Lidar Obstruction Detection Using Zero-Distance Pulse Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies face challenges in accurately detecting and mitigating obstructions on lidar sensors, which can degrade sensor data quality due to environmental factors and internal errors, impacting autonomous vehicle navigation and obstacle detection.
Innovation Solution
The implementation of an obstruction detection component that uses raw sensor data and machine learned models to detect and classify obstructions on lidar sensors, determining an obstruction score to control autonomous vehicle operations and initiate cleaning actions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional sensor data collection methods are used, then the system can operate with simple sensor deployment, but sensor data quality degrades due to undetected obstructions on the lidar sensor surface
Solution Approach 1:
The system performs preliminary obstruction detection by analyzing raw lidar data for pulses returning at substantially zero distance from the sensor. This preliminary check occurs before main processing, allowing the system to identify obstructions on the sensor surface proactively and mitigate data quality degradation before it affects navigation decisions.
Solution Approach 2:
The patent introduces an intermediary obstruction detection component that acts as a mediator between the lidar sensor and the main processing system. This component analyzes raw sensor data to detect obstructions and generates obstruction scores, which then inform subsequent processing decisions, thereby protecting the main system from degraded data quality without requiring complete system redesign.
2Measurement precision
If the system continuously monitors all sensor data for obstructions, then detection accuracy improves, but processing time and computational load increase
Solution Approach 1:
The system extracts only the critical portion of sensor data needed for obstruction detection - specifically raw lidar pulses and their distance measurements - separating this from the full sensor data stream. By taking out only the essential elements (pulse distance and intensity information), the system achieves accurate obstruction detection without the computational burden of processing complete point cloud data.
Solution Approach 2:
The patent applies partial action by monitoring only a subset of lidar returns - specifically those at substantially zero distance - rather than analyzing all sensor data. This selective approach provides sufficient detection accuracy for surface obstructions while significantly reducing processing time and computational resources required compared to comprehensive full-field analysis.
3Productivity
If the system relies on processed lidar data, then navigation decisions can be made quickly, but obstructions on the sensor surface cannot be detected
Solution Approach 1:
The system performs preliminary obstruction detection on raw lidar data before the data enters the main processing pipeline. By checking for obstructions in advance using simple distance thresholding on raw pulses, the system ensures reliable obstruction detection without delaying the main navigation processing, as the check occurs parallel to or before the primary data flow.
Solution Approach 2:
The patent segments the data processing into distinct stages: raw data collection, obstruction detection using zero-distance pulse analysis, and main navigation processing. This segmentation allows obstruction detection to occur independently on a separate data stream, enabling both rapid navigation decisions and reliable obstruction detection without one process blocking the other.
Data Source
AI summary
Techniques for detecting an obstruction associated with a lidar sensor are discussed herein. For example, a computing device can implement an obstruction detection component to detect rain, mud, dirt, dust, snow, ice, animal droppings, etc., on and/or near an outer surface of the lidar sensor. The obstruction detection component can apply one or more heuristics and/or models to the lidar data and/or compare pulse information associated with the lidar data to a threshold to determine a size, a type, or a location of an obstruction blocking a lidar beam.