The development of intelligent service robotic systems is currently an active field of research in the robotics community. For example assistive robots that can aid elderly and disabled people in daily life activities. One emerging requirement for this type of system is the inclusion of the user in the decision process through physical and cognitive collaboration. This human-in-the-loop (HIL) concept allows for the use of the human perception and cognitive abilities in order to safely achieve the tasks that would be too complex to perform in a purely autonomous way. However, the overall human-machine system is complex and may be difficult to analyze. The user and the robot are operating in a closed loop and both are potentially capable of adapting to the other. The users may have a disparate set of noisy channels available for communicating their intended commands to the robot. The robots are typically dexterous and are expected to operate in an unstructured environment. Metrics can help in the analysis, development, and benchmarking of this type of system, by quantifying performance and driving the mutual learning and adaptation process. However, there are currently few such metrics available. Information Theory and related information-based concepts have been applied in disparate fields such as communications, human factors, control theory and cognitive processes. The work presented here attempts to identify metrics based on these concepts for assistive human- in-the-loop cognitive systems.