Evaluasi Kinerja Navigasi Inspeksi Melingkar Otonom Berbasis Finite State Machine pada UAV dengan Variasi Jarak di Webots

Abstract

Autonomous aerial inspection requires reliable navigation mechanisms to execute sequential flight operations while maintaining trajectory accuracy and returning to the initial position. This study evaluates a Finite State Machine (FSM)-based autonomous circular inspection navigation system implemented in the Webots simulation environment. The proposed navigation system consists of sequential states including takeoff, forward navigation, hovering, circular inspection, return navigation, and landing. The effect of inspection distance on navigation performance was investigated using five target-radius configurations of 3, 5, 7, 10, and 15 m, with three repetitions for each configuration. Performance was evaluated based on mission completion time, altitude root mean square error (RMSE), circular trajectory radius RMSE, yaw RMSE, and return-to-home (RTH) error. The results show that increasing the inspection radius leads to longer mission completion times, ranging from 88.61 s at 3 m to 350.54 s at 15 m. In contrast, altitude RMSE, radius RMSE, and yaw RMSE generally decrease as the inspection radius increases, with the lowest values obtained at 15 m of 0.1230 m, 0.1830 m, and 0.1806 rad, respectively. The RTH error exhibits a non-monotonic response, with the lowest value of 0.1615 m obtained at a 5 m inspection radius. These results indicate that inspection distance affects different aspects of autonomous navigation performance, particularly the trade-off between mission duration and trajectory-tracking accuracy. The findings demonstrate the feasibility of FSM-based autonomous circular inspection navigation in Webots and provide a simulation-based baseline for further robustness evaluation and real-world implementation.