Simulation of Real-Time Frontier Exploration in Confined & Cluttered Environment

Abstract

This paper presents a real-time simulation of a frontier exploration robot using a nonholonomic mobile robot’s kinematic model. In MATLAB, three heuristic-based frontier selection methods, namely Randomized Histogram Sector (RHS), Informed Randomized Point (IRP), and Histogram Clustering (HC), were simulated within two constrained space maps, one of which (Map 1) contained more obstacles than the other (Map 2). For the suggested frontier detection, the percentage of successfully mapped area against the ground truth map was determined to be 97.5% for Map 1 and 98.3% for Map 2. These results indicate that the proposed method yields great results. The HC approach is the most effective of the three proposed methods for both maps in terms of the time required to explore the maps and the efficiency of navigation inside the maps.