Low-Power Sensor Stack Tunability for Multi-Radio Support

Abstract

The challenge in realizing a multi-radio stack in sensor node system architectures is optimization for radio protocol processing and low-power considerations. Since smart sensor node platforms typically would have computation blocks that can handle minimal instructions for handling data, the paradigm for extending the sensors' utilities and deployment scope necessitates multi-radio support on a single optimized stack. Our idea is to make an integrated stack tunable and optimized to the sensor node platform architecture. In this paper, we explore tunability of radio protocol stacks from the perspective of multi-radio support while keeping power consumption as a performance metric in sensor nodes. Our approach is based on a scheduler design based on traffic conditioning, along with a reconfigurable frame parser that helps in pre-processing of the radio protocol packets before they are scheduled according to a choice set of MAC and PHY units that form the vertical flow of protocol stack processing run-time. This is currently work-in-progress and we are developing our concepts and test prototype implementation.