LiDAR Self-Hit Data Separation Using Geometric Vehicle Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles (AVs) face challenges in distinguishing self-hit data from environmental data collected by LiDAR sensors, which can lead to confusion in object identification and navigation, as self-hit data is recorded on the vehicle surface rather than the surrounding environment.
Innovation Solution
The use of geometric models, such as CAD models, to approximate vehicle boundaries and create masks for processing sensor data, allowing for the separation of self-hit data from environmental data by comparing collected LiDAR data to a spherical or Cartesian coordinate system-based model, thereby filtering out self-hit points.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If LiDAR sensor collects data from all directions including vehicle surface, then comprehensive environmental data is obtained, but self-hit data contaminates the sensor data causing confusion in object identification
Solution Approach 1:
The patent segments sensor data into two distinct categories: self-hit data (reflections from vehicle surface) and environmental data (reflections from external objects). By creating separate processing pathways for each type, the system maintains comprehensive data collection while ensuring reliable object identification through dedicated processing logic for each data category.
Solution Approach 2:
The patent introduces an intermediary classification mechanism that acts as a mediator between raw sensor data and object identification processing. This intermediary layer analyzes sensor returns to determine whether they originate from self-hit or environmental sources, thereby preventing contamination of object identification data while preserving comprehensive environmental coverage.
2Measurement precision
If geometric models are used to filter self-hit data, then object identification accuracy improves, but data processing complexity increases
Solution Approach 1:
The patent creates a digital geometric model (copy) of the vehicle's physical structure to represent its spatial boundaries. This virtual model serves as a reference for identifying self-hit data without requiring complex physical modifications to the vehicle or sensor system, thereby achieving precise data separation through computational rather than mechanical means.
Solution Approach 2:
The patent replaces complex mechanical or hardware-based filtering systems with a computational approach using geometric models and coordinate system transformations. By substituting physical filtering mechanisms with algorithmic processing, the system achieves accurate self-hit data separation while minimizing additional hardware complexity.
Data Source
AI summary
The subject disclosure relates to ways to identify self-hit data collected by autonomous vehicle (AV) sensors. In some aspects, a method of the disclosed technology includes steps for generating a geometric model of an autonomous vehicle (AV), wherein the geometric model specifies physical boundaries of the AV in three-dimensional (3D) space, collecting sensor data for an environment around the AV, and identifying one or more data points, from among the collected sensor data, that correspond with a surface of the AV. Systems and machine-readable media are also provided.