Over the past decade, wireless sensor network research primarily relied on highly-integrated commercial off-the-shelf radio chips. The rigid silicon implementation of the radio stack restricted access to the lower layers; thus, research focused mainly on the medium access control (MAC) layer and above. SRAM field-programmable gate array (FPGA)-based software-defined radios (SDR), on the other hand, provide a flexible architecture to experiment with any and all layers of the radio stack, but usually require desktop computers and draw high currents that prohibit mobile or longer-term outdoor deployments. To address these issues, we have developed a modular flash FPGA-based wireless research platform, called Marmote SDR, that has computational resources comparable to those of SRAM FPGA-based radio platforms, but at a reduced power consumption, with duty cycling support. We discuss the design decisions underlying Marmote SDR and evaluate its power consumption. Furthermore, we present and evaluate an asynchronous and multiple access communication protocol specifically designed for data-gathering wireless sensor networks

Abstract-In this paper we describe our custom designed, lowpower, intelligent sensor platform, and a novel analysis approach for Structural Health Monitoring (SHM). More specifically we show how Acoustic Emission (AE) signals were recorded during aluminium and steel beam break tests utilizing two channels on our MarmotE platform, and how subsequent low-resource but accurate onset time detection yielded Time Difference of Arrival (TDoA) results. We also demonstrate a new, simplified method to pick valid AE events from a vast set of noisy measurements, and prove the feasibility of our ideas by showing that our approach provided results comparable to widely used industry methods with modest resource requirements.

No abstract

No abstract