Abstract:
There is growing interest in the use of wireless sensor data to control physical systems. Given the operational constraints (especially related to the power usage) that wireless sensors operate under, there has been considerable attention given to approaches for optimising these networks. This has considered the relationship between network behaviours and control system performance (particularly stability), as well as ways in which the network traffic can be reduced. Suprisingly however, there has been very limited consideration of dynamically reducing the underlying data generation (i.e. sampling) rates, based on what data is needed for the control system to perform adequately. In this paper we argue that this approach can potentially minimise the sensor energy usage assocated with sampling. We describe the basic approach and present the results of a preliminary case study that demonstrates the feasibility of the technique.
