Autonomous Intersection Navigation via Sensor Obstruction Profiling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face difficulties navigating intersections due to sensor obstructions like vegetation, buildings, and road signs, as map data lacks information about these obstructions, leading to forced handovers to human drivers.
Innovation Solution
A system and method that use processors to detect moving objects on adjacent roadways, estimate profiles of sensor obstructions in non-roadway regions, combine these estimates to determine if autonomous navigation is feasible, and notify drivers accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If autonomous vehicles rely on conventional map data for navigation, then the system is simple and easy to operate, but the vehicle cannot detect sensor obstructions in non-roadway regions, leading to forced handovers to human drivers at intersections
Solution Approach 1:
The system performs preliminary characterization of intersections by collecting sensor data during layer drives before autonomous navigation is attempted. Map data is pre-updated with sensor obstruction information from multiple vehicle encounters, allowing the vehicle to assess navigation feasibility in advance without real-time handover
Solution Approach 2:
The patent introduces map data as an intermediary that carries sensor obstruction information between vehicles and the autonomous navigation system. The map serves as a shared database that stores composite estimates of sensor obstruction profiles, enabling vehicles to query and use this information for navigation decisions without direct real-time communication
2Measurement precision
If the vehicle collects sensor obstruction information during layer drives, then navigation accuracy improves, but the system complexity and processing requirements increase
Solution Approach 1:
The system merges sensor data from multiple layer drives to create composite estimates of sensor obstruction profiles. By combining information from multiple encounters with the same intersection, the system achieves more accurate and reliable obstruction characterization while distributing the processing load across multiple data collection events
Solution Approach 2:
The system creates simplified representations (copies) of sensor obstructions in the form of profile data stored in map data. Instead of processing complex raw sensor data in real-time, the system uses pre-computed profile copies that capture the essential obstruction characteristics, reducing computational complexity during autonomous navigation
3Reliability
If the system requires multiple layer drives to characterize intersections, then the accuracy of obstruction detection improves, but the time required for reliable navigation assessment increases
Solution Approach 1:
The system performs intersection characterization in advance during layer drives, storing the results in map data before they are needed for autonomous navigation. This preliminary action allows the vehicle to quickly query pre-computed obstruction profiles without requiring real-time data collection, reducing the time loss for navigation assessment
Data Source
AI summary
Systems, methods, and other embodiments for determining whether a vehicle is capable of navigating an intersection in an autonomous driving mode are disclosed. One embodiment detects that a vehicle traveling on a first roadway is approaching an intersection of the first roadway with a second roadway; detects, during one or more layer drives, one or more moving objects traveling on the second roadway; estimates the profile of each of one or more sensor obstructions situated in a non-roadway region abutting the intersection based on the detected one or more moving objects; combines, for each sensor obstruction, the profile estimates obtained from one or more layer drives to produce a composite estimate of the profile of that sensor obstruction; and determines, based at least in part on one or more composite profile estimates, whether navigating the intersection in an autonomous driving mode of the vehicle is achievable.