Predict Energy Consumption of Trigger-Driven Sensor Network by Markov Chains
Wei Zha, Wee Keong Ng · 2011
Markov Model has been proved its feasibility of predicting the energy state of sensor nodes. Thus, user can monitor sensor nodes energy state in real-time without querying them frequently. However, a stationary state transition probability is required to apply Markov Model, which means the prediction is only applicable to schedule-driven sensor networks rather than trigger-driven sensor networks. In this paper, we will introduce how to use Markov Model to make prediction in trigger-driven sensor networks. By considering events distribution and query patterns, our proposed method managed to predict sensor node energy level information of trigger-driven sensor networks. Experimental results show that our proposed model is able to predict sensor node energy state accurately for trigger-driven sensor networks.