Automatic Smart UAV search of lost floating target in ocean environment based on high dense clustering

• Published in: 2020 First International Conference of Smart Systems and Emerging Technologies (SMARTTECH) pp. 116 - 121
• Main Authors: Boulares, Mehrez, Barnawi, Ahmed
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01-11-2020
• Subjects: Delays; Dynamic target prediction; High dense clustering; History; Muti UAV; Sea surface; Surface Ekman current; Trajectory; Unmanned aerial vehicles
• DOI: 10.1109/SMART-TECH49988.2020.00039

In this paper, we present a new rescuing mission based on non-stationary lost target searching methodology exploiting the high dense zone extraction. Our searching method is based on the determination of the critical searching zones (by density) relying on clustering techniques applied on particle trajectories history. Starting from the extracted critical high dense zones, we defined a circular searching location around each critical zone to be used by Multi Unmanned Aerial Vehicle (UAV) for a parallel searching mission. We conducted an experimental study to evaluate our searching methodology by using MeanShift and DBscan clustering techniques covering 68517km 2 surface area. We concluded that Meanshift performs better than DBscan by obtaining 51 % (of 100 different searching mission) are less or equal to 10 KM as closest UAV distance to the lost target with an average of 70 minutes (characterizing the searching delay) which could be considered as a good result.