Guideless Autonomous System for Underground Navigation


Joseph Nsasi Bakambu, Fuyi Xu, Vladimi9r Polotski, Paul Cohen,

It is currently recognized by mining society that autonomously guided underground mining vehicles may greatly increase the efficiency of mining production and improve human safety. Recent results in mobile robotics, data fusion and sensor development provide a solid basis for solving the problems related to autonomous underground navigation. Recently developed commercial systems mostly rely on the infrastructure that is expensive and difficult to maintain. Guideless navigation in unstructured and harsh environment of underground mines is still a challenging task. In this paper, we describe an autonomous navigation system for an underground mining vehicle equipped with on-board dead reckoning, inertial and optical sensors. This system does not require any artificial infrastructure and is capable to guide the vehicle along the complex environment of interconnected underground drifts. It also ensures the detection of obstacles on the vehicle path. A navigation simulator, in which a topological graph of the environment and an approximate map are used to specify the path through the drift network is presented. A laboratory testbed developed for testing the feasibility of the proposed navigation methodology is described and the results of extensive laboratory tests and on-site experiments are presented. The current work addresses the system integration for real equipment, and the system upgrade, adding map building and self localization capabilities.
Keywords: guideless, Autonomous, Underground, path-tracking, Simulation, sensor-based, Control, Navigation, Experiments