A moving body obstruction detection device includes: a detection section that detects a predetermined moving body within an image that is captured by an imaging section provided at a vehicle; and an inferring section that infers a moving body state that relates to the moving body crossing a road, based on a position of a bounding box that surrounds the moving body detected by the detection section.

Herein, a “perception statistical performance model” (PSPM) for modelling a perception slice of a runtime stack for an autonomous vehicle or other robotic system may be used e.g. for safety/performance testing. A PSPM is configured to receive a computed perception ground truth, and determine from the perception ground truth, based on a set of learned parameters, a probabilistic perception uncertainty distribution, the parameters learned from a set of actual perception outputs generated using the perception slice to be modelled. A simulated scenario is run based on a time series of such perception outputs (with modelled perception errors), but can also be re-run based on perception ground truths directly (without perception errors). This can, for example, be way to ascertain whether perception error was the cause of some unexpected decision within the planner, by determining whether such a decision is also taken in the simulated scenario when perception error is “switched off”.

A prediction device including: a recognizer recognizing a road structure and another vehicle in the vicinity of a subject vehicle; and a predictor predicting a running locus of the other vehicle recognized by the recognizer in the future on the basis of the road structure recognized by the recognizer in a predetermined situation, wherein, in the predetermined situation, in a case in which at least a part of the road structure used for predicting the running locus of the other vehicle in the future is not recognizable for the recognizer, the predictor predicts the running locus of the other vehicle in the future on the basis of a running locus of the other vehicle in the past acquired on the basis of a result of recognition in the past that is acquired by the recognizer.

A moving object determination device applied to a vehicle, the device includes at least a predicted time calculation unit, a time difference calculation unit, and a moving/stationary determination unit. The predicted time calculation unit calculates a predicted time from when an obstacle starts to be detected by the first sensor to when the obstacle will no longer be detected by the second sensor based on a traveling speed of the obstacle and a length of the obstacle. The time difference calculation unit calculates a real time from when the obstacle has been detected by the first sensor to when the obstacle has no longer been detected by the second sensor, and calculates a time difference between the real time and the predicted time. The moving/stationary determination unit determines that the obstacle is a stationary object in response to the time difference being equal to or greater than a predetermined value.

Multiple trajectories for a vehicle are generated based on a road segment. Sensor data is received from at least one sensor. The vehicle is traveling the road segment in accordance with a first trajectory of the multiple trajectories. A potential collision is predicted between the vehicle and an object based on the sensor data and the first trajectory. A set of constraints is determined to avoid the potential collision. The set of constraints is determined based on the sensor data. A maneuver is determined for the vehicle by superimposing each constraint of the set of constraints on each other constraint of the set of constraints. The maneuver includes a second trajectory independent of the multiple trajectories. Instructions are transmitted to a control circuit of the vehicle to override the first trajectory and traverse the road segment according to the second trajectory to perform the maneuver.

A control device to be applied to a vehicle equipped with an imaging device, a ranging device, and a safety device is configured to, based on moving-object detection information around the vehicle acquired from images captured by the imaging device, perform a first actuation process directed to moving objects to actuate the safety device, and based on stationary-object detection information around the vehicle acquired from measurements made by the ranging device, perform a second actuation process directed to stationary objects to actuate the safety device. In the control device, a mask region setting unit is configured to set at least either neighborhood-of-stationary-object regions or far-side regions as a mask region. An actuation restriction unit is configured to, in response to the moving object determined to be present around the vehicle being present in the mask area, restrict performance of the first actuation process on the moving object.

Systems and methods for selecting among different driving modes for autonomous driving of a vehicle may: generate output signals; determine the vehicle proximity information that indicates whether one or more vehicles are within the particular proximity of the vehicle; determine the internal passenger presence information that indicates whether one or more passengers are present in the vehicle; select a first driving mode or a second driving mode based on one or more determinations; and control the vehicle autonomously in accordance with the selection of either the first driving mode or the second driving mode.

A system for data-driven, modular decision making and trajectory generation includes a computing system. A method for data-driven, modular decision making and trajectory generation includes: receiving a set of inputs; selecting a learning module such as a deep decision network and/or a deep trajectory network from a set of learning modules; producing an output based on the learning module; repeating any or all of the above processes; and/or any other suitable processes. Additionally or alternatively, the method can include training any or all of the learning modules; validating one or more outputs; and/or any other suitable processes and/or combination of processes.

A vehicle and a method for controlling vehicle is provided. A vehicle includes a camera which detects an inside lane and acquires an image of an object around the vehicle, a sensor which acquires traveling information of the vehicle, a storage which stores learning data on a steering pattern, and a controller which determines whether a preset learning condition is satisfied based on the inside lane and the object around the vehicle detected by the camera and extracts lane deviation information based on the traveling information acquired by the sensor to update the learning data on the steering pattern when the learning condition is satisfied.

The present application discloses a method for determining an automatic driving feature, an apparatus, a device, a medium and a program product, and relates to automatic driving technology in the field of artificial intelligence. A specific solution is: acquiring scenario information of a plurality of driving scenarios and driving behavior information of an automatic driving system in each of the driving scenarios, where the driving behavior information includes a decision made by the automatic driving system and an execution result corresponding to the decision; determining an automatic driving feature of the automatic driving system according to the scenario information of the plurality of driving scenarios and respective driving behavior information. The automatic driving feature determined through the above process can represent a characteristic of an automatic driving strategy adopted by the automatic driving system.