Ethologically Inspired Hybrid Fuzzy-Virtual Force Field Navigation in ROS: A Simulation-Based Study

Robust and adaptive navigation in dynamic environments remains a central challenge in autonomous robotics. Traditional methods such as Virtual Force Field (VFF) navigation are prone to issues like local-minima traps and unstable trajectories in cluttered spaces. This paper presents a simulation-based, ethologically inspired hybrid navigation framework that integrates a Fuzzy Behaviour module with VFF-based motion planning within the Robot Operating System (ROS). The proposed controller encodes biologically inspired internal states such as fear and escape into fuzzy logic rules that modulate repulsive and attractive forces based on real-time sensor data. Unlike the conventional fuzzy-VFF systems, the model interprets environmental cues through emotional analogs to support more context-sensitive behavioural decisions. The fuzzy rules and membership functions are specified using the Fuzzy Behaviour Description Language (FBDL) and implemented with LIDAR sensing in dynamic ROS-Gazebo scenarios. The system was evaluated across 25 simulation trials with varying obstacle and agent configurations. Compared to a standard VFF baseline, the proposed approach achieved a 28.1% reduction in collisions, a 13.3% reduction in task completion time, a 7.5% improvement in behaviour switching latency, and a 14.7% increase in model accuracy. While the results demonstrate significant gains in adaptability and interpretability, the current study is limited to simulation-based validation. The future work will focus on hardware deployment, quantitative validation through statistical testing, and optimization of the fuzzy rule base via learning-based approaches.