Quality evaluation of embedded software in robot software design contest

A robot design contest, called the "Embedded Technology (ET) Software Design Robot Contest," which involves designing software to automatically control a line-trace robot, was held in Tokyo, in 2005. The contest was intended to provide a practical opportunity to educate young Japanese developers in the field of embedded software development. In this paper, we give the results of the contest from the viewpoint of software quality evaluation. We created a framework for evaluating software quality, which integrated the design model quality and the final system performance, and we conducted an analysis using this framework. As a result of the analysis, the quantitative measurement of the structural complexity of the design model was found to have a strong relationship to the qualitative evaluation of the design by the contest judges. On the other hand, no strong correlation between the design model quality evaluated by the judges and the final system performance was found. For embedded software development, it is particularly important to estimate and verify reliability and performance in the early stages, according to the design and analysis models. Based on the results, we consider possible remedies with respect to the models submitted, the evaluation methods used, and the contest specifications. To adequately measure several quality characteristics, including performance, in terms of a model, it is necessary to improve the approach to developing robot software (for example, by applying model-driven development) and to reexamine the evaluation methods.