Occupancy Grid Mapping for Shadowed Drivable Space Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous driving systems face challenges in accurately detecting drivable space due to limitations in sensors such as GNSS/IMU, cameras, and LIDAR scanners, which can be affected by environmental conditions and shadows from stationary objects, leading to errors in occupancy mapping.
Innovation Solution
A device that receives point data from a scanner to determine occupancy probabilities within an occupancy grid, accounting for shadow effects by maintaining the occupancy probability of cells within the grid, thereby improving the accuracy of drivable space detection and reducing latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If occupancy probability is reduced based on new sensor frames to update occupancy mapping, then the occupancy map reflects current sensor data, but cells in shadowed areas may be incorrectly marked as unoccupied due to sensor limitations
Solution Approach 1:
The system performs preliminary actions by detecting occupied cells before processing shadowed cells. When a cell is determined to be occupied, the system proactively identifies and protects shadowed cells behind it from being incorrectly marked as unoccupied. This preliminary protection prevents the harmful effect of shadow-induced detection errors before they can occur.
Solution Approach 2:
The system applies preliminary anti-action by preventing the reduction of occupancy probability in shadowed cells before the normal update process could incorrectly reduce it. By identifying shadowed cells in advance and applying a protective measure (preventing probability reduction), the system counteracts the potential harmful effect of shadow-induced detection errors.
2Loss of time
If the system processes all cells in each sensor frame to maintain real-time occupancy mapping, then the occupancy map remains current, but processing time increases causing latency
Solution Approach 1:
The system extracts and processes only the relevant subset of cells that are actually occupied or potentially affected by shadows, rather than processing all cells in the occupancy grid. This selective extraction of necessary cells reduces processing time and latency while maintaining detection accuracy for the affected areas.
Solution Approach 2:
The system applies partial action by processing only the necessary portion of cells (occupied cells and their shadowed cells) rather than performing excessive processing on all cells in the grid. This partial processing approach reduces computational overhead and latency while maintaining sufficient accuracy for autonomous navigation decisions.
3Productivity
If the scanner continuously scans the environment to detect changes in occupancy, then real-time detection is achieved, but false negatives occur in shadowed areas where the scanner cannot detect objects
Solution Approach 1:
The system uses occupied cells as intermediaries to indirectly detect objects in shadowed areas. Instead of relying on the scanner to directly detect objects in shadowed regions (which it cannot do), the system uses the detected occupied cells in front of the shadows as intermediaries to infer and protect the shadowed regions, thereby maintaining detection accuracy without requiring direct scanner coverage.
Data Source
AI summary
In some aspects, a device may receive point data associated with a cell of an occupancy grid for controlling a vehicle. The device may determine, based on the point data, a characteristic of the cell that is associated with an occupancy probability of the cell, wherein the occupancy probability is determined according to a first technique based on the point data. The device may configure, based on the characteristic, the occupancy probability for the cell, within the occupancy grid, according to a second technique. Numerous other aspects are described.