One of the most important tasks in the anti-terrorism and military
applications is field surveillance and target tracking. Also, surveillance
of community, office building, hospitals, banks, as well as other
important locations demands high detection probability against potential
intruders. Most existing works in the sensor network research assume that
targets are non-intelligent, that is, a target is unaware of existence of
the sensor network. In real life, however, many intruders are smart. These
intelligent targets may try to hide themselves, choose a better moving
path, or intrude the sensor network.

In this talk, we will discuss the characteristics of an intelligent
target, understand its behaviors, and then investigate the methodology to
confront the attack from the intelligent target. In particular, we will
describe how an intelligent target with limited detection range finds the
best path to its destination. We will discuss also how to build a robust
sensor network for surveillance of intelligent targets. Methodologies to
construct sensor networks that are immune from attacks of an intelligent
target and able to intelligently detect the target will be presented:

(1) hiding techniques for sensor networks in the initialization phase,
surveillance phase, and reporting phase;
(2) detection techniques that are designed specifically for intelligent
targets; and
(3) self-healing techniques that repair the damage.