Energy Optimization and Trajectory Planning for Constrained Multi-UAV Data Collection in WSNs

Abstract

Wireless sensor networks (WSNs) deployed in remote areas face a challenge in uploading the collected data to data centers due to limited network coverage. Unmanned aerial vehicles (UAVs) can extend network coverage to remote WSNs by flying and communicating with WSN aggregator nodes to collect data. UAV-assisted data collection systems need to be carefully developed to collect all data efficiently while considering the UAV and WSN constraints. This paper provides an energy-efficient multi-UAV data collection framework for WSNs. We formulate the data collection system as a problem that jointly optimizes the system cost and energy consumption constrained by the communication power, UAV mission time, and memory size. The problem is resolved over two steps: First, the location and number of aggregators needed are determined using a triangulation-based K-means clustering that minimizes the number of aggregators used and the system cost. Second, the dockstation position that minimizes the energy consumption is obtained using the gaining-sharing knowledge (GSK) optimization algorithm. The optimum UAV trajectory for each GSK candidate solution is designed by solving a capacitated vehicle routing problem (CVRP) that combines heuristic and metaheuristic solving techniques. Simulations show that our framework outperforms other recent techniques by minimizing the overall system cost and energy consumption.

Keywords

• Unmanned aerial vehicles
• wireless sensor networks
• clustering
• trajectory planning
• energy optimization
• data collection