An Interaction-Driven Approach to Identifying Functional Behaviors of Service Robot Systems

The service robot domain has gained increasing attention for its significant impact on human society. By its nature, the service robot domain is component-based and distributed and comprising components are usually developed by third parties as black-box. In general, very limited information is available even in execution logs. This makes it difficult for the system integrator to understand and verify the functional behaviors of the system. The existing work on deriving functional behaviors focuses on sequential systems and assumes certain level of information available, which is not suitable for the service robot domain. In this work, we present an interaction-based approach for systematically identifying functional behaviors of a service robot system. This approach allows not only identifying functional behaviors, but also their dependencies as invariants which are useful for verifying system behaviors. We demonstrate the approach using a cognitive game robot and the results show the potential of the approach.