This paper presents a new algorithm for application layer synchronisation of data from multiple sensors arriving to a mobile phone's Bluetooth interface. A system that provides feedback signals to an athlete is one example where it is crucial to synchronise data from several wireless sensors. This paper also discusses synchronisation problems caused by unpredictable Bluetooth transmission performance. 

This paper presents a method to synchronize the clocks in a Bluetooth piconet from the application layer in a mobile phone. It adapts algorithms for time synchronization of distributed systems and the Internet to Bluetooth networks. The performance issues that cause problems for data synchronization between master and slaves in Bluetooth are highlighted. The tests show that the synchronization error is limited to one sampling time.

Efficient algorithms for time synchronization, including compensation for clock drift, are essential in order to obtain reliable fusion of data samples from multiple wireless sensor nodes. This paper evaluates the performance of algorithms based on three different approaches; one that synchronizes the local clocks on the sensor nodes, and a second that uses a single clock on the receiving node (e.g. a mobile phone), and a third that uses broadcast messages. The performances of the synchronization algorithms are evaluated in wireless personal area networks, especially Bluetooth piconets and ZigBee/IEEE 802.15.4 networks. A new approach for compensation of clock drift and a realtime implementation of single node synchronization from the mobile phone are presented and tested. Finally, applications of data fusion and time synchronization are shown in two different use cases; a kayaking sports case, and monitoring of heart and respiration of prematurely born infants.