On the evaluation of real life test expert systems

The authors consider the factors affecting the successful integration of expert systems (ESs) in testing and maintenance environments. The first factor considered has to do with the communication between the ES and the UUT (unit under test) expert and between the ES and the test technician. The authors discuss the importance of embedding in the expert system basic understanding of electronic terms and universal knowledge bases, i.e., knowledge which is not unique to a specific UUT. The second factor considered has to do with the communication between the ES and the ATE (automatic test equipment). It is claimed that artificial intelligence software will not penetrate the real-world test industry, unless it offers very smooth interfaces with test instrumentation and ATE. Several specific requirements are discussed. This work is based on extensive experience in introducing the AITEST expert system for electronic troubleshooting and service management to a wide variety of fields; including aerospace and military, automotive, computers and peripherals, communication and general instrumentation.