A framework for verification and validation of integrated and adaptive control systems

Abstract

This paper summarizes a project that seeks to extend previous work in formal methods, functional languages, adaptive control and hybrid control to establish a methodology and a framework for validation and verification of adaptive/intelligent software systems. We are working to apply recent results in creating logical abstractions in Haskell to build a capability for automatic verification of new or modified components which meet the constraints imposed on the resulting system. As components become ready for use, we intend to apply the specification for the Open Control Platform, currently under development as part of the US Defense Advanced Research Projects Agency (DARPA) Software-Enabled Control project, as the definition of our distributed control platform. Our approach emphasizes the use of off-line verification and validation techniques to establish ranges of parameter variations to be allowed by the adaptive/intelligent control software and online techniques to ensure that approved constraints are being met. We are in the process of developing an engineering plan to demonstrate use of the binary decision diagram (BDD) and Stanford Validity Checker (SVC) extensions to Haskell to automatically validate an adaptive, online control law redesign and implementation to meet performance specifications. This implementation will be designed to demonstrate an initial capability that can be extended to other adaptive and intelligent control systems safety critical tasks.