Roadside Sensor Fusion for Vehicle Perception Gap Coverage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current vehicle-to-everything (V2X) communication technologies face challenges in providing complete and accurate real-time mapping for autonomous driving due to occlusions and limited viewing capabilities of individual vehicles, leading to unreliable sensor data sharing and incomplete environmental sensing.
Innovation Solution
Infrastructure-based sensor networks that request and combine sensor data from both fixed infrastructure sensors and vehicles, using broadcast and multicast communication to minimize overhead and fill perception gaps, ensuring comprehensive and timely data for real-time mapping.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If individual vehicles use their own sensors for environmental sensing, then each vehicle can independently detect objects, but the sensing coverage is limited due to occlusions and viewing angles
Solution Approach 1:
The patent combines sensor data from multiple vehicles and infrastructure sensors into a unified environmental model. The server aggregates sensor data from different sources (vehicles and fixed infrastructure) to create a comprehensive real-time map that overcomes individual vehicle sensing limitations and occlusions.
Solution Approach 2:
The server acts as an intermediary that collects, processes, and distributes sensor data to vehicles. Instead of vehicles directly sharing data with each other, the server mediates the data fusion process, filling perception gaps by combining data from multiple sources and distributing the consolidated environmental information to all participants.
2Loss of information
If vehicles share sensor data continuously, then environmental awareness is improved, but communication overhead increases
Solution Approach 1:
The system extracts only the essential sensor data needed for environmental mapping and shares it through the server. Instead of continuous full-data transmission, vehicles send sensor data on demand or periodically to the server, which then distributes only the necessary environmental information to each vehicle, reducing redundant communication.
Solution Approach 2:
The patent implements periodic or event-triggered data sharing rather than continuous transmission. Vehicles communicate sensor data to the server at specific intervals or when significant environmental changes occur, and the server updates environmental models periodically, reducing communication frequency while maintaining environmental awareness.
Data Source
AI summary
Systems, methods, and computer-readable media are provided for wireless sensor networks (WSNs), including vehicle-based WSNs. A road side unit (RSU) includes one or more fixed sensors covering different sectors of a designated coverage area. The RSU uses the sensors to capture sensor data that is representative of objects in the coverage area, tracks objects (e.g., vehicles) in the coverage area, and determines regions in the coverage area that are not adequately covered by the sensors (e.g., “perception gaps”). When the RSU identifies an object that is in or at a perception gap, then the RSU sends a request to that object for sensor data captured by the object's on-board sensors. The RSU obtains the sensor data from the object, and uses the obtained sensor data to complement the knowledge that the RSU (i.e., “filling the perception gaps”). Other embodiments are disclosed and/or claimed.