UAV Obstacle Detection Using Segment-of-Interest Sensing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current obstacle detection systems for drones or unmanned aerial vehicles (UAVs) face challenges in efficiently detecting and reacting to obstacles due to limited field of view and resource constraints, particularly when navigating complex three-dimensional environments.
Innovation Solution
An apparatus and method that utilize telemetry and motion vector data to determine a segment of interest, process distance data to identify obstacles, and provide imaging data analysis using neural networks to estimate obstacle dimensions and determine evasive paths, while optimizing data processing by discarding unnecessary data and verifying dimensions using external distance data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the field of view of distance sensors is limited to reduce data processing load, then computational resources are saved, but obstacle detection coverage is reduced
Solution Approach 1:
The system segments the detection space into a 'segment of interest' based on projected contour dimensions and motion vectors, focusing computational resources only on relevant areas. This allows the drone to process distance data selectively rather than analyzing all sensor data, improving processing efficiency while maintaining detection reliability for critical obstacles.
Solution Approach 2:
The system performs preliminary projection of the drone's contour dimensions over motion vectors to pre-identify the segment of interest before full obstacle analysis. This preliminary action filters the detection space in advance, allowing the system to focus subsequent detailed analysis only on pre-identified relevant segments, thereby saving computational resources while ensuring obstacles in critical paths are detected.
2Productivity
If imaging data analysis is performed only when obstacles are detected in the segment of interest, then computational resources are optimized, but detection response time may increase
Solution Approach 1:
The system implements periodic action by continuously monitoring distance data in the segment of interest and triggering imaging data analysis only when obstacles are detected. This conditional periodic approach allows the system to maintain readiness for critical events while avoiding continuous heavy processing, optimizing computational resources without significantly impacting response time for actual obstacles.
Solution Approach 2:
The system uses the distance sensor data itself to trigger the more resource-intensive imaging analysis only when necessary. The distance data serves a dual purpose: both as the primary detection mechanism and as a trigger for secondary analysis, eliminating the need for continuous imaging processing and optimizing resource allocation based on actual detection needs.
3Productivity
If distance data outside the segment of interest is discarded to reduce processing load, then data processing efficiency improves, but verification of obstacle dimensions becomes less accurate
Solution Approach 1:
The system extracts and retains specific distance data points that fall within the projected contour dimensions, discarding data outside this segment. This selective extraction maintains processing efficiency by focusing only on relevant data while preserving the precision needed for obstacle dimension verification within the critical detection zone.
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
Enhances the ability of UAVs to detect and avoid obstacles effectively, even with limited field of view distance sensors, by reducing computational resources and ensuring accurate obstacle dimension estimation, thereby improving navigation and safety in complex environments.
Implementation Method 1
The said distance sensor may comprise a LIDAR sensor
Data Source
AI summary
An apparatus, method and computer program is described comprising the following: receiving telemetry and/or motion vector data and distance data from a distance sensor of the drone or unmanned aerial vehicle (32); determining a segment of interest dependent on said telemetry data and/or motion vector data; processing said distance data to determine whether an obstacle falls within said segment of interest (34); receiving imaging data; and providing imaging data analysis in the event that an obstacle is determined to fall within said segment of interest (36).