Autor: Pedro Miranda Pinheiro (Currículo Lattes)
Resumo
This work addresses the trajectory planning of an unmanned hybrid aerial-underwater vehicle called Hydrone, which is being developed by the Intelligent Robotics and Automation Group (NAUTEC) of the Federal University of Rio Grande (FURG). This work is based on a classical technique for trajectory planning of autonomous vehicles, Rapidly-exploring Random Trees (RRT). In addition, two heuristics are proposed to increase the energy efficiency of the vehicle. The first of these heuristics consists of biasing the tree expansion towards the domain with the lowest navigation cost, and the second assigns estimated costs to nodes in the tree and chooses the cheapest paths. This work demonstrates the main characteristics and application potentials of these methods for hybrid vehicle navigation. To evaluate these techniques, simulated experiments were performed in 135 different scenarios using each of the three compared algorithms, the algorithm without heuristics and the two algorithms with the proposed heuristics. A comparative analysis of their performances is also presented. It was identified in these results, that the use of heuristics can contribute significantly to the reduction of energy consumption, and even increase the average speed in the missions performed by these vehicles. Among the main results, it was identified that in 95% of the scenarios, one of the algorithms using the proposed heuristics was responsible for the lowest energy consumption. Finally, in the last part of this work, the findings are summarized and opportunities for future work are identified.