Vehicle Sensor Field Modeling to Exclude Self-Returns
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous vehicles face challenges in accurately distinguishing between sensor data from the vehicle itself and external objects due to obstructions, leading to false positive signals that can impact driving decisions.
Innovation Solution
A refined 3D mapping of the vehicle is developed using sensor data collected under various conditions, creating a model that filters out self-returns and accounts for different configurations and conditions, allowing for accurate sensor data interpretation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a coarse bounding box model is used to represent the vehicle, then the model creation process is simple and fast, but the model cannot accurately account for different vehicle configurations and active operating conditions, leading to false positive signals
Solution Approach 1:
The patent segments the vehicle model creation process into multiple stages: collecting sensor data under various operating conditions, processing the data to identify vehicle boundaries, and creating a refined 3D mesh model that captures different configurations. This segmentation allows the system to achieve high measurement precision while managing the complexity through structured data collection and processing steps.
Solution Approach 2:
The patent performs preliminary data collection under various operating conditions before final model creation. By pre-collecting sensor data across different vehicle configurations and environmental conditions, the system prepares comprehensive information that enables accurate model generation without requiring complex real-time processing during operation.
2Quantity of substance
If sensor data is collected without filtering, then all potential signals are captured, but false positive signals from the vehicle itself and transient obstructions interfere with accurate object detection
Solution Approach 1:
The patent extracts and removes false positive signals from the sensor data by comparing detected objects against the refined 3D vehicle model. The system identifies signals that originate from the vehicle itself or transient obstructions and filters them out, keeping only the relevant external object signals for accurate detection and decision-making.
Solution Approach 2:
The refined 3D vehicle model serves as an intermediary between raw sensor data and object detection decisions. This model acts as a reference framework that mediates the interpretation of sensor signals, helping to distinguish between signals from external objects and those from the vehicle itself or transient obstructions.
3Measurement precision
If a refined 3D mapping model is created to eliminate false positives, then sensor data accuracy is improved, but the data processing complexity and computational requirements increase
Solution Approach 1:
The patent creates a dynamic refined 3D vehicle model that adapts to different operating conditions and vehicle configurations. Rather than using a static coarse model, the system generates and updates the 3D mesh model based on collected sensor data, allowing the model to reflect the actual vehicle state and improve detection accuracy while managing complexity through conditional model generation.
Data Source
Figure 1
Figure 2
Figure 3A
AI summary
The technology relates to developing a highly accurate understanding of a vehicle's sensor fields of view in relation to the vehicle itself. A training phase is employed to gather sensor data in various situations and scenarios, and a modeling phase takes such information and identifies self- returns and other signals (730) that should either be excluded from analysis during real-time driving or accounted for to avoid false positives (Figs. 8A-B). The result is a sensor field of view model for a particular vehicle (Fig. 6B), which can be extended to other similar makes and models of that vehicle. This approach enables a vehicle to determine when sensor data is of the vehicle or something else. As a result, the detailed modeling allowing the on-board computing system to make driving decisions and take other actions based on accurate sensor information.