An ML Approach for Decoding Collision Slots

Abstract

The tag inventory process in an EPCglocal Class-1 Generation-2 (EPCglobal Gen2) long-range Radio Frequency Identification (RFID) system is based on Framed Slotted ALOHA (FSA). Collisions between tags are inevitable in an FSA-based system and limit its maximal throughput. In this work we describe a simple Maximum Likelihood (ML) scheme for jointlydecoding R tag replies in a collision-slot, allowing the reader to decode all of the colliding tag replies and greatly increasing the probability of decoding at least on tag reply correctly.