SOFTWARE RELIABILITY PREDICTION FOR ARMY VEHICLE

Army vehicles are complex due to various on-board mission critical communication devices. The Army cannot afford unreliable software to interact between the devices. The Army vehicle software’s reliability is influenced by multiple factors during or prior to its development. Using complex statistical and mathematical models, software’s reliability can be predicted, but it is dependent on the accuracy and context of the historical software failure data. The cost of developing such complex models does not yield a good return on investment. The data collection process to use these models is very difficult and time consuming. In this paper, we propose reliability metrics based on the current software development and design process factors. We also propose a fuzzy logic based software reliability prediction algorithm using the proposed reliability metrics.