The CSIR working on robots to improve mine safety

Mining is a dangerous occupation, and South Africa’s hard rock (gold and platinum) mining is more dangerous than most (but not all) types of mining. In the period January 1 to August 15 this year, 76 South African miners lost their lives in workplace accidents. The total for the whole of last year was 128, and for 2009 the toll was 168.

The trend is, fortunately, downwards. And the death toll has fallen hugely over the past quarter century – it stood at 865 miners in 1986. But, even if South Africa reduced its mine fatality rate to international standards, that would still work out at 34 deaths a year (assuming no further decline in the size of the country’s mine workforce).

The very great depths of South African gold mines are a key factor regarding these deaths – gold miners work in a remote and hostile environment. Naturally, there is a widespread desire to increase the use of technology in South African mining.

The Rise of the Robots

Robots. A concept created and popularised by science-fiction. But the first real working robot bore no resemblance to the humanoid robots beloved of science-fiction. This was Unimate, developed by the Unimation (Universal Automation) company in the US. Unimate, which entered service on a General Motors assembly line in 1961, was the forerunner of all of today’s industrial robots and, being in the form of a large mechanical arm, also set the format that most such robots still follow.

More recently, however, a new field of robotics has opened up – field robotics. “In science fiction, robots have true artificial intelligence; unlimited working environment, like humans; can learn [are not pre-programmed] and can work cooperatively alongside humans. In reality today, the traditional ‘home’ of robots is a production line, with repetitive work, fully controlled environment and no interaction with humans, because it is not safe. Such robots are unaware of what is around them. They just do their tasks. Field robotics is relatively new. Field robots are in-between industrial robots and science fiction. They operate in unstructured environments but in limited roles. They can act autonomously. Some can operate with people safely. Some have limited learning ability,” explains Council for Scientific and Industrial Research (CSIR) mining robotics project manager Liam Candy.

Internationally, rapid progress has been made in field robotics technology over the past decade, given a huge impetus by the demands of war. (Field robotics is currently focused mainly on unmanned air vehicles, vessels, and ground vehicles.) US technology market research company ABI Research has calculated that $5.8-billion was spent on military robotics worldwide last year and that this figure will increase by 70% to more than $8-billion in 2016. In addition, there is lesser but still significant civilian investment in field robotics, for cargo handling, mining and agricultural applications, among others.

South Africa cannot match the scale of the investments, military or civil, being made overseas. Yet the country cannot ignore field robotics either. It is becoming too important, too dynamic, a field, relevant to many areas of human endeavour. As a result, the CSIR is undertaking research into field robotics, through the formation of the Mobile Intelligent Autonomous Systems (Mias) group. “This is an emerging research area for the CSIR, so it is different from a competency area,” explains Mias group leader Dr Simukai Utete. “Our group targets niche areas which address national needs, niche areas which are of relevance to [our] society – for example, mining. We are very concerned about capacity development in robotics.”

“Regarding robotics in mining, the [world] leader is probably the Australian Centre for Field Robotics (ACFR),” reports Candy. * “It has developed a Load Haul Dump (LHD) vehicle that is partially autonomous – after being loaded, it proceeds autonomously to exit from the mine. The ACFR has also developed a partially autonomous dragline, an automated bucket loader and automated charging of blast holes.”

Other leading players in mining robotics research are The Robotics Institute of Carnegie Mellon University (in Pittsburgh in Pennsylvania in the US), with its Groundhog autonomous vehicle (used to do three dimensional mapping of abandoned coal mine tunnels in conditions too dangerous for humans); Sandvik of Sweden, which has developed an autonomous LHD vehicle; Atlas Copco, also of Sweden, which has also developed an almost fully autonomous LHD (called a Scooptram in North America); and Anglo American plc, as part of its 2030 Mine project. Most of the systems developed to date are for use in open cast operations, because these pose far fewer difficulties than underground operations.

From PackBot to MineBot

The CSIR Mias group has four research areas – perception, planning, navigation and machine learning. Its projects are carried out in the areas of mining robotics, “mule” robotics (unmanned cargo carrying ground vehicles – see Engineering News September 16, 2011), intelligent manipulation and active vision for autonomous systems.

“We develop the autonomy (intelligence),” says Utete. “We often buy in the platforms. We are very much about codevelopment with other research and industrial partners. Our research is not focused on sensor design. We take data from different sensors and integrate it using sensor fusion for different applications.” Some of the sensors employed (bought in) by the CSIR Mias team are very expensive, but it is expected that, as the global demand for them increases, the costs of the sensor packages will fall.

“In mining robotics, we work on robot systems that could be applied in a mine environment,” she explains. “It’s a project we work on jointly with the CSIR Centre for Mining Innovation (CMI) and with the CSIR Mechatronics and Micro-Manufacturing (MMM) competence area. Jeremy Green of the CMI is principal investigator for the Mine Safety Platform project (which enters its final year next year). Mias is working on intelligent autonomy for the platform while CMI develops the sensor suite and the MMM group develops a low-cost platform.”

“There are few mines in the world like South African gold and platinum mines,” highlights Candy. “Hard rock mining happens in a cycle – drill, blast and clean. In between when they blast and when they remove the ore (clean) they have to do a safety inspection. At the moment that’s conducted by people and it’s quite a dangerous task. What we are hoping to do is have a robot that can go in and automatically do a preliminary inspection and generate a risk map for the miners who have to go in after the blast. Basically, we want to take a human being out of a really dangerous, dirty, job and put him in a safe area.”

To put if more formally, the CSIR Mias group is seeking to develop a fully autonomous robot that can evaluate hanging wall (tunnel ceiling) stability. This is not easy. Such a robot needs to know where it is and where it wants to go (summed up by the term localisation), what the environment around it is like (mapping), how to plan a route through that environment (path planning or navigation), and how it should proceed (motion planning).

To save time and money, the research group acquired an in-production field robot of a proven, reliable, tough design to act as the initial platform for their proposed machine – the abovementioned Mine Safety Platform (MSP). The robot they bought is a PackBot 510, manufactured by US group iRobot. PackBot 510s are military tactical robots, normally employed for explosive ordnance disposal, surveillance and reconnaissance, route clearance and hazardous materials detection missions. They can climb stairs, climb over rubble and operate in narrow passages. They operate either by remote control or semi-autonomously and more than 3 500 have been produced so far. However, the CSIR MMM competency area is currently developing a low-cost robot platform that could be used in mines in the future, instead of having to employ adapted overseas designs.

Turning a semi-autonomous robot which operates above ground or in shallow tunnels into a fully autonomous robot that operates deep underground is quite a challenge. “You can imagine, if you’re doing above ground robotics, you have, for starters, the advantage of GPS [Global Positioning System]. Even though GPS, even differential GPS, struggles to tell you with pinpoint accuracy where you are, even the crudest GPS can tell you your location within 100m or 200 m – it can give you a global location,” states Candy. “When you’re underground, you have nothing like that. You’re in an environment that has no beacons, very few distinguishing features and and not much in the way of landmarks. Basically, it’s dark, it’s wet and it’s muddy. It’s difficult to identify one position underground from another.”

An underground mine environment cannot, in practice, be instrumented to provide a control and navigation network for robots. “We need the intelligence to be on the machine,” he affirms. “We’re doing the preliminary work on solving the problems around mapping, path planning, motion planning and localisation.” To fulfill these functions, the Mias team is focusing on using laser sensors (which function similarly to radar, but on a different part of the electromagnetic spectrum), Time of Flight cameras (which provide three-dimensional surface information of an object in real time) and inertial sensors.

“And then we’ve got some novel localisation schemes, involving RFID [radio frequency identification] tags, for example. Some of these are quite new techniques and some of them are quite cutting edge techniques.” One concept being developed is Simultaneous Localisation and Mapping using RFID, or Slam RFID. RFID tags, used in shops as security tags attached, for example, to clothing, are cheap any easy to acquire. They could be scattered throughout a mine and the MSP could build up a map by detecting and remembering the individual tags. After blasting, when the ore was removed, the surviving RFID tags would go with it, which would allow mine management to track the ore and know what ore came from what part of the mine.

Of course, it is not just a matter of hardware. There is also the need to write the software which will allow the MSP to process the data it receives and decide what to do – to think in a very basic manner (like, perhaps, an insect).

The first demonstrator model of the MSP is due to undergo initial underground trials within the next few weeks (before of the end of the year). “This will be our baseline MSP, not a final production model, and this will be our first milestone,” avers Candy.

Gas! gas!

The Mias has a second mine safety project that would make use of the MSP. This is concerned with the remote underground monitoring of toxic gases in underground mines. “Underground mine terrains are restricted and can be dangerous due to poisonous gases and dust following blasting,” points out project leader Dr Isaac Osunmakinde. Examples are Black Damp, which is a mixture of dust, carbon dioxide and molecular nitrogen, which is heavier than air and displaces oxygen, and so can cause asphyxiation; and methane, which is often released during coal mining and is lighter than air – it also displaces oxygen and is highly explosive.

There are, of course, effective safety measures and procedures in place, but these can always be improved. But achieving such improvement requires better knowledge of the behaviour these gases and dust in the confined spaces of underground mines. “The objective of this project is to understand the suspension of toxic gases and dust in air,” he explains. “This will generate the knowledge to answer the question of whether toxic gas concentrations are high, low or have disappeared, and allow the creation of maps to show which areas of a mine are safe and which are dangerous.” In turn, this will allow the development of better accident prevention measures and improve the safety of miners.

To achieve these objectives, the Mias group is developing a real time gas monitoring system, which will have static and mobile versions. The intent is that the mobile monitors will be mounted on the same basic robot as the MSP, and it is likely the same basic robot will be used, when necessary, to deploy static monitors as well.

The fundamental aim of CSIR Mias mining robotics research is to increase the safety of miners and reduce the death toll in the industry. It is not aimed at displacing workers with machines, unlike in other countries which have high labour costs and career options that are much more attractive than mining. 

* Subsequent to the publication of this article, Mining Weekly was informed that the LHD automation, including dragline, shovel and explosive loading, was actually carried out by the Autonomous Systems Laboratory of the ICT Centre of the Commonwealth Scientific and Industrial Research Organisation, located near Brisbane in Queensland, Australia.