Optimized path planning of a swarm of drones for massive data collection

This Master project is part of laboratory research consisting in optimizing path planning of a swarm of drones for massive traffic data collection. The main goals aim to determine the optimal number of drones to use given the available fleet and to generate the optimal path and initial position for each vehicle. Two methods have been implemented to solve the problem: Mixed-Integer Linear Programming and the Column Generation Method. Different road networks are considered to test the algorithm robustness and adaptability. Moreover, an intuition and a Mathematical model have also been proposed to consider collision and hovering. The results suggest that a suitable solution is obtained with the MILP but the running time is problematic. The Column Generation method performs significantly faster. Accurate results are obtained for a small scaled network. Simulations are also performed on a part of San Francisco road network and valid, yet less accurate, results are obtained. Finally, several areas of improvement are suggested to strengthen and complete this research.