Dual-Return LiDAR Vapor Detection for False Object Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles often incorrectly detect vapor as objects due to opaque lidar data, leading to unnecessary stops or maneuvers, as conventional approaches fail to differentiate vapor from other objects using single return lidar data.
Innovation Solution
The use of dual return lidar data, including a strongest and farthest lidar return, to identify vapor in the driving environment, with a neural network determining object types and controlling vehicle systems to ignore vapor, allowing the vehicle to pass through without stopping.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional single return lidar data is used to detect objects, then the autonomous vehicle can detect objects in the driving environment, but the autonomous vehicle incorrectly identifies vapor as objects leading to unnecessary stops
Solution Approach 1:
The patent segments the lidar return data into multiple components (strongest return, farthest return, and intermediate returns) to analyze different reflection sources. By examining the distribution and characteristics of these segmented returns, the system can distinguish vapor reflections from actual object reflections, thereby improving detection accuracy and avoiding unnecessary stops
Solution Approach 2:
The patent introduces an additional dimension of analysis by comparing multiple lidar returns (not just the strongest one) along the range dimension. This multi-dimensional approach to analyzing return intensity distributions enables the system to identify vapor patterns that single-point detection cannot detect, resolving the contradiction between detection reliability and time loss
2Reliability
If the autonomous vehicle stops to avoid detected vapor, then the vehicle may avoid potential hazards, but the vehicle loses navigation efficiency and may require extended stopping periods
Solution Approach 1:
The patent implements feedback by continuously monitoring lidar return patterns and using neural networks to classify detected features as vapor or actual objects. This feedback mechanism enables real-time differentiation between safe vapor detections and hazardous objects, allowing the vehicle to maintain navigation efficiency while ensuring safety through accurate classification
3Difficulty of detecting and measuring
If the autonomous vehicle uses lidar data to detect all reflective surfaces, then comprehensive object detection is achieved, but vapor is incorrectly classified as solid objects requiring avoidance
Solution Approach 1:
The patent changes the parameters used for object classification by analyzing multiple characteristics of lidar returns including intensity distribution across multiple returns, range values, and reflection patterns. These parameter changes enable the neural network to distinguish vapor from solid objects with high precision while maintaining comprehensive detection of all reflective surfaces
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach effectively mitigates false object detection by vapor, enabling autonomous vehicles to navigate through steam and other vapors without unnecessary stops, improving navigation efficiency and reducing human intervention.
Implementation Method 1
Vapor can reflect a light beam emitted by the lidar sensor system, which can result in a lidar return having a relatively high intensity corresponding to a location of the vapor
Data Source
AI summary
Various technologies described herein pertain to detecting data in a lidar point cloud representative of vapor and controlling an autonomous vehicle based on such detection. The lidar point cloud is outputted by a lidar sensor system of the autonomous vehicle. The techniques set forth herein utilize dual return lidar data outputted by the lidar sensor system. The dual return lidar data includes two lidar returns received responsive to a light beam emitted (e.g., at a particular azimuthal angle) by the lidar sensor system into a driving environment of the autonomous vehicle. The data in the lidar point cloud detected as being caused by vapor can be removed from downstream processing; accordingly, the autonomous vehicle can be controlled such that the autonomous vehicle need not stop for or maneuver around vapor in the driving environment.