relate to the fields of artificial intelligence technology, automatic driving technology, intelligent transportation technology and deep learning technology. The specific implementation solution is: in case of acquiring the sample deviation data, a first driving behavior parameter of a vehicle and a second driving behavior parameter of the vehicle in a simulated automatic driving mode are respectively acquired in the manual driving mode; and it is determined whether there is a deviation between the first driving behavior parameter and the second driving behavior parameter, and if there is a deviation, the vehicle is controlled to acquire the sample deviation data within a preset time period.

A safety system for a vehicle may include one or more processors configured to determine uncertainty data indicating uncertainty in one or more predictions from a driving model during operation of a vehicle; change or update one or more of the driving model parameters to one or more changed or updated driving model parameters based on the determined uncertainty data; and provide the one or more changed or updated driving model parameters to a control system of the vehicle for controlling the vehicle to operate in accordance with the driving model including the one or more changed or updated driving model parameters.

Method and device for autonomous driving of a vehicle on a roadway in a direction of travel. A trajectory for driving on the roadway in the direction of travel is determined. A bending strip of limited length defines the trajectory, wherein the bending strip is fixed at one end thereof in a node which defines a starting point of the trajectory. A course of the trajectory is determined, starting from the starting point, in dependence on a bending line of the bending strip, which line extends, starting from the node, to the other end of the bending strip. A representation of a roadway boundary defines a boundary condition for the determination of the trajectory. A quality measure is defined in dependence on a property of the bending strip. The bending line which satisfies the boundary condition and for which the quality measure has an extremal value is determined.

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.

Provided are methods, systems and computer program products for predicting vehicle crossing and yielding, which can include receiving sensor information indicating an object surrounding a vehicle. Some methods also include determining a future position of the vehicle based on a first trajectory of the vehicle, determining a future position of the object based on a second trajectory of the object, and determining a vehicle control based on the future position of the vehicle and the future position of the object. The methods also include training a model using the vehicle control, the first trajectory of the vehicle, and the second trajectory of the object.

A method for vehicle control, including: obtaining a slip ratio in a current control cycle of a vehicle; calculating a road friction coefficient in the current control cycle of the vehicle by invoking a corresponding calculation strategy according to the slip ratio; and controlling the vehicle in real time by inputting the road friction coefficient into a vehicle control optimization model to obtain a control instruction for the current control cycle of the vehicle.