Energy-efficient traffic-aware station grouping for low-power dense wireless networks.

Existing wireless technologies often exhibit poor performance in very dense networks, where hundreds or even thousands of stations need to connect to the same access point. This is mostly caused by the increased probability of two devices transmitting data at the same time, which causes the data packets to collide and be lost. Recently, station grouping has been proposed as a new method for collision-free data transmission in these dense environments. The basic idea is that stations are split into groups and each group is given a specific time interval during which only its members can transmit. This limits the maximum simultaneous transmissions, and therefore potential collisions. A station grouping configuration has many degrees of freedom: the number of groups, their duration, and which stations belong to each group. Several algorithms have been proposed to determine the optimal configuration as a function of the number of stations and their traffic demand. However, they all have several shortcomings that we aim to address in this project: they assume very specific and static traffic patterns, they do not optimise the trade-off between energy consumption and performance, and they cannot avoid interference among multiple overlapping networks. In this project, we will develop novel accurate mathematical models, based on Markov chains and supervised machine learning, that accurate estimate the energy consumption and throughput performance of specific station grouping configurations. This will be used to develop real-time station grouping algorithms that can handle heterogeneous stations and traffic, and adapt to changes in the traffic patterns. Finally, the algorithm will be extended to multiple access points, and used to implement interference avoidance mechanisms. The resulting solution will be evaluated using simulation to assess scalability, and will be implemented on real hardware to assess real-time execution time constraints.

Keywords:WIRELESS NETWORKS, SENSOR NETWORKS, INTERNET OF THINGS

Disciplines:Communication networks, Performance modelling, Wireless communications