Authors: Muhammad Aria Rajasa Pohan, Jana Utama, Budi Herdiana
DOI: 10.17509/ajse.v4i3.75343
Abstract
This study aims to propose a new motion planning strategy for autonomous vehicles (AV) using fuzzy logic to improve safety. The strategy mitigates risky behaviors of other road users, such as zigzagging vehicles, sudden braking, pedestrians emerging from blind spots, and responding to sudden lane changes. The proposed method combines a novel fuzzy inference system and a configuration space map with adaptive dynamic object bounding boxes. These bounding boxes adjust in size according to the risk level of the dynamic object’s movement. The simulation test was conducted using four scenarios involving risky behavior from other road users. The proposed method was compared with conventional methods, with safety costs used to measure performance. The results showed that the proposed algorithm achieved better safety costs across all four scenarios; this improvement is due to the integration of the fuzzy system with the adaptive configuration space map, which accounts for uncertainties. These findings suggest that the proposed method improves AV motion planning safety in dynamic and unpredictable environments. This research contributes to developing safer, more reliable autonomous driving systems. © 2024 Universitas Pendidikan Indonesia.
Author keywords
Autonomous vehicles; Fuzzy logic; Motion planning; Movement uncertainties; Risky behaviors; Safety enhancement
This article can be accessed at https://www.scopus.com/pages/publications/85215819061



