Probabilistic data association for wireless passive body sensor networks

Abstract

Passive communications based on modulated backscatter can dramatically extend the lifespan of wireless sensor networks. Because wireless passive sensor networks reduce the dependence on the batteries, they have great potential to be used in body sensing applications. In a typical sensor network deployment, multiple passive sensor nodes modulate a common carrier. The superimposed reflected signals can be resolved by using code division multiple access (CDMA) principles. In practical systems however, a strong carrier component will leak to the receiver, causing offset to the constellation points and as a result it will destroy cross correlation properties of the CDMA codes. To address this problem, we propose a code shift modulation instead of vanilla CDMA. We develop maximum likelihood receiver and analyse its performance. Since the complexity of ML detection strategy scales exponentially with the number of sensors, we develop a reduced complexity detection technique based on Probabilistic Data Association (PDA). We demonstrate via numerical simulations that our reduced complexity detector approaches the ML performance.