Topology Formation in the Internet of Flying Things
Project no. 32182 · Status: Kapanmış · 2024
Summary
<p>UAVs are used in applications such as agriculture, healthcare, military operations and search-rescue, and maintaining connectivity is vital for coordination. However, factors such as environmental conditions, limited batteries and drone losses can lead to drone failures, increasing the risk of network fragmentation. In order to increase resilience, k-connected networks, designed in such a way that a maximum of k-1 drone failures will not disrupt the connectivity, offer a solution providing multiple communication paths between drones. The k-connectivity restoration problem aims to provide k-connectivity by repositioning drones with minimum movement. In this project, solutions are proposed for this NP-Hard problem. Unlike existing k-connected restoration algorithms that are limited to moving drones to predefined points, our model allows drones to be positioned freely in the mission area; increasing flexibility while expanding the solution space and complexity. In the project, three center-based algorithms (ORIGIN, CENTROID and CENTFAR) are proposed that direct drones to different center points calculated from the network structure. In addition, an MST-based algorithm is presented that aims to provide and theoretically guarantee k-connectedness by moving drones along the MST. In addition to methods that check k-connectedness after each individual movement, group-based variants have also been developed, where all drones move simultaneously and k-connectedness is checked later. Simulations are performed with different numbers of drones, network sizes, k values, and transmission distances to evaluate the effectiveness and scalability of the proposed algorithms. The results show that in low k, high range, and small-scale scenarios, the MST algorithm minimizes the total movement and running time, while increasing the maximum movement; while CENTFAR performs better by reducing the total and maximum movement in larger, denser, and high k networks. Group-based strategies slightly increase the amount of movement, but reduce the running time.</p>