Fuzzy system dynamics risk analysis (FuSDRA) of autonomous underwater vehicle (AUV) deployment in the Antarctic
Autonomous Underwater Vehicles (AUVs) are self-powered robotic devices that operate underwater with a propulsion system. In recent years, the maturing of autonomous technology and commercialization of AUVs has led to the rapid expansion of AUV types and capabilities. As a result, they are now being...
Main Author: | |
---|---|
Format: | Thesis |
Language: | unknown |
Published: |
University of Tasmania
2020
|
Subjects: | |
Online Access: | https://dx.doi.org/10.25959/100.00035098 https://eprints.utas.edu.au/id/eprint/35098 |
Summary: | Autonomous Underwater Vehicles (AUVs) are self-powered robotic devices that operate underwater with a propulsion system. In recent years, the maturing of autonomous technology and commercialization of AUVs has led to the rapid expansion of AUV types and capabilities. As a result, they are now being used in many scientific, commercial and military applications, such as underwater mine-clearing operations, feature tracking, cable or pipeline inspection, deep ocean exploration and air crash investigations. In a relatively new development, there has been a growing interest in the use of AUVs for under-ice marine science research in the Antarctic. For many years, researchers have limited access to investigate these ice-covered waters, but the use of AUV aims to change that. Concealed under the Antarctic’s sea ice lies a vast amount of intrinsic scientific information across a wide range of scientific disciplines, offering insights from Earth’s climate system to human biology. In particular, understanding the impact of climate change on Antarctic’s sea ice is critically important because of its contribution to global sea level rise and its role in regulating the world’s climate system. However, the deployment of AUVs in the Antarctic for under-ice marine science research is not an easy undertaking. It is a complex operation involving thorough and careful planning, collaboration with multiple stakeholders, working in adverse environmental conditions, and often fraught with logistical, financial and technical challenges. During the mission itself, additional considerations are needed to account for ice cover, accessibility and emergency abort procedures. Therefore, it comes as no surprise that there is an increased risk of losing an AUV during operations in the Antarctic when compared to open water missions in other relatively benign environments. The loss of an AUV is not only financially costly due to the resulting higher insurance premium for all (if it is insured, or loss/rebuild costs if it is not), it can also delay research projects, damage the reputation of the AUV community, cause the loss of valuable research data and a possibility of harming the delicate Antarctic environment. It is therefore imperative that the risk of loss be analysed and managed effectively for deployment of AUVs in the Antarctic. Significant developments had been made over the years in risk analyses methodologies to better analyse and manage the risk of AUV loss during deployment. Early efforts focused on the prevention of technical failures to improve reliability and increase life span of AUVs. Later, proactive and systematic risk analysis approaches emerges, to predict the likelihood of loss by analysing historical performance data of the AUV. Gradually, the scope of risk analysis broadens from analysing historical performance of an AUV to other operating uncertainties and phases of deployment. In recent development, more attention has been devoted to the role of organisational and human factors in the overall risk of AUV loss during deployment. Despite improvement in risk analysis approaches to AUV deployments, predicting the risk of loss remains a highly uncertain exercise heavily dependent on historical performance data. Two main areas for improvement were identified to develop a more comprehensive and effective risk analysis methodological framework for Antarctic AUV deployment. First, the time-dependent nature of risks and the complex interrelationships between risk variables of an AUV program needs to be examined collectively as a whole. Second, to reduce dependency on historical performance data by accounting for vagueness and ambiguity in elicitation of expert’s opinion. To address the first shortcoming, a dynamic systems-based risk analysis framework facilitated by system-dynamics methodology is proposed. The use of system dynamics enables modelling of the complex, interrelated and dynamic systems behind an AUV program which may influence the risk of AUV loss during an Antarctic deployment. For the second shortcoming, a fuzzy-based risk analysis framework based on expert’s judgement is suggested. The use of a fuzzy logic overcomes limitations due to the lack of empirical data and accounts for the uncertainties about causal relationships between risk variables. Lastly, a hybrid Fuzzy System-Dynamics Risk Analysis (FuSDRA) framework is proposed. Leveraging strengths while overcoming limitations of both fuzzy logic and system dynamics, the novel approach provides a structured, robust and effective solution for risk analysis of Antarctic AUV deployment. The usefulness of the FuSDRA framework was demonstrated in a case study based on the University of Tasmania’s (UTAS) nupiri muka AUV program. Supported by the Australian government through the Antarctic Gateway Partnership initiative, the objective of the program is to develop a polar capable AUV for the acquisition of high-quality underwater data. The explorer-class AUV was delivered in May 2017 with its first Antarctic deployment in December 2018. Using information sought primarily from interviews of domain experts in UTAS and supported by other knowledge sources, FuSDRA models were developed and tested. Scenario analysis was performed on the models to understand the behaviour of the risk of loss under different circumstances. This included: 1) Knowledge loss due to departure of critical employee, 2) reducing government support and increasing alternatives to the AUV, and 3) increasingly dysfunctional interpersonal dynamics. Simulation results from model testing and scenario analysis were then used to derive a set of policy recommendations to better manage the risk of loss. The importance of implementing an effective budget management system, obtaining diversity in funding, reducing risk of obsolescence, optimizing recruitment strategy and improving interpersonal dynamics and stress awareness were highlighted. This dissertation lays the foundation for structured risk analysis frameworks with the eventual goal of reducing risk of AUV loss during Antarctic deployment. The main contribution, the FuSDRA approach may also be applicable to other types of AUV operations or complex technological systems. To enhance the usability and ability to solve real-world problems, further work is proposed on the FuSDRA framework. |
---|