Two Algorithms Based on Iterative Optimization for UAV Path Planning

In this paper, we propose two algorithms based on iterative optimization for UAV path planning. In the first one, a nonlinear programming (NLP) problem of unmanned aerial vehicle (UAV) path planning is transformed into a general quadratically constrained quadratic programming (QCQP) problem by a numerical optimization method and solved through iteration. This method can converge to the optimal solution without giving the initial value. It overcomes the problems that NLP solvers are dependent on the initial guess, slow to converge, and sometimes can't converge to a feasible solution. In the second one, we improve the optimized rapidly exploring random tree algorithm (RRT*) and add UAV flight constraints to it. The improved RRT* algorithm reduces the length of the path and makes the path smoother.