Online estimation of context dynamics and its impact on context sensitive applications

Context management systems allow services and application easy access to remote dynamic context elements, in turns allow these to adapt to the users context. The fact that most information is dynamic and changes value over time, leads to potential risk of obtaining information mismatching the true value due to the communication delay. The probability for this we call the mismatch probability. Obtaining the mismatch probability is not trivial, as it involves knowledge of both the delay process and description of the dynamics of the information element. This paper addresses online estimation of the parameters related to the information element needed to derive the mismatch probability, and the effect that this estimation has on a context sensitive service discovery for Personal Networks.