Evaluation of control strategies for fixed-wing drones following slow-moving ground agents

There are many situations where fixed-wing drones may be required to track ground moving agents, such as humans or cars, which are typically slower than drones. Some control strategies have been proposed and validated in simulations using the average distance between the target and the drone as a performance metric. However, besides the distance metric, energy expenditure of the flight also plays an important role in assessing the overall performance of the flight. In this paper, we propose a new methodology that introduces a new metric (energy expenditure), we compare existing methods on a large set of target motion patterns and present a comparison between the simulation and field experiments on proposed target motion patterns. Using this new methodology we examine the performance of three control strategies: the Lyapunov Guidance Vector Field strategy, the Bearing-only strategy and the Oscillatory strategy. Among the three strategies considered, we demonstrate that the Lyapunov Guidance Vector Field strategy has the best performance for all target motion patterns. Field experiments with fixed-wing drones provide additional insights into the benefits and shortcomings of each strategy in practice. We compare target tracking strategies on a large set of target motion patterns.We propose a new methodology that introduces energy expenditure as a new metric.Optimal parameters are set in each test and Pareto fronts are used for comparison.LGVF strategy shows superior performance both in simulation and field experiments.