A measurement study of predicting throughput from LQI and RSSI

Wireless sensor networks (WSN) commonly use ZigBee tocommunicate, especially when low power consumption is demanded. Zig-Bee may however provide unpredictable throughput although transmis-sion distances are short. This is especially evident in difficult environ-ments with complicated reflections and various materials through whichradio signals need to pass through. Distributed scheduling based on cog-nitive networking principles may improve both network predictabilityand overall throughput. This paper presents measurements of key pa-rameters for such cognitive scheduling, and discusses their potential forpredicting suitable per-node transmission rates. Results include variabil-ity of throughput, RSSI and LQI observed for different transmission pow-ers, transmission ranges, and number of transmitting nodes.Godkänd; 2012; 20130929 (krzwol)CASTT - Centre for Automotive Systems Technologies and Testin