Enhanced Perception and Control in Autonomous Robot Using Proximity-Based Fuzzy Logic Sensor Fusion

Reliable navigation in unknown environments is a fundamental challenge in autonomous robotics. This paper presents an optimized sensor fusion framework for nonholonomic mobile robots operating without prior maps. We integrate multi-directional proximity sensors emulating ultrasonic and IR systems via a fuzzy logic-based fusion engine that interprets uncertain, overlapping sensory inputs in real time. The system leverages low level proximity data and converts it into navigation commands by dynamically adjusting input weighting based on environmental complexity. The architecture is implemented in MATLAB and deployed in a co-simulation environment with CoppeliaSim. We evaluate the fusion strategy against single-sensor control baselines using traversal time, collision rate, and smoothness metrics. Results show that the fused system improves both safety and efficiency, achieving in maze-like obstacle courses and reducing path oscillations. Tested configurations, Level 2 emerged as the most efficient and reliable across key metrics, while Level 3 offered maximum safety with 100% task success. These findings suggest the viability of fuzzy-logic-driven sensor fusion for lightweight, real-time navigation on resource-constrained robotic platforms.