I am working with Prof Demosthenis Teneketzis on sequential
decentralized optimization problems that arise different application domains
and analyze them using a common conceptual framework. This framework can be
generalized to arbitrary sequential decentralized systems. I have
worked on the following problems.
Real-Time Communication
In modern decentralized systems the design of a communication scheme is not a
stand alone problem but an integrated component of a larger system; the
objective is no longer just to communicate information but also to
characterize how the flow of information affects the overall system
performance. In such systems communication is delay sensitive, since the data
is temporal in nature and loses its significance if not communicated and
decoded in time. This motivates the need to study communication systems with
a hard constraint on communication delay. Such communication systems are
called real-time communication systems.
Real-time communication problems are drastically different from classical
information theoretic formulations which does not model communication delay.
The fundamental concepts of information theory such as source entropy, rate
distortion, and channel capacity are asymptotic concepts and do not provide
much insight for the real-time communication problem. So, we need a new
mathematical formulation to analyze real-time communication. In my research, I
formulate the real-time communication problem as a sequential decentralized
optimization problem and develop a solution methodology to determine optimal
communication schemes. I have considered two variations of the real-time
communication for point-to-point communication system: (i) real-time
communication with no feedback in [J3], [C2],
and [C1], and (ii) real-time communication with
noisy feedback in [J5] and [C4].
Control over limited noisy communication
In the classical formulation of the control system, it is assumed that the
sensors and the controllers are co-located with the plant. Recently, there has
been a lot of interest in network controlled systems (NCS) where the plant,
sensor, and controller are spatially distributed and their operation is
coordinated via a communication network. NCS are a good model for many
distributed control applications such as sensor networks,
micro-electromechanical systems, cellular telephony, and industrial control
networks, where the communication channel between the components does not have
a large bandwidth which makes it imperative to consider an integrated approach
for the design of both control and communication components.
NCS bring out the fundamental conflicts between the design philosophies of
control and communication theories. Communication theory is concerned with the
reliable communication of information from one point to another, and does not
care about the meaning or usefulness of the communicated information at the
receiving node; control theory, on the other hand, is concerned with using the
information at the controller to achieve some performance objective.
Communication theory does not model the delay incurred during transmission of
information while timeliness of information is of paramount importance in
control theory. These conflicts pose fundamental questions for both fields
which makes an integrated control and communication design a challenging
problem. In my thesis, I provide partial answers to these questions. I have
considered the problem of optimal design of communication and control
strategies for a NCS with a noisy communication link between the sensor and
the controller in [J2] and [C3].
Decentralized diagnosis with communication between diagnosers
Fault detection is an important task in the automated control of large scale
systems. Large-scale systems have local diagnosers at various nodes of the
system each having partial local information. It is usually technologically
demanding and economically expensive to collect all this decentralized
information at one single node. Yet it is often possible for the nodes to
communicate limited information to one another. This has motivated the study
of how to communicate information efficiently amongst nodes to ensure that
faults are diagnosed, in particular how to minimize communication needed to
detect faults. Such decentralized fault diagnosis problems arise in various
application areas including document processing systems, heating, ventilation,
and air-conditioning systems, intelligent transportation systems, chemical
process control, and telecommunication networks. The design of such
communication policies is intricate due to the complex interdependence between
the communication and diagnosing policies of the diagnosers. In my thesis I
have considered this problem using the discrete event system
(DES) modelling formalism with the assumption of synchronous
diagnosers. A paper documenting this work is currently in preparation.
Book Chapters
-
Aditya Mahajan and Demosthenis Teneketzis
Multi-Armed Bandits
in
Foundations and Applications of Sensor Management
Springer-Verlag, 2007, pp. 121–151.
(pdf)
Journal Publications
-
Aditya Mahajan and Demosthenis Teneketzis
On the Design of Globally Optimal communication strategies for Real-Time Noisy Communication with Noisy Feedback
IEEE Journal on Selected Areas in Communication
vol. 26, no. 4, pp. 580–595, May 2008.
(abstract)
(paper)
A real-time communication system with noisy feedback is considered. The system
consists of a Markov source, forward and backward discrete memoryless
channels, and a receiver with limited memory. The receiver can send messages
to the encoder over the backward noisy channel. The encoding at the encoder
and the decoding, the feedback, and the memory update at the receiver must be
done in real-time. A distortion metric that does not tolerate delays is given.
The objective is to design an optimal real-time communication strategy, i.e.,
design optimal real-time encoding, decoding, feedback, and memory update
strategies to minimize the total expected distortion over a finite horizon.
This problem is formulated as a decentralized stochastic optimization problem
and a methodology for its sequential decomposition is presented. This results
in a set of nested optimality equations that can be used to sequentially
determine optimal communication strategies. The methodology simplifies the
search for determining an optimal real-time communication strategy.
-
Aditya Mahajan and Demosthenis Teneketzis
Globally Optimal Performance of Feedback Control Systems with Limited Communication over Noisy Channels
SIAM Journal of Control and Optimization submitted for publication.
(abstract)
(paper)
A discrete time stochastic feedback control system with a noisy
communication channel between the sensor and the controller is
considered. The sensor has limited memory and at each time, it
transmits an encoded symbol over the channel and updates its memory. The
controller receives a noise-corrupted copy of the transmitted symbol. It
generates a control action based on all its past observations and all
its past actions. This control action action is fed back into the
system. At each time instant t, an instantaneous cost, which is
a function of the state of the system at t and the control
action at t, is incurred. The objective is to choose encoding,
memory update and control strategies to minimize: an expected total cost
over a finite horizon, or an expected discounted cost over an infinite
horizon, or an average cost per unit time over an infinite horizon. For
each case a sequential decomposition of the global optimization problem
is obtained. The results are extended to the case when the sensor makes
an imperfect observation of the state of the system.
-
Aditya Mahajan and Demosthenis Teneketzis
On Globally Optimal Encoding, Decoding and Memory Update for Noisy Real-Time Communication Systems
IEEE Transactions on Information Theory submitted for publication
(Revision May 9, 2008).
( abstract)
(paper)
(code)
(pretty printed code)
The design of optimal joint source-channel communication strategies for a
real-time communication system, i.e., a sequential communication system in
which information must be transmitted and decoded withing a fixed-finite
delay, is considered. First, a system which runs for a finite horizon and
consists of a first-order Markov source, a real-time encoder, a memoryless
noisy channel, a real-time decoder with finite memory, and distortion
metric that accepts zero delay, is considered. The design of optimal
real-time communication strategies is formulated as a decentralized
stochastic optimization problem. There is no existing solution methodology
to solve general decentralized stochastic optimization problems over
finite and infinite horizon. This paper develops a systematic methodology,
based on the notions of information structure and information state, to
sequentially obtain globally optimal real-time encoding, decoding, and
memory update strategies. Such a sequential decomposition results in a set
of nested optimality equations whose solution determines an optimal
communication strategy. This methodology is extended to two classes of
infinite-horizon systems, where optimal communication strategies are
determined by the solution of an appropriate functional equation. The
methodology is also extended to systems where distortion metric accepts a
fixed-finite delay, to systems with higher-order Markov sources, and to
systems with channels with memory. Thus, this paper develops a
comprehensive method to study different variations of real-time
communication.
-
Aditya Mahajan, Manu Agarwal, and Ajit K. Chaturvedi
A novel method for down-conversion of multiple bandpass signals
IEEE Transactions on Wireless Communication vol. 5,
no. 2, pp. 427–434, February 2006.
( abstract)
(paper)
Simultaneous down-conversion of multiple bandpass signals is desirable for
a number of wireless applications. Bandpass sampling technique can be used
for this purpose, but it is difficult to implement and has several
drawbacks. In this paper we propose a novel front-end technique to
directly down-convert multiple frequency division multiplexed (FDM)
signals separated by certain minimum frequency. A special down-conversion
function is derived to achieve simultaneous down-conversion of the received
signals. The technique requires simpler bandpass filters and the ADC has a
baseband input as compared to bandpass sampling, which imposes strict
requirements on bandpass filters and requires an ADC which can handle RF
inputs. The performance of the method has been evaluated by simulating a
BPSK receiver employing this technique.
-
Baquer Mazhari and Aditya Mahajan
An improved interpretation of depletion approximation in p-n junctions
IEEE Transactions on Education vol. 48, no. 1,
pp. 60–62, February 2005.
(abstract)
(paper)
The conventional treatment of depletion approximation in
p-n-junctions often leaves a student with an erroneous impression
that the approximation essentially involves complete neglect of the
fraction of the space charge region (SCR) where charge density makes a
transition to zero. This work describes a simple analytical model for
clarifying the relationship of depletion approximation to the SCR.
Conference Publications
-
Mahta Moghaddam, Dara Entekhabi, Leila Farhadi, Yuriy Goykhman, Mingyan Liu, Aditya Mahajan, Ashutosh Nayyar, David Shuman, and Demosthenis Teneketzis
Initial Analyses and Demonstration of a Soil Moisture Smart Sensor Web
Proceedings of the Earth Science Technology Conference (ESTC)
East Adelphi, MD, June 24–26, 2008.
(paper)
-
Aditya Mahajan and Demosthenis Teneketzis
On-time diagnosis in discrete event systems
Proceedings of the 9th International Workshop on Discrete Event Systems (WODES)
pp. 382–387, Gothenberg, Sweden, May 28–30, 2008.
(paper)
(examples)
(code)
-
Aditya Mahajan and Demosthenis Teneketzis
Real-Time Communication Systems with Noisy Feedback
IEEE Information Theory Workshop
pp. 283–287, Lake Tahoe, CA, September 2–6, 2007.
(paper)
-
Aditya Mahajan and Demosthenis Teneketzis
Optimal Performance of Feedback Control Systems with Limited Communication over Noisy Channels
Proceedings of the 45th IEEE Conference on Decision and Control
pp. 3228–3235, San Diego, CA, December 13–15, 2006.
(paper)
-
Aditya Mahajan and Demosthenis Teneketzis
Fixed Delay Optimal Joint Source-Channel Coding for Finite-Memory Systems
Proceedings of the International Symposium of Information Theory
pp. 2319–2323, Seattle, WA, July 9–14, 2006.
(paper)
-
Aditya Mahajan and Demosthenis Teneketzis
A decision theoretic approach to real-time communication
Proceedings of the 43rd Annual Allerton Conference on Communication, Control and Computing
September 2005.
(paper).
Thesis
-
Aditya Mahajan
Sequential decomposition of sequential dynamic teams: applications to real-time communication and networked control systems
University of Michigan
September, 2008.
(thesis) (presentation)
Talks and Presentations