A real-time algorithm for variable-objective motion planning over terrain and threats

Abstract

A major issue required to enhance the autonomy level of unmanned vehicles is real-time motion planning. In this context, optimal trajectories need to be generated online considering the vehicle’s dynamic potentials and constraints. However, autonomous air vehicles often need to plan and execute their missions with varying objectives that may even be dictated in flight. Therefore, the current study introduces and focuses on the new concept of variable-objective motion planning. In this regard, a new dynamic multi-objective heuristic optimization algorithm is developed for path and motion planning of autonomous air vehicles in presence of deterministic terrain obstacles as well as random threats. Effectiveness of the proposed algorithm is demonstrated via several simulated scenarios and verified against past research. The results are promising and indicative of a good potential of the proposed algorithm in producing optimal real-time optimal feasible trajectories with multi-variable objectives.