Physical layer design and implementation of distributed arrays in packet networks

Abstract

Multi-input multi-output (MIMO) is well known to improve the throughput and reliability of wireless links, and has thus become an essential element for wireless communication standards, such as IEEE 802.11, Long Term Evolution, and WiMAX. A conventional MIMO link has an array antenna (i.e., several antennas on a common hardware platform, sharing the same local oscillator and clock) at one or both ends of the link. For example, in multi-user MIMO, the base station or access point has multiple antennas. However, distributing the antennas across multiple hardware platforms without a common local oscillator has several advantages, such as larger coverage area for the same number of antenna elements, more flexible network topology, and lower cost of individual hardware platforms. The main challenge in the implementation of a distributed or virtual array is synchronizing the platforms in time, frequency, and in some cases, phase. For example, state-of-the-art synchronization techniques fail at extreme ranges, far beyond single-input-single-output (SISO) range. In wireless packet networks, which includes local area networks, sensor networks, and ad hoc networks, synchronization depends on the preamble of the packet. This dissertation explores how new packet designs and new signal processing algorithms can adequately synchronize and operate distributed arrays to achieve range extension and to achieve higher data rates in dense deployments. Both OFDM-based wideband transmission techniques and MSK low data rate transmission techniques are developed. All techniques are demonstrated using software defined radio (SDR).

The main contributions of this dissertation are as follows. For distributed arrays in range extension scenarios, novel synchronization strategies with diversity gain and robustness against interference are presented. One of these is a novel OFDM preamble design and synchronization method with diversity gain, presented for time division cooperative transmission (TDCT). For extending range for spatially multiplexed multi-user MIMO communications, a new relay self-selection scheme is presented for amplify-and-forward (AF) relaying. With the aim of enabling transmit-only wireless sensor networks with long battery life, a novel interference-insensitive minimum shift keying (MSK) synchronization scheme is presented, along with a new statistical model of the residual remaining after interference cancelation. The last contribution aims to facilitate high throughput wireless LANs, through dense deployments of co-channel wireless LAN access points. That contribution is a hybrid method of selective application of nonlinear precoding (NLP) on top of interference alignment (IA) beams, over multi-user, multi-access-point MIMO networks. The Tomlinson-Harashima NLP scheme is implemented on SDR to demonstrate real-time packet decoding performance.