LOCALIZED MOVEMENT CONTROL CONNECTIVITY RESTORATION ALGORITHMS FOR WIRELESS SENSOR AND ACTOR NETWORKS

Wireless Sensor and Actor Networks (WSANs) are gaining an increased interest
because of their suitability for mission-critical applications that require autonomous
and intelligent interaction with the environment. Hazardous application environments
such as forest fire monitoring, disaster management, search and rescue, homeland
security, battlefield reconnaissance, etc. make actors susceptible to physical damage.
Failure of a critical (i.e. cut-vertex) actor partitions the inter-actor network into
disjointed segments while leaving a coverage hole. Maintaining inter-actor
connectivity is extremely important in mission-critical applications of WSANs where
actors have to quickly plan an optimal coordinated response to detected events. Some
proactive approaches pursued in the literature deploy redundant nodes to provide fault
tolerance; however, this necessitates a large actor count that leads to higher cost and
becomes impractical. On the other hand, the harsh environment strictly prohibits an
external intervention to replace a failed node. Meanwhile, reactive approaches might
not be suitable for time-sensitive applications. The autonomous and unattended nature
of WSANs necessitates a self-healing and agile recovery process that involves
existing actors to mend the severed inter-actor connectivity by reconfiguring the
topology. Moreover, though the possibility of simultaneous multiple actor failure is
rare, it may be precipitated by a hostile environment and disastrous events. With only
localized information, recovery from such failures is extremely challenging.
Furthermore, some applications may impose application-level constraints while
recovering from a node failure.
In this dissertation, we address the challenging connectivity restoration problem while
maintaining minimal network state information. We have exploited the controlled
movement of existing (internal) actors to restore the lost connectivity while
minimizing the impact on coverage. We have pursued distributed greedy heuristics.
This dissertation presents four novel approaches for recovering from node failure. In
the first approach, volunteer actors exploit their partially utilized transmission power
and reposition themselves in such a way that the connectivity is restored. The second
approach identifies critical actors in advance, designates them preferably as noncritical
backup nodes that replace the failed primary if such contingency arises in the
future. In the third approach, we design a distributed algorithm that recovers from a
special case of multiple simultaneous failures. The fourth approach factors in
application-level constraints on the mobility of actors while recovering from node
failure and strives to minimize the impact of critical node failure on coverage and
connectivity. The performance of proposed approaches is analyzed and validated
through extensive simulations. Simulation results confirm the effectiveness of
proposed approaches that outperform the best contemporary schemes found in
literature.