Artificial Intelligence
Intelligent decision support feedback using Multi-Agent System in a defence maintenance environment
International Journal of Intelligent Information and Database Systems
Hi-index | 0.00 |
In this paper we describe the successful introduction of the model-based technology in a troubleshooting support tool for aircraft wiring under industrial conditions at Airbus, where strict cost/benefit analyzes decide over success or failure of a tool. Aircraft built at Airbus Deutschland GmbH are highly customized products: virtually no two aircraft are identical. Troubleshooting support cannot be built in the traditional way of compiling static decision trees once and for all aircraft. Rather, diagnosis dialogs must be computed on demand for each individual aircraft. Our solution is based on dynamically generating the diagnosis dialogs from on-line generated electrical signal nets. Generating diagnosis dialogs on demand is a considerable technical challenge. We have solved this with a combination of intelligent document extraction services and model-based diagnosis technologies. However, the paper focuses less on technical issues and more on the non-technological aspects that highlight the challenges of deploying well established intelligent technologies in industry. For example, while existing engineering data is correctly interpretable by “fault tolerant and flexible” humans, it remains a challenging input for algorithms. Furthermore, some psychological aspects remain a challenge to the acceptance of the proposed tool.