WiFi coverage planning and router position optimization using machine learning

In this paper, a regression based supervised learning algorithm Mels (Modified Extensible Lattice Sequence) combined with Global response search method (GRSM) optimization routine is used for (i) design of a WiFi antenna for IEEE 802.11 WLAN protocol, and (ii) radio coverage planning and optimization of number of routers to maximize coverage using the designed antenna for an office space. A dual slot antenna is designed to operate at 5.2 GHz (S46 or S54 band) with a gain of 6 dBi and used as a representative router antenna for coverage planning. Using machine learning, the number of routers is optimized such that maximum coverage with at least a minimum date rate of 13 Mbps is achieved over a 15000 square foot office space.