In one embodiment, a plurality of obstacles is sensed in an environment of an automated driving vehicle (ADV). One or more representations are formed to represent corresponding groupings of the plurality of obstacles. A vehicle route is determined in view of the one or more representations, rather than each and every one of the obstacles individually.

A method for driving-related object detection, the method may include receiving an input image by an input of an object detector; and detecting, by an object detector, objects that appear in the input image. The detecting includes searching for (i) a first object having a first size that is within a first size range and belongs to a four wheel vehicle class, (ii) a second object having a second size that is within a second size range and belongs to a subclass out of multiple four wheel vehicle subclasses, (iii) a pedestrian, and (iv) a two wheel vehicle; wherein a maximum of the first size range does not substantially exceed a minimum of the second size range.

An apparatus for post-processing of a decision-making model of an autonomous vehicle receives a decision-making model including a plurality of states. The model is processed using multivariate data that comprises values for at least three observations of a vehicle operational scenario. A slice of the model decision space is generated by fixing values of all except two observations, and modifying the values of the two observations to obtain multiple alternative solutions for the model. The alternative solutions and the modified values form the slice. Each alternative solution is associated with a respective first value of a first observation and a respective second value of a second observation. The apparatus also generates a solution to a modified decision-making model that is the model modified by, for at least one state and at least one of the two observations, modifying a probabilistic transition matrix, a probabilistic observation matrix, or both.

An example operation may include one or more of identifying a person as a roadway obstruction via a transport moving along the roadway, determining a first threat level of the person at a first time, via the transport, when the first threat level is above a threshold, indicating via the transport, to alert at least one of an occupant of the transport and the person, detecting a gesture, via the transport, performed by the person, and the gesture indicates the transport should proceed, responsive to detecting the gesture, determining a second threat level at a second time is below the threshold, and responsive to the second threat level being below the threshold, proceeding, by the transport, along the roadway.

A the method for controlling a vehicle: recognizing at least a position of a traffic participant around a vehicle and a road environment around the traffic participant, setting a risk region for the traffic participant based on at least the recognized position of the traffic participant, correcting the set risk region based on a width of a sidewalk where the traffic participant is present or a width of a roadway around the traffic participant which is the recognized road environment, and controlling a speed and steering of the vehicle based on the corrected risk region.

A method for controlling a vehicle including specifying an object, an attribute of the object, and the strength and direction of wind influencing the object based on an image captured by an imager imaging the surroundings of a vehicle, setting a risk region for the object based on the attribute of the object and the strength and the direction of the wind that are specified, and controlling a speed and steering of the vehicle based on the risk region set by the setter.

A system for trajectory prediction using a predicted endpoint conditioned network includes one or more processors and a memory that includes a sensor input module, an endpoint distribution module, and a future trajectory module. The modules cause the one or more processors to the one or more processors to obtain sensor data of a scene having a plurality of pedestrians, determine endpoint distributions of the plurality of pedestrians within the scene, the endpoint distributions representing desired end destinations of the plurality of pedestrians from the scene, and determine future trajectory points for at least one of the plurality of pedestrians based on prior trajectory points of the plurality of pedestrians and the endpoint distributions of the plurality of pedestrians. The future trajectory points may be conditioned not only on the pedestrian and their immediate neighbors' histories (observed trajectories) but also on all the other pedestrian's estimated endpoints.

An example operation includes one or more of receiving data, by a transport, from a device in proximity to the transport, the data comprises one or more of a speed, a direction, and a distance of the device from the transport, determining, by the transport, a dangerous situation based on the data, obtaining consensus, by the transport, to validate the dangerous situation from one or more of: one or more other devices proximate to the transport, and one or more other transports proximate the transport, and notifying the device, by the transport, based on the consensus, the notification comprises the dangerous situation.

A driving support system executes a risk avoidance control for reducing a risk of collision with an object in front of a vehicle. A risk potential field represents a risk value as a function of position. An obstacle potential field is a risk potential field in which the risk value is maximum at a position of the object and decreases as a distance from the object increases. A vehicle center potential field is the risk potential field in which a valley of the risk value extends in a lane longitudinal direction from a position of the vehicle. A first risk potential field is the sum of the vehicle center potential field and the obstacle potential field. The driving support system executes a steering control such that the vehicle follows the first valley of the risk value represented by the first risk potential field.

A mobile object control device includes a first controller that recognizes a surrounding situation of a mobile object based on an output of a detection device having a space around the mobile object as a detection range and generates a first movement plan for the mobile object in a first period based on the recognized surrounding situation of the mobile object, and a second controller that generates a second movement plan for the mobile object in a second period shorter than the first period, and when the second controller generates label data in which label information indicating different values depending on at least the presence or absence of a moving object is imparted to each of division elements obtained by dividing the space around the mobile object into a finite number, and generates the second movement plan based on the label data.