Can a robot improve mine safety?

Abstract

Safety in mines is of paramount importance, especially in the labour intensive operations of South Africa, where upward of 300 000 people are employed on a daily basis in an environment that is inherently dangerous. On average approximately 50 people die annually in underground rock fall related incidents with an estimated economic cost of R800 million. This paper explores the potential for deploying an autonomous robotic platform with specialised sensors to measure the roof integrity between blasting and human re-entry. The Platform will attempt to prevent the rockfall incidents by identifying the high risk areas. The CSIR has developed a hard hat mounted electronic sounding device to assist the miner in analysing the sound. It has also been shown that a temperature differential exists between stable and unstable areas. A robotic platform will use a thermal imager to determine potentially loose areas, and then a modified sounding device will probe the hanging wall to delineate the high risk zones and indicate them on a map. The automation of this process removes any variance or ambiguity that is present with a human operator thereby providing consistent results. Localisation is paramount to the success of this robotic application. The dynamic environment in the mine would make any pure SLAM (Simultaneous Localisation And Mapping) application extremely complex. A CSIR developed ultrasonic beacon system will be adapted to meet the need of a disposable low range localisation system. The localisation system will enable the roof integrity measurements to be geo-stamped and therefore superimposed on a map of the area. A 3D map will be composed by the autonomous robot using a time of flight (TOF) camera fused with position and attitude data. The map will then be used for path planning and exploration planning to ensure that the entire area has been inspected, and a complete map is made available to the miners once they re-enter the mine after a blast, thereby enhancing mine productivity by pre-inspecting the stopes.