A cost model for testing unmanned and autonomous systems of systems

Abstract

The evolutionary nature of Unmanned and Autonomous Systems of Systems (UASoS) acquisition needs to be matched by equally evolutionary test capabilities in the future. There is currently no standard method to determine what is required to make programs safe for deployment, nor is there the ability to make effective contingency plans should testing requirements change. Spending too much effort designing goals when causal understandings are still in flux is inefficient. As such, policy making and enforcing policies on the deployment of UASoS becomes very problematic. Testing is required especially for UASoS to identify risk, improve capabilities and minimize unpleasant surprises. It needs to be effective and focused, determining the issues and working towards ensuring the risks of the UASoS are known. It is important to have adequate feedback loops, a culture of information sharing and learning from best practices, as well as the development of metrics and/or performance indicators that adequately reflect the effectiveness of the test process. This thesis describes a model that is part of a larger Prescriptive and Adaptive Testing Framework (PATFrame), which uses knowledge acquisition to minimize risk through a decision support system. This work presents the cost and risk considerations for UASoS T&E and provides the preliminary parameters to conduct trade-off analyses for T&E. It also provides guidance on how the DoD can adopt such tools to transform the DoD T&E enterprise. The model is a combination of information collected from various normative and descriptive views of testing based on literature review, surveys, and interviews with members of the Department of Defense (DoD) T&E community A cost estimation model can have significant impacts on how the DoD currently does testing and would help maximize the use of the resources available. It is a model based method for calculating effort for test and evaluation and forms a baseline for strategic decision making in DoD acquisition programs. The intent is to predict within a certain probability that a test program can be completed within a certain budget given the assumptions used in characterizing the UASoS and the T&E process.