An LTE Random Access Channel Model for Wireless Sensor Network Applications

In this thesis we investigate the scheduling aspect of using 3GPP Long Term Evolution(LTE) as a radio transport in Wireless Sensor Network (WSN) applications.A possible performance bottleneck in LTE transmission scheduling is identified where atypical WSN would interfere with traditional cellular network users. All users are allotedequal resources in a specialized signalling channel (PUCCH), which are to be used forscheduling requests. This indicates that a WSN with a high number of nodes and alow traffic intensity would skew that signalling channel in such a way that regular userswould see increased delays and thus latency. We suggest a subtle change in the LTEscheduling request procedure to allow such low-intensity nodes to use Random Access(RACH) exclusively, foregoing their resources in the regular scheduling channel.We use a probabilistic network-level model of Multichannel Slotted ALOHA[1] as thebasis for our own mathematical model of LTE RACH. This is done by identifying thedecision points of the RACH protocol and devising ways to model them using probabilitytheory. several key changes to the base model are proposed to better suit our purpose.Finally, the results are verified using a discrete-time event simulator. The operationand structure of the simulator are described in detail and its results compared and usedto correct misalignments in the original model.We conclude that the throughput and delay of the LTE Random Access Channel canbe accurately modelled as a Slotted ALOHA network with multiple independent channelsand suggest methods to model the two success/failure selection stages.Validerat; 20120509 (anonymous