Proposal and Overview

(The final project report can be found under the 'Files & References' section)

Overview: Many mobile wireless sensor network applications require a high bandwidth data link back to a remote host. One approach to deliver high SNR with over a long distance is considered in this project, using clusters of nodes as relays. This is motivated by two observations: 1) Cooperative beamforming among N nodes allows for directivity up to N2, 2) the spatially distributed receivers are expected to add robustness to beamforming techniques. 

Goals: The purpose of this project is to consider whether the spatial diversity of a receive cluster (as opposed to a single receive node) allows for greater robustness to synchronization error on the transmitter side. A simple FDM system will be used as a baseline for comparison.

Challenges: Optimal cooperative beamforming requires perfect channel information, which can be considered as perfect knowledge of the relative positions of the transmitters and receivers. In static networks, this is possible (though often impractical), but in mobile networks the best localization techniques only provide centimeter level localization. Timing and carrier frequency synchronization also present significant error since each antenna is run by separate clock generation circuitry.

Prior work: These challenges make it impractical to assume perfect beamforming, and several studies consider the effects of uncertainty on the quality of the main beam. In [1], the authors identify significant degradation of the beam quality in random arrays due to uncertainty in phase. They present a probabilistic model for predicting the degradation as a function of wavelength, distance to transmitter and phase variance. In [2], a series of algorithms are proposed to overcome the issues described above for cooperative beamforming. The authors demonstrate promising results for communications to a single destination node from a static wireless sensor network. 

System Description: The system consists of two clusters, one transmitting and one receiving, which are spatially separated by a significant distance. Transmitting clusters aim to maximize SNR per unit power in order to communicate at a large data rate. Receive clusters may perform limited distributed processing on the received signal. It is assumed that channels change slowly compared to the symbol time (no significant fading effects).  The uncertainty models will be based heavily on the work in [1] and [2]. Time permitting, the effect of node formation will also be considered (as opposed to random placement models from the references). It is worth noting that the primary design criteria for this system is not energy efficiency, as often the lifetime of nodes is limited by actuation costs. Instead, the goal is to maximize SNRs given limited radio power.

Collaborative Beamforming: Nodes broadcast the same message simultaneously. Transmit nodes adjust phase to steer the beam to the centroid of the receive cluster. The receive SNR at each node is recorded. Monte Carlo simulations will be run to determine the average maximum receive SNR for a variety of uncertainty parameters, cluster sizes, and cluster separations. 

Frequency Division: A naive frequency division multiplexing method will be tested along-side the beamforming approach. Each node is given a portion of the data and transmits over a designated frequency bin. Due to the simplicity of this approach, it is expected to perform worse than collaborative beamforming. However, if the collaborative beamforming does not outperform, then it is clearly not a viable method of communication under the system model.

   Feedback:  Time permitting, limited feedback schemes will be extended for the multi-receiver case and tested via simulation. Numerous papers such as [3] and [4] demonstrate that feedback can improve the beam quality with little overhead. Extension to the multi-receiver case is not obvious and may not preserve the properties previously demonstrated.

A pdf version of this proposal is available under 'Files and References'


[1] H. Ochiai, P. Mitran, H. V. Poor, and V. Tarokh, "Collaborative beamforming for distributed wireless ad hoc sensor networks," IEEE Trans. Signal Process., vol. 53, no. 11, pp. 4110-4124, 2005. 

[2] M. F. Urso, S. Arnon, M. Mondin, E. Falletti, F. Sellone, "A distributed and collaborative beamforming algorithm for a self-organizing wireless sensor network," Physical Communication, vol. 4, iss. 3, pp. 172-181, Sept. 2011

[3] R. Mudumbai, B. Wild, U. Madhow, and K. Ramchandran, “Distributed beamforming using 1 bit feedback: from concept to realization,” in Proc. 44th Allerton Conference on Communication Control and Computing, Sept. 2006. 

[4] M.-O. Pun , D. R. Brown IIIand H. V. Poor  "Opportunistic collaborative beamforming with one-bit feedback",  IEEE Trans. Wireless Commun.,  vol. 8,  no. 5,  pp.2629 -2641 2009