A simple method for improving time synchronisation suitable for use in ad-hoc sensor networks where processing power is limited is presented. The technique is local and uses Bayesian estimation. A simple simulation shows an improvement of about four times.
