Occupancy Grid Hazard Detection for Multi-Sensor Depth Uncertainty
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional hazard detection systems for autonomous and semi-autonomous vehicles face inaccuracies in determining the location and distance of hazards due to the 2D nature of camera systems and the sparsity of LiDAR and RADAR data, leading to inefficient and unreliable hazard detection.
Innovation Solution
The use of occupancy grids to plot hazard indicators from multiple sensors, combining confidence levels to accurately determine the presence of hazards, leveraging the strengths of different sensor modalities while accounting for their weaknesses, and using ego-motion to interpolate or extrapolate sensor data over time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If camera detection systems are used for hazard detection, then the system can detect hazards in the driving environment, but the depth estimate accuracy is low leading to imprecise location determination
Solution Approach 1:
The patent combines multiple sensor modalities (camera, LiDAR, RADAR) to create a unified hazard detection system. The camera provides comprehensive hazard detection coverage while LiDAR and RADAR contribute depth and distance information, resolving the contradiction between detection reliability and measurement precision by merging complementary sensor strengths.
Solution Approach 2:
The patent transitions from 2D camera images to 3D world space representation by integrating depth information from LiDAR and RADAR sensors. This dimensional transformation enables accurate depth estimation and precise hazard location determination while maintaining the comprehensive detection capability of the camera system.
2Measurement precision
If LiDAR systems are used for depth estimation, then more accurate depth estimates are provided, but the sample density is sparse for small objects making size and shape determination insufficient
Solution Approach 1:
The patent merges LiDAR depth information with camera image data to compensate for sparse sampling. The camera provides dense visual information about object size and shape, while LiDAR provides accurate depth estimates, creating a complementary relationship that resolves both the precision and information loss issues.
Solution Approach 2:
The patent creates a multi-functional sensor system where each sensor type serves multiple purposes. The camera not only detects hazards but also provides object size and shape information, while LiDAR not only provides depth estimates but also contributes to hazard detection. This universal approach ensures that no single sensor modality is overwhelmed by its limitations.
3Measurement precision
If RADAR systems are used for hazard detection, then depth information is provided, but the sample density is insufficient for determining hazard size, shape, and classification
Solution Approach 1:
The patent combines RADAR depth information with camera visual information to achieve both accurate depth estimation and comprehensive hazard classification. The camera system provides detailed visual data for hazard size, shape, and classification, while RADAR contributes reliable depth information, resolving the contradiction through sensor fusion.
Data Source
AI summary
In various examples, a hazard detection system plots hazard indicators from multiple detection sensors to grid cells of an occupancy grid corresponding to a driving environment. For example, as the ego-machine travels along a roadway, one or more sensors of the ego-machine may capture sensor data representing the driving environment. A system of the ego-machine may then analyze the sensor data to determine the existence and/or location of the one or more hazards within an occupancy grid—and thus within the environment. When a hazard is detected using a respective sensor, the system may plot an indicator of the hazard to one or more grid cells that correspond to the detected location of the hazard. Based, at least in part, on a fused or combined confidence of the hazard indicators for each grid cell, the system may predict whether the corresponding grid cell is occupied by a hazard.