Details

Autor(en) / Beteiligte

• Green, Jeremy

Titel

Requirements Elicitation Techniques for the Development of an Underground Remote Piloted Aerial System

Ort / Verlag

ProQuest Dissertations & Theses

Erscheinungsjahr

2017

Link zum Volltext

• ProQuest

Quelle

ProQuest Dissertations & Theses A&I

Beschreibungen/Notizen

• In South Africa, mining research funding has been dwindling since the mining heyday of the 1980s. More and more companies protect their research and development (R&D) and strategic partnerships as they represent a competitive advantage. The industry as a whole therefore does not have any longterm blue sky research or any common-goal research of which the outputs are available to the entire mining industry. The only such open research is conducted by the universities and the science councils in South Africa. By government dictate, university and science council research should be at least partially owned by the state due to its contribution. As such, it is considered in the public domain and no longer protected. This removes the competitive advantage that the mining houses seek, resulting in a significant reduction in available industry funding for mining research. The impact of this shift in the industry has been so severe that the Council for Scientific and Industrial Research (CSIR) closed its Centre for Mining Innovation in 2014 due to a lack of industry funding. The common-goal research has thus been left to academic institutions. Student-executed academic projects and thesis topics are not always well defined. Typically, a potential supervisor or lecturer would compile a list of possible topics from which students can choose based on their interest, or more likely, based on the likelihood of succeeding. A requirement elicitation process is seldom executed to improve the problem statement and requirements definition. The research outputs focus on a technical result, and therefore the effort focuses on technology development.In the mining environment there are many opportunities where technology can improve safety and productivity. One notable example that is explored in this work as a case study is the use of an underground remote piloted aerial system to scan and inspect the vertical rock passes in underground mines. These rock passes cannot be inspected once they have been constructed without significant disruption to production. Ore flow must be halted for the duration of the inspection, resulting in production losses. The inspection requires either the dangerous entry of a person from the top, or the construction of a support structure from the lower entry point. Ore passages periodically become blocked for a variety of reasons, and when this happens, explosives are used to unblock them. Explosive delivery varies depending on the nature and location of the blockage, but all methods are dangerous to the miners involved. Every year the unblocking activities result in injuries and fatalities in the industry. Production is also halted in the area above the blockage because the ore cannot move through the mine.This research developed a guideline for a requirement elicitation process for new technology development projects in an academic or research environment. Requirement elicitation techniques were matched to the generic project characteristics for the underground RPAS development, and compiled into a guide that can be used for similar projects. Execution of a requirement elicitation phase in a new product development (NPD) project makes the research more relevant and ensures that the research outcomes match industry needs and expectations. This makes it more likely that the industry would adopt and implement these outcomes on completion of the study. The guide was tested by executing it in the development of the underground remote piloted aerial system (RPAS). The utilisation of the underground RPAS for the temporal scanning of rock passes can help to identify potential future failure sites and the possible causes of such failures. This can serve as a tool to plan scheduled maintenance to prevent blockages. This would eliminate production interruptions and remove the need for the dangerous process of unblocking the rock pass using explosives. In the event where a blockage is not prevented, the underground RPAS can be used to scan the blockage to determine its structure and cause and to determine an unblocking strategy. The strategy could include using the underground RPAS to sacrificially deliver an explosive charge directly to a targeted point on the blockage. This would mean that blockages can be removed in a much safer way than what current techniques allow. The blockage could also be removed quicker than the current methods allow, resulting in a significant decrease in production losses due to blockages in underground mines.The requirement elicitation techniques were developed to suite the generic project characteristics of an academic research activity that is technology-driven, with initially unspecified and unknown stakeholders, spread over a geographically diverse area and across time zones, with a limited budget and a flexible timeline. The resulting techniques of semi-structured one-on-one interviews with a brainstorming session involved a questionnaire, unified modeling language (UML) coding of use cases, business cases and multiple deployment scenarios, and subsystem prototypes as research tools.The subsystem prototypes addressed the initially identified technology challenges of the application. A Hubsan X4 mini-UAV was demonstrated to familiarise the interviewee with the technology. A platform preservation system based on an array of ultrasonic sensors with basic data visualisation was demonstrated to illustrate the detection and avoidance of walls and obstacles. A mapping and operator visualisation system based on an ASUS Xtion Pro LIVE depth sensor was demonstrated to illustrate data streaming and map-building. A drift sensor was described to counter the ventilation air, causing the aerial platform to drift down a tunnel. The final subsystem prototype was a basic mockup of the operator consul, prompting discussion regarding the type of operator and the person’s skill level, the data to be presented on the screen, the level of autonomy that may be needed and expected, and the possible operation philosophy for executing the various possible missions of the system.The stakeholder interviews were analysed to determine if there was sufficient representation in the three identified areas of expertise, namely: of mining, surveying or mapping, and unmanned aerial systems. The interview transcripts were then analysed to produce a first round of requirements,which included identifying conflicting and uncertain performance criteria. The outcomes were used to develop a full size integrated RPAS prototype. A quadcopter was built, incorporating the anticipated 1kg load capacity for a sacrificial explosive delivery. Instrumentation included a gimbal mounted depth sensor for real-time operator visualisation on a glasses mounted display, and for 3D data collection for offline map creation. The platform preservation subsystem was adapted to be an orthogonal array of six ultrasonic sensors. The prototype quadcopter was then demonstrated to a select stakeholder group in a mining research tunnel for requirements triage. Agreement was obtained on the conflicting requirement performance requirements, and thus the actual requirements of the system were discovered.The requirement elicitation guideline is successfully demonstrated in the development of a prototype underground RPAS that is demonstrated in a realistic mining environment. The requirement elicitation guide improve the research outputs of academic institutions when there is no specific industry partner that can assist in the project. Industry stakeholders volunteer small amounts of time to assist in developing a realistic requirement specification for technology development, thus improving the quality and applicability of the research outputs, and the likelihood of industry adoption and support of the new technology. The underground-RPAS prototype demonstrated has the potential to improve underground mine safety by preventing the injuries and fatalities that occur every year in the inspection and unblocking of rock passes. It also has the potential to improve the industry productivity, not only by avoiding the cost of a fatality, which is noted to be as much as R31 million, but also avoiding production losses of approximately R600 000 per day.