Multi-Layer Road Grid Modeling for Sensor Fusion Uncertainty
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing road-perception systems are inadequate in accurately modeling roadways and quantifying uncertainty, often relying on inaccurate parametric methods and failing to fuse data from multiple sensors, which limits their reliability for autonomous and assisted driving systems.
Innovation Solution
A grid-based road model with multiple layers is developed, using data from various sources to generate a road model with layer hypotheses, mass values, belief parameters, and plausibility parameters, allowing for accurate representation and uncertainty quantification of roadway attributes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If parametric methods are used to represent lane boundaries and lane centers, then the road model can be generated with simpler computation, but the accuracy of the road model deteriorates
Solution Approach 1:
The patent divides the continuous roadway into discrete grid cells, where each cell independently represents a specific roadway attribute (lane boundary, lane center, barrier, etc.). This segmentation allows the system to avoid complex parametric fitting while achieving accurate representation of road geometry through the collective state of individual cells.
2Reliability
If data from multiple sensors is fused to build the road model, then the reliability of the road model improves, but the complexity of data processing increases
Solution Approach 1:
The patent applies segmentation by dividing the sensor data fusion problem into cell-level decisions. Each grid cell independently evaluates evidence from multiple sensors to determine its roadway attribute state, avoiding the need for complex global optimization while achieving reliable road model construction through distributed local decisions.
Solution Approach 2:
The patent introduces mass values as an intermediary representation that quantifies the strength of evidence from multiple sensors for each cell hypothesis. This intermediary allows conflicting sensor data to be systematically combined and evaluated, facilitating reliable road model construction while maintaining manageable processing complexity through a standardized evidence fusion framework.
3Adaptability or versatility
If multiple layers representing different roadway attributes are used, then the comprehensiveness of the road model improves, but the complexity of the model increases
Solution Approach 1:
The patent segments the road model into multiple independent layers, where each layer represents a specific roadway attribute (e.g., lane boundaries, lane centers, barriers). This segmentation allows the system to comprehensively represent different aspects of the roadway while managing complexity by treating each layer as an independent classification problem with its own grid cell states.
Solution Approach 2:
The patent adds a dimensional aspect to the road model by introducing multiple layers that represent different roadway attributes. This dimensional expansion enables the model to capture comprehensive roadway information (geometry, markings, obstacles) while maintaining manageable complexity through the grid-based cellular structure that organizes information systematically across layers.
4Reliability
If uncertainty quantification is implemented in the road model, then the safety of autonomous driving improves, but the computational overhead increases
Solution Approach 1:
The patent segments uncertainty quantification into cell-level operations, where each grid cell independently computes belief and plausibility parameters based on local sensor evidence. This segmentation enables safety-critical uncertainty assessment without requiring complex global computations, as each cell's uncertainty can be determined independently through straightforward evidence evaluation.
Solution Approach 2:
The patent implements self-service uncertainty quantification where each grid cell autonomously evaluates its own hypothesis confidence using locally available sensor data and mass values. This self-service approach enables safety improvement through uncertainty awareness while minimizing computational overhead by avoiding centralized uncertainty calculations.
Data Source
AI summary
This document describes techniques, apparatuses, and systems for a grid-based road model with multiple layers. An example road-perception system generates a roadway grid representation that includes multiple cells. The road-perception system uses data from multiple information sources to generate a road model that includes multiple layers. Each layer represents a roadway attribute of each cell in the grid and includes one or more layer hypotheses. In this way, the described techniques and systems can provide an accurate and reliable road model and quantify uncertainty therein.