Great Wall Motor Company Limited

Aptiv Technologies Limited

Aptiv Technologies Ag

Kai Zhang

Michael H Laur

Krzysztof Kogut

Jakub Porebski

Aniello Sorrentino

Amith Somanath

Nanhu Chen

Aron Sommer

Mohamed A Moawad

Alexander Ioffe

Uri Iurgel

Vladimir Djapic

Ceyhan Karabulut

Damjan Karanovic

Duong Le

Zhichao Liu

James Nicholas Nickolaou

Walter Kirk Kosiak

Xiaohui Wang

This document describes methods and systems for vehicle localization based on radar detections. Radar localization starts with building a radar reference map. The radar reference map may be generated and updated using different techniques as described herein. Once a radar reference map is available, real-time localization may be achieved with inexpensive radar sensors and navigation systems. Using the techniques described in this document, the data from the radar sensors and the navigation systems may be processed to identify stationary localization objects, or landmarks, in the vicinity of the vehicle. Comparing the landmark data originating from the onboard sensors and systems of the vehicle with landmark data detailed in the radar reference map may generate an accurate pose of the vehicle in its environment. By using inexpensive radar systems and lower quality navigation systems, a highly accurate vehicle pose may be obtained in a cost-effective manner.

Vehicle localization based on radar detections

Methods and systems are described that enable radar reference map generation. A high-definition (HD) map is received and one or more HD map objects within the HD map are determined. Attributes of the respective HD map objects are determined, and, for each HD map object, one or more occupancy cells of a radar occupancy grid are indicated as occupied space based on the attributes of the respective HD map object. By doing so, a radar reference map may be generated without a vehicle traversing through an area corresponding to the radar reference map.

Radar reference map generation

Methods and systems are described that enable vehicle routing based on availability of radar-localization objects. A request to navigate to a destination is received, and at least two possible routes to the destination are determined. Availabilities of radar-localization objects for the possible routes are determined, and a route is selected based on the availabilities of the radar-localization objects. Furthermore, while traveling along a route, the vehicle is localized based on radar detections of radar-localization objects. A radar-localization quality of the localizing is monitored, and a determination is made that the radar-localization quality has dropped or will drop. Based on the radar-localization quality dropping, the route is modified and/or an operation of a radar module is adjusted. In this way, availabilities of radar-localization objects may be used to select an optimal route and to adjust a current navigation along a route to minimize driver takeover.

Vehicle routing based on availability of radar-localization objects

This document describes change detection criteria for updating sensor-based maps. Based on an indication that a registered object is detected near a vehicle, a processor determines differences between features of the registered object and features of a sensor-based reference map. A machine-learned model is trained using self-supervised learning to identify change detections from inputs. This model is executed to determine whether the differences satisfy change detection criteria for updating the sensor-based reference map. If the change detection criteria is satisfied, the processor causes the sensor-based reference map to be updated to reduce the differences, which enables the vehicle to safely operate in an autonomous mode using the updated reference map for navigating the vehicle in proximity to the coordinate location of the registered object. The map can be updated contemporaneously as changes occur in the environment and without hindering performance, thereby enabling real-time awareness to support controls and to improve driving-safety.

Change detection criteria for updating sensor-based reference maps

Methods and systems are described that enable radar detection prioritization based on downstream feedback. As a vehicle is traveling in a travel corridor, radar detections are received by a processor and feedback related to a downstream function is received from another processor that is executing the downstream function. For the radar detections that are determined to be within the travel corridor, respective priorities are assigned based on locations of the radar detections and the feedback related to the downstream function. The radar detections with the assigned priorities are then stored in a detection buffer along with indications of their respective assigned priorities. In this way, the detection buffer may be optimized for the downstream function by filtering out certain detections and prioritizing others. Consequently, computational loads on the downstream function are reduced while ensuring that important detections are received by the downstream function early.

Radar detection prioritization based on downstream feedback

This document describes techniques, apparatuses, and systems that can implement auction-based cooperative perception for autonomous and semi-autonomous driving systems. The described techniques, apparatuses, and systems cooperatively use perception systems of an autonomous or semi-autonomous vehicle (AV) in a fleet of connected AVs to provide perception data to the entire fleet. An AV of the fleet is selected to act as an auction system (e.g., an auctioneer) and sends an announcement offering perception tasks to the fleet for bidding. The AVs of the fleet determine whether they have the communication and computational capabilities to perform the tasks and, if so, submit bids to perform one or more tasks. The auctioneer awards the tasks, and the bid-winning AV(s) perform the tasks and update the fleet. In this way, the described techniques, apparatuses, and systems can provide perception services with increased coverage and quality, which can make autonomous and semi-autonomous driving systems safer.

Auction-based cooperative perception for autonomous and semi-autonomous driving systems

Disclosed herein are methods and systems for motion planning in an autonomous vehicle (AV) that separate path planning and velocity planning and may use reference lines in combination with motion planners to determine paths. The method may include reference lines to project planning data into a S-L coordinate system. A motion planner algorithm uses the reference lines and previous path planning history to generate a path in the S-L coordinate system. A velocity is determined for the path. An AV controller is updated with the path and the velocity. Motion planning computations use a dynamic vehicle look-up table to determine possible vehicle motions based on an initial state and control input.

Motion planning methods and systems for autonomous vehicle

Disclosed herein are methods and systems for efficient optimal control with dynamic modeling for an autonomous vehicle (AV). The method may include acquiring vehicle status information for the AV, determining a longitudinal velocity of the AV, determining a driving style factor, wherein the driving style factor is dependent on at least road scenarios, obtaining an optimal control factor from a look-up table (LUT) using the determined longitudinal velocity and the determined driving style factor and providing an updated control command (such as a steering command) based on the obtained optimal control factor. The driving style factor may be determined from at least vehicle status, desired trajectory, current linear velocity and like parameters and ranges between a gentle driving mode and an aggressive driving mode.

Efficient optimal control with dynamic model for autonomous vehicle

A localization system uses an auction-based cooperative localization method to determine local pose estimations or local pose state corrections. The auction-based cooperative localization method is triggered by receipt of relative position measurements by a member of a cooperative and is cooperatively processed by a bid winning member of the cooperative, where the cooperative refers to a collection of autonomous vehicles and passengers with location/position determinable devices that are in communication with each other for the purposes of sharing localization data, position data and the like. The bid winning member determines local pose estimations or local pose state corrections for itself and provides local pose estimations or local pose state corrections to the other members of the cooperative.

Growing community of inventors

Carmel, IN, United States of America

Kai Zhang