LiDAR Based Detect and Avoid System for UAV Navigation in UAM Corridors
Система обнаружения и предотвращения столкновений на основе LiDAR для навигации БПЛА в коридорах городской воздушной мобильности
2022-07-22
SCID: 54.1/75rh2jxa
Discuss with AI
LiDAR-based detect and avoidSecond Order Cone ProgramUAV navigationUrban Air Mobility corridorspoint-cloud object detection
Figures from the paper
Abstract (AI)
In this work, a Detect and Avoid system is presented for the autonomous navigation of Unmanned Aerial Vehicles (UAVs) in Urban Air Mobility (UAM) applications. The current implementation is designed for the operation of multirotor UAVs in UAM corridors. During the operations, unauthorized flying objects may penetrate the corridor airspace posing a risk to the aircraft. In this article, the feasibility of using a solid-state LiDAR (Light Detecting and Ranging) sensor for detecting and positioning these objects was evaluated. For that purpose, a commercial model was simulated using the specifications of the manufacturer along with empirical measurements to determine the scanning pattern of the device. With the point clouds generated by the sensor, the system detects the presence of intruders and estimates their motion to finally compute avoidance trajectories using a Second Order Cone Program (SOCP) in real time. The method was tested in different scenarios, offering robust results. Execution times were of the order of 50 milliseconds, allowing the implementation in real time on modern onboard computers.
Key Findings
1
A LiDAR-based Detect and Avoid system was developed for autonomous multirotor UAV navigation in Urban Air Mobility corridors.
2
Testing across different scenarios produced robust results, with execution times around 50 milliseconds on modern onboard computers.
3
The proposed processing speed supports real-time onboard implementation for UAV corridor operations.
4
The study evaluated the feasibility of a commercial solid-state LiDAR for detecting and positioning unauthorized airborne intruders.
5
The system uses LiDAR point clouds to detect intruders, estimate their motion, and compute collision-avoidance trajectories with a real-time Second Order Cone Program.
Research Object
multirotor UAVs operating in Urban Air Mobility corridors with unauthorized flying objects intruding into the corridor airspace
Research Subject
LiDAR-based detection, motion estimation, and real-time collision-avoidance trajectory generation for intruding flying objects
Publication Details
Publication Date
2022-07-22
Journal
Publisher
ISSN
Cited by
77
Open access PDF
Access Type
Author Information
Download PDF
Subscribe to digest