Jie (Archer) Lin 
|
|
jie.lin
[at] xeroxlabs.com Imaging
and Xerox
Innovation Group |
Background:
I finished my PhD in Electrical Engineering from Yale University in 2004 in the area of coordination
control and emergent behaviors. My thesis advisor is Prof. A. S. Morse. After graduation,
I joined the Imaging and
Services Technology Center in Xerox
Innovation Group as a Senior Member of Research located at Webster, NY. For more
academic info about me, please refer to my CV.
As prejudice as I could be, I
would like to first refer you to the website of ACSSY (Association of Chinese Students and
Scholars at Yale) where I have served as President from 2001-2002
and made lots of friends. Check out their website, lots of cool stuff and of
course intelligent people. Being prejudice one more time, I would recommend you
to check out Fudan University where I
spent 4 years for my undergrad. By the way, 2005 is Fudan’s centennial anniversary.
As good as it gets, I have a wonderful
family with my wife Haipo and my little Alex born in 2005. Here are some family pictures and Alex’s
first studio set!
Research Interests:
My
doctoral dissertation is concerned with understanding the natural dynamics of
large systems of autonomous agents and how to design local interactions to
bring about particular global goals. Therefore, I am interested in the area of
analysis and synthesis of multi-agent distributed coordination control with
applications to formation control, wireless sensor networks (maybe mobile as
well), distributed intelligent systems, mobile/grid computing and robotics. I
am also interested in distributed resource sharing systems, applications of
structured peer-to-peer overlay networks in distributed resource discovery and
self-organization. In general, I am fascinated by all aspects self-organizing
networks and the kinds of distributed interaction rules that will give rise to
those amazing self-organizing behaviors.
At
Xerox, my work focuses on Enterprise Performance Optimization, Dynamic
Scheduling, Agent-based modeling and optimization, Workflow Automation and
Optimization, Self-Organizing Peer-to-Peer Network Architecture,
Trust/Security/Reputation system in P2P/Wireless/Ad Hoc Networks using social
influence (This is a joint work with Prof
Yang Richard Yang at Yale). I also maintain a UAC gift from Xerox
Foundation to support Prof A. S.
Morse’s research on distributed control and its industrial applications.
Publications:
Here
is a list of my publications.
Research on the Multi-agent
Flocking and Multi-agent Rendezvous Problem:
There
is a growing interest in understanding on the one hand, how various animal
aggregations such as fish schools, bird flocks, deer herds, etc. coordinate
their collective motions to perform useful tasks and on the other, how groups
of mobile autonomous agents such as AUV(Autonomous Underwater Vehicles)
schools, UAV(Unmanned Aerial Vehicles) flocks, etc., might be instructed to
cooperate in a similar manner. Here is a
link to my PhD thesis. A pleasant surprise
recently came through. Part of my thesis work on the flocking behavior
(described below) has won the 2005 IEEE Control System Society George S. Axelby Outstanding
Paper Award based on its originality and potential impact on the
theoretical foundations of control. Here are some
citations to our flocking paper that you can find by standing on Google Scholar’s shoulder.
Not
surprisingly, this rapidly growing research area is extremely challenging.
Because, first of all, a good understanding of the biological principles and
physics underlying the fish schooling and birds flocking is needed.
Furthermore, abstract mathematical models have to be extracted from the animal
behaviors to be applied on the man-made machines. In order to for the mobile
autonomous agents like AUVs and UAVs to coordinate and perform some useful
tasks, the agents must communicate to each other and follow certain sets of
motion rules to plan their individual motions. But the communication is very
costly in a sense that it consumes lots of precious power resource of the
mobile agents. Thus, it is very desirable to have only limited communication
region for each mobile agent to conserve power, for instance only communicating
to its immediate neighbors. This is so called local interaction. Moreover, as
the agents move, every agent can be communicating to different agents at every
instance of time. Thus, a provably correct algorithm has to be imposed on each
agent in order for them to perform a useful task, like foraging and escorting,
etc.. Therefore, it is not least surprising that this research area requires knowledge
from and can be applied in many different fields, such as animal biology,
robotics, automatic control, communications and sensor networks, artificial
intelligence, dynamical systems, algorithms, etc. It is exactly this very
cross-disciplinary aspect of the research that makes it highly interesting but
extremely challenging.
The
outcomes of this research area will undoubtedly have impacts on many facets of
our daily life. For example, the mobile agents that we have considered can be
treated as a set of mobile sensors (hopefully intelligent), a user carrying a
cell phone, or a software agent running on the internet, or a investor in the
stock market communicating to his clients and other investors, all coordinated
to perform some tasks. Moreover, many important and useful applications for software
agents require multiple agents on a network that communicate with each other.
Such agents must find each other and perform a useful joint computation without
having to know about every other such agent on the network. Since the graphical model
studied in my research captures precisely the dynamic nature and the local
aspect of the interacting user environment, it’s extremely suitable for those
studies.
Two
problems that I have worked on and are still exploring are Multi-agent Flocking problem and Multi-agent
Rendezvous problem (conference versions here: flocking,
rendezvous) which I will summarize as follows:
In
the flocking problem, we are interested in regulating a group of autonomous
agents, each with limited sensing region, to move towards a common heading in a
distributed manner. By distributed we mean each agent makes decision on its own
based on the information it receives from its neighbors. No central commander
in the scheme. The same heading problem was first introduced by Craig Reynold in his boid model and later on Vicsek et al. in their physics review
paper performed experiments on the heading rule similar to Reynold’s and
provided a variety of interesting simulation results which demonstrate that the
nearest neighbor rule they are studying can cause all agents to eventually move
in the same direction despite the absence of centralized coordination and
despite the fact that each agent’s set of nearest neighbors change over time.
While Toner and Tu
construct a continuous “hydrodynamics” model of the group of agents and proved
the emergent behavior of the system, in our flocking
paper, we provide a direct theoretical explanation for Vicsek’s
discrete-time model for this observed behavior. To analyze such model, a
discrete-time system recursion equation is set up using the knowledge from
Graph Theory and it proves useful to appeal to well-known result by Wolfowitz
on the convergence of infinite product of ergodic matrices. The study of
infinite matrix products is ongoing. For interested readers please refer to the
reference [19]-[26] in our flocking paper and the
references therein. There are numerous papers following up on our flocking
paper and if you go to scholar.google.com you will find plenty of them.
In
the rendezvous problem, as in [1],
we are interested in regulating a similar capacitated group of autonomous
agents to rendezvous at a single unspecified point in the plane. Again, we
would like to accomplish this task in a distributed manner and without any
global knowledge of the group and the final rendezvous site. We introduce a
sequence of “stop-and-go” maneuvers for
each agent and solve the problem both in synchronous and a truly asynchronous
case. It’s worth noting that many papers existed in the literature on the
coordination control have claimed that their result fits in an asynchronous
setting. But in fact, most of them are actually delayed synchronous version in
that not all agents perform at the same time and the performing agents are
synchronized. In our rendezvous paper (part I, II), we discovered an important principle exists in many of
the emergent behaviors of a multi-agent system, i.e. sticking to a local
property will cause the global property to hold. Also, because of the asynchronous
nature of the problem and the “stop-and-go” maneuvers we introduced, agents’
neighbors position under study becomes obscure. To account for this, we
introduced a concept of “registered neighbors” with “registered positions”
which helped to solve the problem. We also used “Analytic Synchronization” technique to create a deterministic
synchronous discrete system to analyze the asynchronous behavior which is
applicable in other asynchronous analysis.
It’s worth mentioning that we initially used a non-deterministic model to analyze the asynchronous rendezvous
problem which is intrinsically more complex. You can get it here if you are interested in learning the ordeal
that we went through.
Also
lately, I have been working on applying the concept of “Controlled Mobility” to
the area of wireless networking to optimize the energy consumption. For
interested readers, please refer to our recent paper titled “Towards
Mobility as a Network Control Primitive” submitted to Mobihoc 2004. More
interesting topics utilizing the Controlled Mobility in the context of wireless
network will come along in the near future. Stay tuned.
Most recently, I started to get into the area of Small World and its application in physical distributed networks such as wireless networks. Publication will be updated here as I progress.
Proceedings of Block
Island Cooperative Workshop 2003, click
here.
My Previous Papers from Creighton University:
·
My Talk on the IEEE Real Time 99 meeting at
·
My Master’s Thesis at
For Phy. 93, Fudan
Univ.
Enter here
Other Links:
·
Homepage created on
12/02/1997, last modified on 01/12/2004.
·
Please email jie.lin at
aya.yale.edu for any suggestions and/orrrections.
Disclaimer: This document in no way represents anyone else.
All opinions and errors are mine alone.
This page has been viewed times.