Hardware Offset Multi-User Detection for Ultra-Low-Complexity Wireless Sensor Nodes

Wireless sensor nodes need a drastically reduced technical complexity to fit constraints of future applications. Current multi-user detection requires large amounts of technical complexity, leading to increased sensor node size, cost and energy consumption. Instead of deterministic resource allocation we propose an ultra-low-complexity communication scheme using truly asynchronous direct random access. The randomness in this scheme solely stems from manufacturing tolerances, further decreasing cost, size and energy consumption. While these tolerances have to be compensated for at great expense in classic digital communication, we use them specifically to enable multi-user detection with sensor nodes that do not differ in design or configuration. After studying the necessary steps for a measurement reconstruction at the base station we propose an algorithm, based on a time-frequency transformation in combination with a peak search to implement the first of these steps. We derive an analytical lower bound on the user-capacity of the scheme and compare the results with the performance of the proposed detection algorithm.