Techniques for a perception system of a vehicle that can detect and track objects in an environment are described herein. The perception system may include a machine-learned model that includes one or more different portions, such as different components, subprocesses, or the like. In some instances, the techniques may include training the machine-learned model end-to-end such that outputs of a first portion of the machine-learned model are tailored for use as inputs to another portion of the machine-learned model. Additionally, or alternatively, the perception system described herein may utilize temporal data to track objects in the environment of the vehicle and associate tracking data with specific objects in the environment detected by the machine-learned model. That is, the architecture of the machine-learned model may include both a detection portion and a tracking portion in the same loop.

Disclosed are methods, systems, and non-transitory computer readable media that control an autonomous vehicle via at least two sensors. One aspect includes capturing an image of a scene ahead of the vehicle with a first sensor, identifying an object in the scene at a confidence level based on the image, determining the confidence level of the identifying is below a threshold, in response to the confidence level being below the threshold, directing a second sensor having a field of view smaller than the first sensor to generate a second image including a location of the identified object, further identifying the object in the scene based on the second image, controlling the vehicle based on the further identification of the object.

A driving determination device includes an acquirer configured to acquire at least a captured image of a driving body in a driving direction and information that changes with movement of the driving body; a driving level calculator configured to calculate a driving level for evaluating a driving method for the driving body for each predetermined determination item, using at least one of the acquired captured image and the acquired information that changes with the movement of the driving body; an itemized calculator configured to calculate values based on a plurality of the calculated driving levels for each determination item; and an evaluation result calculator configured to calculate a value for comprehensively evaluating the driving method for the driving body, using the values based on the driving levels for each determination item.

Embodiments include apparatus and method for collecting observation data for updating a geographic database. An initial observation is collected by a first mobile device, a first vehicle, or a first sensor. Along with the geographic position, data indicative of the first observation is send to a server. The central server may analyze of the initial observation data to determine if additional observations should be made and define a bounding box from the geographic position of the first mobile device and the analysis of the initial observation data. A request for additional observations is generated and sent to at least one second mobile device, second vehicle, or second sensor based on the bounding box.

A vehicle control apparatus comprising an actuator used for traveling, an output device outputting an information, and a microprocessor. The microprocessor is configured to perform capturing an image ahead of the vehicle, recognizing a road sign included in an imaging range of the image captured in the capturing, determining, when the road sign is recognized in the recognizing while the vehicle is traveling on a merging lane merging into a main lane, whether or not the road sign is only applicable to the vehicle traveling on the merging lane based on a road curvature at a recognition position of the road sign, and controlling at least one of the output devices and the actuator based on an information of the road sign when it is determined in the determining that the road sign is only applicable to the vehicle traveling on the merging lane.

A traffic hand signal detection system includes: an imaging unit configured to acquire a photographed image from a camera photographing a predetermined range; an image classifier configured to classify an arm motion from the photographed image provided from the imaging unit by imparting a class; a detection module configured to detect the arm motion from the photographed image classified by the image classifier and generate a traffic hand signal sequence converted into a number; and an analysis module configured to receive the traffic hand signal sequence converted into the number from the detection module and determine a type of a traffic hand signal.

The present application discloses a method and apparatus for constructing a testing scenario for a driverless vehicle. An embodiment of the method includes: acquiring a traffic image including a scenario object, the scenario object comprising a roadway object, a traffic sign object, a vehicle object, or a pedestrian object; acquiring attribute information of the scenario object based on the traffic image; and constructing the testing scenario for the driverless vehicle based on the scenario object and the attribute information. The present application enables constructing the testing scenario for the driverless vehicle with real attributes of a roadway object, a traffic sign object, a vehicle object, or a pedestrian object. Therefore, a real traffic condition is restored and employed as the testing environment for the driverless vehicle, and the accuracy of the driverless vehicle testing is improved.

A computer-implemented method for recognising a printed character string is provided. The method includes receiving an image comprising the character string, the character string comprising a plurality of characters, determining a readability quality for each character in the character string and selecting at least one anchor character based at least in part on the readability quality of the characters in the character string. The identity of the at least one anchor character is determined using a character recognition algorithm and the identity of the character string recognised based on the at least one identified anchor character.

Apparatuses, methods and storage medium associated with traffic sign recognition, are disclosed herein. In some embodiments, an apparatus includes an orchestrator, disposed in a CA/AD vehicle, to receive a classification and a location of a traffic sign, while the CA/AD vehicle is enroute to a destination. In response, the orchestrator query a remote sign locator service or a local database on the CA/AD vehicle for a reference description of the traffic sign, determine whether the classification is correct, and output a result of the determination. The classification of the traffic sign is generated based at least in part on computer vision, and the orchestrator includes an anomaly detector to detect anomalies between the classification and the reference description, and determine whether the classification is correct based at least in part on an amount of anomalies detected. Other embodiments are also described and claimed.

A method ensures that a vehicle can safely pass a traffic light. The vehicle includes a processor and a light sensor. The method includes receiving traffic data, establishing a speed profile, establishing a control distance, the processor establishing, in accordance with the speed profile, a control distance at which braking ensures that the vehicle stops safely before the position of the traffic light, adjustment according to the speed profile, detecting the state of the traffic light when the vehicle is at the control distance from the traffic light, and the processor activating braking if the traffic light is red.