Low-Cost Independent Clock Estimation

Abstract

This thesis investigates developing an independent clock estimation algorithm for low-cost wireless sensor nodes. This would allow each sensor node to maintain time synchronization by estimating its own clock frequency based on environmental factors without requiring periodic transmissions between nodes. An algorithm was researched and investigated using representative testbeds. The testbeds were subjected to different environmental stimuli to determine if temperature was the main contributing factor for the low-cost clocks used. Several different testbeds were subjected to different temperature conditions and a clock estimation algorithm was designed around the collected data. The estimation algorithm was then directly implemented on several testbed micro-controllers and real-time experiments were conducted in a number of different temperature environments. The results show that the algorithm holds potential for increasing frequency stability and compensating for clock drift in applications where transmissions between sensor nodes is limited or undesirable.