A system for providing an alert to an occupant of a vehicle may include one or more processors and a memory. The memory may store a free space detection module, a target detection module, a path prediction module, an activation threshold module, and an alert module. The modules include instructions that cause the one or more processors to determine one or more dimensions of a free space located adjacent to a side of the vehicle, determine one or more dimensions of one or more targets, determine one or more predicted paths of one or more targets, selectively adjust an activation threshold for providing an alert according to the one or more predicted paths, and activate the alert to inform the occupant of a hazard associated with the one or more targets according to whether the one or more predicted paths satisfies the activation threshold.

Among other things, we describe techniques for traffic light estimation using range sensors. A planning circuit of a vehicle traveling on a first drivable region that forms an intersection with a second drivable region receives information sensed by a range sensor of the vehicle. The information represents a movement state of an object through the intersection. A traffic signal at the intersection controls movement of objects through the intersection. The planning circuit determines a state of the traffic signal at the intersection based, in part, on the received information. A control circuit controls an operation of the vehicle based, in part, on the state of the traffic signal at the intersection.

A deep neural network(s) (DNN) may be used to detect objects from sensor data of a three dimensional (3D) environment. For example, a multi-view perception DNN may include multiple constituent DNNs or stages chained together that sequentially process different views of the 3D environment. An example DNN may include a first stage that performs class segmentation in a first view (e.g., perspective view) and a second stage that performs class segmentation and/or regresses instance geometry in a second view (e.g., top-down). The DNN outputs may be processed to generate 2D and/or 3D bounding boxes and class labels for detected objects in the 3D environment. As such, the techniques described herein may be used to detect and classify animate objects and/or parts of an environment, and these detections and classifications may be provided to an autonomous vehicle drive stack to enable safe planning and control of the autonomous vehicle.

An autonomous vehicle can obtain state data associated with an object in an environment, obtain map data including information associated with spatial relationships between at least a subset of lanes of a road network, and determine a set of candidate paths that the object may follow in the environment based at least in part on the spatial relationships between at least two lanes of the road network. Each candidate path can include a respective set of spatial cells. The autonomous vehicle can determine, for each candidate path, a predicted occupancy for each spatial cell of the respective set of spatial cells of such candidate path during at least a portion of a prediction time horizon. The autonomous vehicle can generate prediction data associated with the object based at least in part on the predicted occupancy for each spatial cell of the respective set of spatial cells for at least one candidate path.

An apparatus for controlling a vehicle includes an object selection device configured to select an object intersecting the vehicle at an intersection existing on a driving path of the vehicle, a risk determination device configured to determine a risk during driving of the vehicle based on a predicted path of the object, and a driving control device configured to determine a driving method of the vehicle based on a risk determination result.

A control system of a vehicle that can travel in a first state in which travel control is performed based on a position of a white line on a travel lane and in a second state in which travel control is performed based on a travel position of another vehicle. Periphery information is obtained of the vehicle. It is determined, based on the periphery information obtained, whether an emergency vehicle is approaching. A control unit configured to perform control so that travel control in the first state is prioritized when it is determined that the emergency vehicle is not approaching. Travel control in the second state is prioritized when it is determined that the emergency vehicle is approaching.

A vehicle behavior evaluation device: sets multiple possible behaviors of an own vehicle when the own vehicle travels along a planned route; generates, using a function device, traveling situation related reward data by learning; calculates, using the function device, a reward to each possible behavior of the own vehicle by considering a traveling state of different vehicle; and evaluates each possible behavior of the own vehicle based on the calculated reward. The traveling situation related reward data is generated by: simulating multiple combinations of situations of the own vehicle and the different vehicle under different environments; assigning a first reward to a first situation of the own vehicle when the own vehicle avoids a contact with the different vehicle; and assigning a second reward that is lower than the first reward to a second situation of the own vehicle when the own vehicle contacts with the different vehicle.

A control device includes: an acquisition unit configured to acquire information indicating a location of a warning target, and the number of the warning targets, the warning target being recognized from an image captured by an image capture device mounted on a moving body; a reception control unit configured to perform a control to receive, from a plurality of external terminals existing near the location of the warning target acquired by the acquisition unit, trajectory information indicating a past movement trajectory of each of the plurality of external terminals; a selection unit configured to select, from among the plurality of external terminals, one or more external terminals that are transmission targets of warning information based on the past movement trajectory; and a transmission control unit configured to perform a control to transmit the warning information to the one or more external terminals selected by the selection unit.

Methods by which an autonomous vehicle may predict actions of other actors are disclosed. A vehicle will assign either a high priority rating or a low priority rating to each actor that it detects. The vehicle will then generate a forecast for each of the detected actors. Some of not all high priority actors will receive a high resolution forecast. Low priority actors, and optionally also some of the high priority actors, will receive a low resolution forecast. The system will the forecasts to predict actions for the actors. The autonomous vehicle will then use the predicted actions to determine its trajectory.

The technology involves to pickups performed by autonomous vehicles. In particular, it includes identifying one or more potential pullover locations adjacent to an area of interest that an autonomous vehicle is approaching. The vehicle detects that a given one of the potential pullover locations is occupied by another vehicle and determines whether the other vehicle will be vacating the given pullover location within a selected amount of time. Upon determining that the other vehicle will be vacating the given potential pullover location within the timeframe, the vehicle determines whether to wait for the other vehicle to vacate the given pullover location. Then a driving system of the vehicle either performs a first action in order to wait for the other vehicle to vacate the given pullover location or performs a second action that is different from the first action when it is determined to not wait.