Gaussian-mixture based potential field approach for UAV collision avoidance

This paper deals with a collision avoidance method for UAVs based on potential field. We derive potential vectors by using a Gaussian distribution function to design an avoidance trajectory and use a Gaussian Mixture Model (GMM) to represent two-dimensional complex shaped obstacles. In addition, we apply the Expectation-Maximization (EM) algorithm to modify the potential field in order to update the information using measured data on the obstacles. For this purpose, after briefly introducing the potential field, we construct a GMM with the EM algorithm and conduct guidance simulations using a point-mass UAV model in order to demonstrate the performance of the algorithm. Finally, we discuss the implications of the current approach and future research direction.