METHODS OF RELIABILITY-CONTROL FOR AUTONOMOUS UNDERWATER VEHICLES

Lead Research Organisation: Imperial College London
Department Name: Dept of Computing

Abstract

Many autonomous underwater vehicles (AUVs) have been recently developed and successfully deployed for various oceanographic, industry and defense applications. Two key concerns of AUVS are that they are expensive to produce and that they are mostly one-off products thereby disallowing quality control techniques and analysis used in mass production. This proposal aims to analyse techniques and to build a system aimed at reducing the chance of AUVs malfunctioningor becoming irrecoverably lost; hence, indirectly, the proposal aims to reduce the effective operational cost of AUVs. The project is to build on the many years of experience of designing, building and operating AUVs at the National Oceanography Centre at Southampton University. The main parts of the proposed system are and fault assessment system (IFAS), design optimization tool for overall operational reliability (DOPTOR) and human interface for daily operation (HIDO) .The cost of development and then loss of AUVs is very high relative to the design system proposed. The methodology developped in the project would pay for itself in a single vehicle and many times over should it be replicated. Collateral benefits of the proposal are that the methodology generated will largely be transferable to users other than the collaborators and also to other types of costly autonomous vehicles on land and in air.
 
Description In this research we have developed a methodology for assessing the impact of faults in autonomous systems at runtime. We have devised techniques that enable us to predict the consequences of specific faults (sensors, engines, components, etc) and assess the reliability and resilience of autonomous systems against faults. We have developed the theoretical underpinnings for doing so and released an open-source platform which can be used by engineers to carry out the process. We have applied the method to study the resilience of the AUTOSUB6000, an experimental autonomous submarine.
Exploitation Route Engineers can download and use the toolkit produced to evaluate the resilience of their design.

Researchers can improve the technique further to support a wider range of faults.
Sectors Aerospace, Defence and Marine,Digital/Communication/Information Technologies (including Software),Transport

 
Description The outcome of this research have been used to assess the reliability and diagnosability properties of an autonomous submarine. The methodology has been released; other groups will be able to replicate the results and adapt the methodology for other studies.
First Year Of Impact 2010
Sector Aerospace, Defence and Marine,Digital/Communication/Information Technologies (including Software),Security and Diplomacy
Impact Types Cultural,Economic