Knowledge-based interpretation of foetal phonocardiographic signals

A knowledge-based system (KBS) is presented for the processing and analysis of foetal phonocardiographic signals. Conventional analytic techniques of signal processing are inappropriate for such a volatile signal environment where little information about the signal events is known beforehand. The essential feature of this KBS is its hierarchical organisation of procedurally based knowledge areas. This formulation ensures that the knowledge is consistent and also allows an efficient search for applicable rules. An analysis of the signal is effected using a solution island driving approach. The performance of the system is comparable to that achieved by a domain expert.