Multi-Perspective Sensor Data Fusion for Automated Driving
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current automated driving systems primarily perceive scenes from a single perspective, leading to incomplete, inaccurate, or indeterminate scene information due to the inability to leverage additional viewpoints from other vehicles, resulting in poorer driving decisions and increased accidents.
Innovation Solution
A sensor system that generates and shares digital data among automated vehicles via wireless networks, including sensor measurements, object classifications, and confidence values, to reconcile discrepancies and create a combined heatmap, enhancing the perception capabilities of automated vehicles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If automated driving systems use single-perspective sensor data, then device complexity is reduced, but measurement precision and reliability deteriorate due to occlusions and limited field of view
Solution Approach 1:
The patent combines sensor data from multiple vehicles to create a fused perception model. Each vehicle's sensors detect objects from their respective perspectives, and the system merges these detections to overcome individual occlusions and field-of-view limitations, thereby improving measurement precision without requiring each individual vehicle to have complex multi-perspective sensor systems
Solution Approach 2:
The patent introduces communication infrastructure and data fusion servers as intermediaries. These intermediaries collect, process, and fuse sensor data from multiple vehicles, enabling each vehicle to benefit from multi-perspective information without directly implementing complex inter-vehicle sensor coordination systems
2Reliability
If automated driving systems integrate data from multiple vehicles, then reliability and scene information completeness improve, but loss of information increases due to data transmission and reconciliation challenges
Solution Approach 1:
The patent performs preliminary data processing and object detection at each vehicle before transmission. By pre-processing sensor data to extract relevant object information and confidence metrics, the system reduces data transmission requirements and minimizes information loss during communication, while still enabling reliable multi-vehicle fusion at the server level
3Productivity
If automated driving systems use single-point perspective, then ease of operation is maintained, but productivity and safety deteriorate due to increased accidents and collisions
Solution Approach 1:
The patent implements self-service through automated data fusion algorithms that automatically reconcile multiple vehicle perspectives without requiring manual intervention. The system autonomously processes sensor data from multiple sources, resolves conflicts, and generates unified perception results, maintaining operational simplicity while dramatically improving driving safety through enhanced scene awareness
Data Source
AI summary
The disclosure includes embodiments for a sensor system for multiple perspective sensor data sets. In some embodiments, a method executed by an ego vehicle includes receiving, from a set of remote vehicles, a set of sensor data describing a set of preliminary heatmaps for a roadway environment. The method includes reconciling discrepancies in the set of heatmaps and a preliminary heatmap of the ego vehicle to form a combined heatmap that describes the objects in the roadway environment as collectively observed by the onboard sensors of the set of remote vehicles and the ego vehicle. The method includes providing the combined heatmap to one or more of the remote vehicles included in the set of remote vehicles. The method includes modifying an operation of the ego vehicle based on the combined heatmap. For example, an operation of an autonomous driving system of the ego vehicle is modified.