Methods and systems are provided for determining a road surface condition. In one embodiment, a method includes: receiving vehicle data; constructing, by the processor, a driver behavioral model based on the vehicle data; determining, by the processor, a surface condition based on the driver behavioral model; and generating a signal based on the surface condition.

The technology relates to identifying and addressing aberrant driver behavior. Various driving operations may be evaluated over different time scales and driving distances. The system can detect driving errors and suboptimal maneuvering, which are evaluated by an onboard driver assistance system and compared against a model of expected driver behavior. The result of this comparison can be used to alert the driver or take immediate corrective driving action. It may also be used for real-time or offline training or sensor calibration purposes. The behavior model may be driver-specific, or may be a nominal driver model based on aggregated information from many drivers. These approaches can be employed with drivers of passenger vehicles, busses, cargo trucks and other vehicles.

A method is provided for driving model calibration. The method clusters a plurality of vehicle trajectories into a plurality of datasets for different driving styles based on a score. The score is calculated for each vehicle trajectory by an objective entropy weight method. The method trains, for each of the plurality of datasets for the different driving styles relative to an existing target driving model, a respective neural network which inputs a respective one of the plurality of datasets and outputs a respective parameter for the existing target driver model to obtain a plurality of trained neural networks. The existing target driver model is for simulating human driving behaviors. The method performs, for each trained neural network, an online adaptation of the existing target driving model based on a respective output of each of the plurality of trained neural networks to obtain a plurality of adapted driver models.

An information processing system is provided with a modeling unit. The modeling unit models, based on a degree of similarity among individual travel histories of a plurality of drivers, individual travel histories of at least one or more drivers from among the individual travel histories of the plurality of drivers to construct a driver model indicating a relationship between conditions of the at least one or more drivers and travel environments of at least one or more vehicles.

A system includes a computer programmed to identify, from a first vehicle, one or more second vehicles within a specified distance to the first vehicle. The computer is further programmed to receive data about operations of each of the second vehicles, including trajectory data. Based on the data, the computer is programmed to identify, for each of the second vehicles, a distribution of probabilities of each of a set of potential planned trajectories. The computer is further programmed to determine a planned trajectory for the first vehicle, based on the respective distributions of probabilities of each of the set of potential planned trajectories for each of the second vehicles. The computer is further programmed to provide an instruction to at least one controller associated with the first vehicle based on the determined planned trajectory.

By way of example, the technology disclosed by this document may be implemented in a method that includes aggregating local sensor data from vehicle system sensors, detecting a driver action using the local sensor data, and extracting features related to predicting driver action from the local sensor data during the operation of the vehicle. The method may include adapting a stock machine learning-based driver action prediction model to a customized machine learning-based driver action prediction model using one or more of the extracted features and the detected driver action, the stock machine learning-based driver action prediction model initially generated using a generic model configured to be applicable to a generalized driving populace. In some instances, the method may also include predicting a driver action using the customized machine learning-based driver action prediction model and the extracted features.

An information processing system capable of estimating a driving conduct suited to a driver includes: a history obtainer that obtains a personal driving environment history of each of a plurality of drivers, each of the personal driving environment histories indicating one or more vehicle behaviors selected by the driver, and a driving environment associated with each of the one or more behaviors, the driving environment being a driving environment of the vehicle at a point in time of selection of the behavior it is associated with; and a modeler that models, from a driving environment history including the personal driving environment histories of the plurality of drivers, the personal driving environment history of at least one of the plurality of drivers, to build a driver model indicating a relationship between a behavior and a driving environment for the vehicle of the at least one of the plurality of drivers.

By way of example, the technology disclosed by this document may be implemented in a method that includes receiving stored sensor data describing characteristics of a vehicle in motion at a past time and extracting features for prediction and features for recognition from the stored sensor data. The features for prediction may be input into a prediction network, which may generate a predicted label for a past driver action based on the features for prediction. The features for recognition may be input into a recognition network, which may generate a recognized label for the past driver action based on the features for recognition. In some instances, the method may include training prediction network weights of the prediction network using the recognized label and the predicted label.

Method and system are provided for vehicle accident avoidance carried out with respect to a host vehicle by modeling behavior. The method includes: monitoring a surrounding environment of the host vehicle and detecting other vehicles in a vicinity of the host vehicle by at least one visual sensor. The method further includes: estimating a speed and direction of each of the detected vehicles; calculating one or more projected paths of each of the detected vehicles based on their current estimated speed and direction, the current monitored surrounding environment, and other vehicle projected paths; estimating a probability of intersection of each projected path with the host vehicle; and providing an alert or action to the host vehicle if there is a high probability of intersection.

The disclosure relates to controlling a deceleration request in a one-pedal-driving mode of a vehicle. The deceleration request can comprise a deceleration level being a function of a position of a drive pedal of the vehicle. A corresponding method can comprise receiving, by a system comprising a processor, a vehicle stability information, the vehicle stability information being indicative for driving stability of the vehicle, comparing, by the system, the vehicle stability information to a vehicle stability information threshold, and triggering, by the system, a reduction of the deceleration level of the deceleration request if the vehicle stability information is determined to exceed the vehicle stability information threshold.