3D Trajectory Planning of Agricultural Unmanned Aerial Vehicles Based on GA–SA Algorithm

Abstract

The trajectory optimization problem in agricultural unmanned aerial vehicles (UAVs) inspection of agricultural facilities and equipment shares similarities with the traveling salesman problem. Traditional heuristic algorithms can encounter challenges such as overlapping flight paths and local optima due to complex terrain, numerous obstacles, and external environmental influences. To optimize the agricultural UAV detection process, a Genetic Algorithm and Simulated Annealing (GA–SA) algorithm based on GA was proposed to calculate 3D flight paths for agricultural UAVs. By establishing a mathematical model that combines the UAV threat environment with physical constraints, the cost function of the traditional SA algorithm was optimized to improve search efficiency. Additionally, trajectory smoothing techniques were applied to ensure smooth transitions between trajectory points. Experimental results demonstrate that the GA–SA algorithm overcomes the limitations of the traditional SA algorithm in node search efficiency and trajectory smoothing, enabling the planning of realistic and optimal flight paths in 3D environments. The GA–SA algorithm exhibits an 18.82% improvement in convergence time GA–SA compared with the traditional GA algorithm and a 13.20% improvement compared with the SA algorithm. Moreover, it shortens the optimal track distance by 2.244 km and 1.793 km, respectively, validating the effectiveness of the proposed method.