Analysis of a self-optimizing wireless data network using autonomous mobile wireless routers

Abstract

A self-optimizing wireless data network that can dynamically optimize the network performance by itself is presented. The wireless data network consists of autonomous mobile wireless routers that can change their locations to optimize the connections with users. The performance improvement is demonstrated by packet-level simulations, where the packets are passing through all seven OSI layers. IEEE 802.11 wireless LAN protocol is used in the simulations. The physical layer (PHY) is modeled by transceivers operating at 2.472 GHz frequency with 1 Mbps data rate, BPSK modulation, 15 dBm transmit power, and 100 m coverage range. Different scenarios and navigation algorithms have been investigated and a navigation algorithm combining the trajectory prediction, and the modified circular Hough transform is proposed. By using the algorithm, four mobile wireless routers each has coverage range of 100 m can efficiently cover a wide area of 500 m /spl times/ 500 m. Compared to a conventional network using fixed wireless routers, the proposed network using mobile wireless routers and center of gravity navigation algorithm achieves 38% improvement in dropped packets and 40% improvement in outage time. The trajectory prediction and the modified circular Hough transform further reduce dropped packets by 40% when the mobile wireless router is moving at the same speed as the mobile users.