A driving assistance apparatus can execute deceleration assistance of decelerating a host vehicle independently of an operation by a driver. The driving assistance apparatus is provided with: an acquirer configured to obtain surrounding information associated with a surrounding situation of the host vehicle; and a controller programmed to reduce a deceleration assistance amount associated with the deceleration assistance when execution of the deceleration assistance is released. The controller is programmed to quickly reduce the deceleration assistance amount when execution of the deceleration assistance is released, if a surrounding situation indicated by the obtained surrounding information is a first situation in which the host vehicle can be required to accelerate, in comparison with a second situation in which the host vehicle cannot be required to accelerate.

A collision probability calculation device is provided with: a setting unit that sets a two-dimensional area composed of direction components of a first direction and a second direction, and including a first estimated position where a vehicle is estimated to arrive in the future from a reference position; a first variance calculation unit for calculating a first variance value of first position information in the two-dimensional area; a second variance calculation unit for calculating a second variance value of second position information on a second estimated position where an object is estimated to arrive in the future at the time when the vehicle arrives at the first estimated position; and a probability calculation unit for calculating the probability of a collision between the vehicle and the object, using the two-dimensional area, the first position information, the second position information, the first variance value and the second variance value.

A vehicle computing device may implement techniques to predict behavior of objects or predicted objects in an environment. The techniques may include using a model to determine whether a potential object will emerge from an occluded region in the environment. The model may be configured to use one or more algorithms, classifiers, and/or computational resources to predict an intersection point and/or an intersection time between the potential object and the vehicle. Based on the predicted intersection point and/or the predicted intersection time, the vehicle computing device may control operation of the vehicle.

Techniques for determining a trajectory for an autonomous vehicle are described herein. In general, determining a route can include utilizing a search algorithm such as Monte Carlo Tree Search (MCTS) to search for possible trajectories, while using temporal logic formulas, such as Linear Temporal Logic (LTL), to validate or reject the possible trajectories. Trajectories can be selected based on various costs and constraints optimized for performance. Determining a trajectory can include determining a current state of the autonomous vehicle, which can include determining static and dynamic symbols in an environment. A context of an environment can be populated with the symbols, features, predicates, and LTL formula. Rabin automata can be based on the LTL formula, and the automata can be used to evaluate various candidate trajectories. Nodes of the MCTS can be generated and actions can be explored based on machine learning implemented as, for example, a deep neural network.

Techniques and methods for determining regions. For instance, a vehicle may determine a trajectory of the vehicle and a trajectory of an agent, such as a pedestrian. The vehicle may then determine one or more contextual factors. In some examples, the one or more contextual factors are associated with a location of the agent with respect to a crosswalk, a location of the vehicle with respect to the crosswalk, a state of the crosswalk, and/or the like. The vehicle may then determine the region using the trajectory of the vehicle, the trajectory of the agent, and the one or more contextual factors. Additionally, using a time buffer value and a distance buffer value associated with the region, the vehicle may determine whether to yield to the agent within the region.

A driving assistance device calculates the field of view of a driver based on an output signal of a camera that captures an image of the driver. The driving assistance device calculates a monitoring required region that requires monitoring when driving the vehicle based on information of the periphery of the vehicle. The driving assistance device determines whether the field of view calculated in the field of view calculation process encompasses the monitoring required region. When determined that the calculated field of view does not encompass the monitoring required region, the driving assistance operates predetermined hardware device to cope with the situation.

System, methods, and other embodiments described herein relate to improving trajectory forecasting in a device. In one embodiment, a method includes, in response to receiving sensor data about a surrounding environment of the device, identifying an object from the sensor data that is present in the surrounding environment. The method includes determining category probabilities for the object, the category probabilities indicating semantic classes for classifying the object and probabilities that the object belongs to the semantic classes. The method includes forecasting trajectories for the object based, at least in part, on the category probabilities and the sensor data. The method includes controlling the device according to the trajectories.

Provided are methods and devices for loss evaluation to automated driving. The method includes: taking classes or localizations of observations as tasks of an automated driving model; correcting loss of each of the observations based on real-world scenarios in driving practice. In the present disclosure, the evaluation of algorithms in automated driving can be set with true realistic value in real world scenario; and rectify the misalignment from using of generic evaluation methods to algorithms used in automated driving scenarios.

Described herein is a method and system for assisting a driver of a vehicle (1) to drive with precaution. Vehicle environment monitoring sensors (3a, 3b) determines other road users and particular features associated with a traffic situation of the vehicle (1) and hypotheses are applied related to hypothetical threats that may arise based thereupon. A driver level of attention, required to handle the hypothetical threats, and a time until that level will be required is estimated. A current driver level of attention is derived, from driver-monitoring sensors (4). If determined that the estimated required driver level of attention exceeds the current and the time until the estimated driver level of attention will be required is less than a threshold-time (t.sub.thres), there is produced at least one of visual (5), acoustic (6) and haptic (7) information to a vehicle driver environment, and/or triggered at least one of automated braking (8) and steering (9) of the vehicle (1).

Systems and methods communicate an intent of an autonomous vehicle externally. In one implementation, scan data of a field around a travel path of an autonomous vehicle is obtained. The scan data is captured using at least one sensor. An object in the field around the travel path is determined from the scan data. The object is determined to be mutable or immutable. A navigation condition associated with the object is determined based on whether the object is mutable or immutable. The navigation condition is correlated to a portion of the travel path. Control operation(s) of the autonomous vehicle is determined for the portion of the travel path in response to the navigation condition. A representation link between the control operation(s) of the autonomous vehicle and the object is generated. A representation of the field around the travel path is rendered and includes the representation link.