Intelligent Drones Trajectory Generation for Mapping Weed Infested Regions Over 6G Networks

Unmanned Aerial Vehicles (UAVs), in conjunction with 6G, are a potential tool for monitoring agricultural lands and various agricultural applications. There have been several trajectory generation algorithms proposed for surveying agricultural land. However, in most situations, the accuracy of the trajectory is hampered by several factors, namely the complex geographical topography, performance, connectivity with the Ground Control Station (GCS) and positioning error of the UAV during flight, amongst others. Therefore, in this paper, we propose Drones Trajectory Generation employing an improved Genetic Algorithm (GA) and Non-Uniform Rational P-Splines (NURPS) based optimizer (DTG-GN). The improved GA utilizes a novel dual fitness function parameter to select an optimal path to map the weed-infested regions. The path chosen is often impeded by the high number of sudden turns, affecting the UAV's speed profile and the path's continuity. Therefore in the NURPS optimization, the computation of the intermediate knots vector between the control points improves the smoothness of the path. Furthermore, the accuracy of detecting the weed-infested area and the path length are employed to optimize the path. Besides, a 6G network is utilized for communicating the path between the GCS and the UAV to ensure seamless connectivity. Thus the time taken by DTG-GN to generate the optimal trajectory reduces by 38.815% for 50 control points. DTG-GN also reduces the average trajectory length by 45.67% for 50 control points, establishing its supremacy over conventional trajectory generation algorithms.