A formal methods-based Rule Verification Framework for end-user programming in campus Building Automation Systems

Abstract

Building Automation Systems have recently exposed programming interfaces for occupants to dynamically control and personalize indoor workplace environments. To this end, the If-This-Then-That paradigm has been proposed as a user-friendly programming approach. However, conflicts, both within the occupant comfort rules and between the rules in open environment and the overall system policy, often result in reduced energy efficiency and other undesired outcomes. In this paper, we propose a software framework based on formal methods to detect and resolve rule conflicts called Rule Verification Framework. The proposed framework builds upon state-of-the-art Satisfiability Modulo Theories and utilizes a suite of algorithms to ensure system correctness and safety. The proposed framework was experimentally evaluated using a prototype implementation developed using Java and the Web Services technology. The proposed framework detected more conflicts and was faster than the previous solutions. Conflict detection algorithms had to work iteratively because conflict resolution resulted in new conflicts.