Hybrid Evolutionary Approach to Multi-objective Path Planning for UAVs

Abstract

The goal of Multi-Objective Path Planning (MOPP) is to find Pareto-optimal paths for autonomous agents with respect to several optimization goals like minimizing risk, path length, travel time, or energy consumption. In this work, we formulate a MOPP for Unmanned Aerial Vehicles (UAVs). We utilize a path representation based on Non-Uniform Rational B-Splines (NURBS) and propose a hybrid evolutionary approach combining an Evolution Strategy (ES) with the exact Dijkstra algorithm. Moreover, we compare our approach in a statistical analysis to state-of-the-art exact (Dijkstra's algorithm), gradient-based (L-BFGS-B), and evolutionary (NSGA-II) algorithms with respect to calculation time and quality features of the obtained Pareto fronts indicating convergence and diversity of the solutions. We evaluate the methods on a realistic 2D urban path planning scenario based on real-world data exported from OpenStreetMap. The examination's results indicate that our approach is able to find significantly better solutions for the formulated problem than standard Evolutionary Algorithms (EAs). Moreover, the proposed method is able to obtain more diverse sets of trade-off solutions for different objectives than the standard exact approaches. Thus, the method combines the strengths of both approaches.

Multi-objective 3D Path Planning for UAVs in Large-Scale Urban Scenarios

“…Different Pareto solutions are achieved by varying the saturation weight for the input map. However, changing weights might in general not always lead to solely non-dominated solutions in the objective space, as shown in a recent study [7]. Moreover, the path is only two-dimensional.…”

“…Like in a recent study [7], we use NURBS curves as a path representation. Here, we adapt them to represent 3D paths by adding a third component to the control point vectors.…”

“…1) Noise: Similar to a recent study [7], this objective is designed to minimize noise pollution in the city. Torija et al [5] show that people perceive the same level of drone noise less loud and annoying in environments with higher road traffic noise.…”

“…On the downside, EAs need high computational resources and can not guarantee optimality. In a recent study [7], it was shown that the performance of EAs in two-dimensional MOPP problems can be increased significantly by an advanced initialization step. However, in order to obtain practically feasible paths for aerial vehicles while making use of the available physical degrees of freedom, it is necessary to search paths in the 3D space.…”

“…The second part contains the main optimization step. As main optimizers, we use a basic Evolution Strategy (ES) [7], the state-of-the-art multiobjective Non-dominated Sorting Genetic Algorithm (NSGA2) [9] as well as the state-of-the-art Multi-Objective Covariance Matrix Adaptation Evolution Strategy (MO-CMA-ES) [10]. For the sake of completeness, we also include the gradientbased Limited-memory Broyden-Fletcher-Goldfarb-Shanno algorithm with Bounds (L-BFGS-B) [11] in our comparison.…”