Undoubtedly, automotive industry witnesses a phenomenal development, thanks to highly sophisticated electronic products. On one side, this leads to a tremendous increase in safety and comfort level. However, on the other side, complexity of corresponding test practices is enormously augmented.
Accordingly, automotive industry witnesses also an emerging interest to develop cost-effective test systems, without any quality compromises. Hereby, the ultimate incentive is an economical one since faulty products lead to formidable costs and reputation loss.
In an effort to reduce test burdens, this thesis investigates the possibility to enrich traditional test systems with cognitive abilities, e.g. learning, reasoning, planning, generalization, etc. Hereby, core assumption is the competence of experienced personnel to orchestrate test sessions, which meet a plausible trade-off between costs and benefits. So, the conclusive goal is to realize several cognitive abilities of skilled personnel on a classical test system.
Toward this target, a learning environment is constructed to observe experienced personnel, during test sessions. Consequently, observed test sequences are interpreted, modeled, and stored in its knowledge base.
Asem Eltaher
Testing artificial intelligence cognitive systems test experts