A data-driven synchronization technique for cyber-physical systems
Conference Paper
Overview
Identity
Additional Document Info
Other
View All
Overview
abstract
2015 ACM. In this paper, we present a method to synchronize data from multiple sensors in a cyber-physical system without any software or hardware modifications to the sensors. This method allows for synchronization of low-power embedded systems in heterogeneous sensor networks, regardless of accuracy of individual sensor clocks by using the events in the physical world to drive the synchronization in the cyber world. We propose two methods to select portions of sensor data streams to drive the synchronization: one leveraging the notion of known templates and the other using an information theoretic approach. Using the events as well as cues from the delay models, we determine alignment points between the data streams. These alignment points are used to synchronize the data. This novel approach is based solely on the sensor data for synchronization, and it can be applied post-deployment on systems of heterogeneous sensors that are not well designed and lack effective synchronization. Experiments show an average accuracy improvement from 12000ppm to 2400ppm for a template basedmethod and from 12000 to 277ppm and 445ppm for information theoretic methods when comparing the synchronized (corrected) clock data to an ideal clock source.
name of conference
Proceedings of the Second International Workshop on the Swarm at the Edge of the Cloud
