Communications Research Colloquium and Informal Seminar

School of Electrical and Computer Engineering -- Cornell University

Fall 2002 -- Meetings: Tuesdays, 1-2pm, 310 Rhodes (coffee and cookies available before)

If you would like to give a talk, please get in touch with Sergio Servetto.

Date	Speaker	Title
Tue 9/10	Joao Barros	On the Capacity of the Reachback Channel in Wireless Sensor Networks.
Tue 9/17	Sergio Servetto	Dry-run for two talks on sensor networks, for ACM MobiCom/WSNA 2002.
Tue 9/24	Srihari Adireddy	CANCELLED, moved to 11/05.
Tue 10/01	Anna Scaglione	Opportunistic Large Arrays
Tue 10/08	Brian Evans	Time-Domain Equalization for Maximizing Bit Rate in ADSL Transceivers.
Tue 10/15	Fall break	No seminar
Tue 10/22	Geert Ysebaert, Koen Vanbleu	Per-tone Equalization for ADSL Transceivers.
Tue 10/29	Rick Martin	Infinite Length Results and Design Implications for Time-Domain Equalizers.
Tue 11/05	Srihari Adireddy	On the Use of Channel State Information for Slotted ALOHA in Fading Channels.
Tue 11/12	a Speaker	a Title
Tue 11/19	a Speaker	a Title
Tue 11/26	Thanksgiving	No seminar
Tue 12/03	a Speaker	a Title

Speaker:

Joao F. Barros.

Title:

On the Capacity of the Reachback Channel in Wireless Sensor Networks.

Abstract:

Wireless sensor networks offer flexible and effective ways to acquire information on a certain process occuring in a designated area. In many applications, the sensors must transmit this information to a central base station, which assembles the collected data for subsequent processing. The protocols and transmission schemes that sensors use to reach back and convey their information pose a number of challenging research problems. In this talk, we will show how network information theory can be used to characterize the ultimate performance limits set by the reachback channel in wireless sensor networks.

First, we consider a problem in which multiple sources want to send correlated data to a single receiver over independent channels. For this case, we give an exact characterization of the joint source/channel capacity region, i.e., which conditions on the sources and the channels guarantee that the information can be transmitted with arbitrarily small probability of error. These conditions generalize in a very meaningful way the condition that H(X) When the conditions for reliable communication are not met, the best we can hope for is to minimize some measure of distortion, which leads us to a rate-distortion problem with correlated sources and independent encoders. We will conclude our talk with some new results on this long standing problem.

Joint work with Sergio Servetto.

Speaker:

Sergio D. Servetto.

Title:

Dry-run for two talks on sensor networks, for ACM MobiCom/WSNA 2002.

Abstract:

This will be my first rehearsal for two talks that I will present next week at MobiCom. One of the papers is co-authored with G. Barrenechea, entitled Constrained Random Walks on Random Graphs: Routing Algorithms for Large-Scale Wireless Sensor Networks, and has the following abstract:

We consider a routing problem in the context of large scale networks with uncontrolled dynamics. A case of uncontrolled dynamics that has been studied extensively is that of mobile nodes, as this is typically the case in cellular and mobile ad-hoc networks. In this paper however we study routing in the presence of a different type of dynamics: nodes do not move, but instead switch between active and inactive states at random times. Our interest in this case is motivated by the behavior of sensor nodes powered by renewable sources, such as solar cells or ambient vibrations. In this paper we formalize the corresponding routing problem as a problem of constructing suitably constrained random walks on random dynamic graphs. We argue that these random walks should be designed so that their resulting invariant distribution achieves a certain load balancing property, and we give simple distributed algorithms to compute the local parameters for the random walks that achieve the sought behavior. A truly novel feature of our formulation is that the algorithms we obtain are able to route messages along all possible routes between a source and a destination node, without performing explicit route discovery/repair computations, and without maintaining explicit state information about available routes at the nodes. To the best of our knowledge, these are the first algorithms that achieve true multipath routing (in a statistical sense), at the complexity of simple stateless operations.

The other paper is co-authored with A. Scaglione, entitled On the Interdependence of Routing and Data Compression in Multi-Hop Sensor Networks, and has the following abstract:

We consider a problem of broadcast communication in a multi-hop network, in which samples of a random field are collected at each node of the network, and the goal is for every node to obtain an estimate of the entire field within a prescribed distortion value. The main idea we explore in this paper is that of jointly compressing the data generated by different nodes as this information travels over multiple hops, to eliminate correlations in the representation of the sampled field. Our main contributions are (a) that we obtain, using simple network flow concepts, conditions on the rate/distortion function of the random field, so as to guarantee that any node can obtain the measurements collected at every other node in the network, quantized to within any prescribed distortion value; and (b) we construct a large class of physically-motivated stochastic models for sensor data, for which we are able to prove that the joint rate/distortion function of all the data generated by the whole network grows slower than the bounds found in (a). A truly novel aspect of our work is the tight coupling between routing and source coding, explicitly formulated in a simple and analytically tractable model---to the best of our knowledge, this connection had not been studied before.

Both papers can be downloaded from http://people.ece.cornell.edu/servetto/.

Speaker:: Anna Scaglione.
Title:: Opportunistic Large Arrays.
Abstract:: In this talk, we introduce a new scheme that allows transmission from a set of asynchronous transmitters to a remote destination which, otherwise, would not be capable to forward the data individually. In our scheme the small devices form an Opportunistic Large Array (OLA) by reacting to the signal sent by a special node in the network, called the leader. The problem solution provides as a by-product a communication technique for the network itself, based on a unique physical layer design where the signals are broadcasted to the other nodes without contending for the channel.

Speaker:

Brian Evans.

Title:

Equalization for ADSL Transceivers.

Abstract:

Asymmetric Digital Subscriber Line (ADSL) systems divide a 1.1 MHz broadband channel into 4.3 kHz narrowband subchannels and transmit quadrature amplitude modulation signals on the narrowband channels via multicarrier modulation implemented by the inverse fast Fourier transform (FFT). The number of message bits carried on each subchannel is specified by a bit allocation table that is returned by the receiver during modem initialization. Before a bit allocation table can be created, the ADSL receiver uses training sequences to train for:

time-domain equalization (TEQ) to shorten the effective duration of the channel impulse response, and
frequency-domain equalization (FEQ) to compensate for magnitude and phase distortion in the channel.

Equalizer design is the key to maximizing bit rate in an ADSL modem. This talk reviews three ADSL equalizer structures and design methods for them. For the conventional equalizer, which is a cascade of a finite impulse response TEQ, FFT, and single-tap FEQ, we propose the

Maximum Bit Rate method to maximize the bit rate at the TEQ output,
Minimum Inter-Symbol Interference method to find a near-optimal bit rate at the TEQ output, which is amenable to implementation in real-time fixed-point software.

We also propose a dual-path TEQ structure, and apply the above design methods to maximize the achievable bit rate at the TEQ output. We compare the tradeoff between achievable bit rate at the FEQ output and implementation complexity against the per-tone equalizer. The per-tone equalizer resembles a filter bank approach and provides the best reported bit rates in the open literature at the FEQ output.

We have implemented design methods for conventional, dual-path, and per-tone equalizers in a TEQ Design Toolbox for Matlab. Download slides.

Speaker:: Geert Ysebaert, Koen Vanbleu.
Title:: Per-tone Equalization for ADSL Transceivers.
Abstract:: Per-tone equalization provides an alternative to time-domain equalization in ADSL receivers, which is typically steered by a channel shortening criterion. Per-tone equalization is steered by SNR optimization for each tone separately and hence provides improved performance (capacity) under all circumstances. An additional advantage is that synchronization delay tuning becomes a significantly less critical task. Per-tone equalization provides this improved performance without any transmission complexity increase. Initialization complexity, however, as well as the memory requirement are increased significantly. Recent work has focussed on cheap (i.e. affordable from an implementation point of view) initialization schemes, as well as pruned schemes with a reduced memory requirement. In addition, the per-tone approach has been applied to echo cancellation, and combined with windowing techniques and non-linear filtering techniques for improved interference cancellation.

Speaker:: Rick Martin.
Title:: Infinite Length Results and Design Implications for Time-Domain Equalizers.
Abstract:: Time-domain equalization is crucial in reducing channel state dimension in maximum likelihood sequence estimation, and inter-carrier and inter-symbol interference in 802.11a and ADSL multicarrier systems. A time-domain equalizer (TEQ), which is a finite impulse response (FIR) filter, placed in cascade with the channel produces an effective impulse response of v+1 samples that is shorter than the channel impulse response. We analyze two families of TEQ design methods amenable to cost-effective real-time implementation: minimum mean squared error (MMSE) and maximum shortening SNR (MSSNR) methods. For infinite length TEQs, MMSE target impulse responses have all v zeros on the unit circle. Hence, finite-length MMSE TEQs will eventually yield decreasing bit rate with increasing length. For infinite length TEQs, MMSE target impulse responses and MSSNR unit norm TEQ (UNT) constraint TEQs are symmetric. A symmetric TEQ reduces FIR implementation complexity in half, enables the frequency-domain equalizer and TEQ to be trained in parallel, and exhibits only a small loss in bit rate over non-symmetric TEQs. In addition, a symmetric MSSNR-UNT TEQ reduces TEQ training computational complexity by a factor of 4 and doubles the length of the TEQ that can be designed using fixed-point arithmetic.

Speaker:: Srihari Adireddy.
Title:: On the Use of Channel State Information for Slotted ALOHA in Fading Channels.
Abstract:: We consider the use of channel state information to vary the transmission probability for Slotted ALOHA in fading channels. We first introduce a general reception model for the physical layer, that takes into account the channel states of the transmitting users and also allows for the modeling of multiple simultaneous successful receptions. We then introduce the notion of asymptotic stable throughput for Slotted ALOHA. We will propose a transmission control that achieves significant gains in the asymptotic stable throughput. As an example, the results of the general theory will then be applied to the particular case of CDMA networks. This application serves to give more insight into the general theory and also shows that significant gains can be achieved.