A Comprehensive Study on Knowledge Management Applied to Embedded System Testing Processes

Abstract

Numerous commercial electronic devices such as computers and mobile phones are composed of embedded systems. In general, these systems present complex functionalities, making the process of developing and testing a challenging task for engineers and designers. Testing is essential for validating any embedded device in order to guarantee that their features are operational and reliable. Improved testing methods could increase the reliability and quality of the devices. In this paper, a test methodology for evaluating commercial embedded devices is proposed. The proposed approach is based on the use of historical knowledge acquired by analyzing and modeling a growing test database composed by information on prior regular test routines. This approach aims to track the similarities among test cases applied in past products in order to optimize the analysis of test cases necessary for the evaluation of recent products that will be submitted to testing processes. A comparative case study was conducted involving the proposed test methodology and a standard test methodology based on test cases in a test routine involving embedded devices produced by a company in the energy solutions business. The empirical results showed a significant improvement in the efficiency of failure and defect detection in the embedded devices, allowing improved management of the processes of development, test, and production.