The thesis presents research on decision-making in the Tactical Air Combat Environment (TACE). Decision-making occurs often in situations such as business, marketing, medicine and management. In some cases the decision can be quickly made if we have sufficient information or a clear need for the work. In complex cases involving uncertain information, the decision making process is hard and ambiguous. It is difficult to choose or to provide good solutions in areas such as medical treatment, business management or in the battlefield. The model of the tactical air combat environment decision support system is used as the trial model for the Decision Support System (DSS) for the Airborne Early Warning Command and Control (AEW&C). This system is currently designed and developed in the Air Operation Department of the Defence Science Technology Organisation (DSTO) in Australia. The cognitive work analysis (CWA) theory has been investigated and developed in recent years to analyse and develop the human system interaction process to support decision making in TACE. The situation Awareness (SA) theory is also investigated. The thesis introduces theories of decision making and the intelligent techniques that can support the decision making process. Fuzzy Logic or Expert Systems will be used to implement the heuristic knowledge. The training process derived from experience or object recognition will be good useful for the decision making process. Neural network using the back propagatino learning algorithm and its optimisation approaches will be used for this task. Usually a decision support system is made to solve problems where multi-criteria decision are involved. The database is the vital part of the decision support which contains the information or data used in the decision making process. This is where engineers and scientists use several heuristics and soft computing techinques such as learning, search and modelling of imprecise information to obtain optimal decisions. The thesis proposes hybrid intelligent techniques using a fuzzy genetic system and a fuzzy neural system to obtain decision rules automatically. The fuzzy inference system is used to process the imprecise information. Some simulation results demonstrate the difficulties in deciding the optimal quantity of membership functions; shape and parameters are also given. The last part of the thesis explicates a combination of unsupervised learning techniques for clustering the data that is proposed in order to develop decision regions for the fuzzy c mean clustering and self organisation map. It uses a feed forward neural network to classify the decision regions accurately. The clustered data is used for the inputs to the multi-layered feed forward neural network, which is trained using several higher order learning paradigms. Experimental results obtained show the proposed method is efficient. / Thesis (PhDElectronicEngineering)--University of South Australia, 2004
| Identifer | oai:union.ndltd.org:ADTP/267566 |
| Date | January 2004 |
| Creators | Tran, Cong Minh |
| Source Sets | Australiasian Digital Theses Program |
| Language | English |
| Detected Language | English |
| Rights | copyright under review |
Page generated in 0.0036 seconds