Event
Stream Processing in
Wireless Ad hoc Sensor Networks
Administrative
Data
Members
Project Summary
In recent years, technical advances in wireless communication have
enabled the development of many applications of sensor networks. In addition,
sensor nodes and small-scale embedded computing devices used in wireless sensor
and actuator networks (WSANs) have become
progressively intelligent with advances in processors, memories, and
communication facilities. WSANs can be employed in
diverse applications such as fire detection, environment monitoring, object tracking, and so on.
However, WSANs have some characteristics
that are distinct from those of ad hoc wireless networks, for example, limited
resources, large and dense deployment, dynamic topology, etc. Due to the
presence of low-rate and error-prone wireless links in the sensor networks,
communication overheads should be minimized, particularly when the sensor nodes
continuously produce a large volume of data. To achieve this goal, event processing systems should be designed so as to enhance
the flexibility and responsiveness of the infrastructure of a wireless sensor
network. Furthermore, the event processing systems should be able to increase
the efficiency of resource utilization, to enable on-demand computing and to
suppress the collection of unnecessary data from the physical world.
Most applications involving WSANs deal
with long-standing queries, which are tasks to be
accomplished in the networks. It is important to assign the tasks
correctly in the networks and perform the assigned tasks in an energy-efficient
manner since the sensor nodes can afford to use only a limited amount of energy
from a battery. To achieve this, WSANs need efficient
protocols for task assignment and load balancing.
¡¤
Goals
- Design of a Distributed Event Processing Framework
First, we identify use cases for the WSAN and requirements for the
use cases. Next, we design a novel distributed event
processing framework for the WSAN and define its functional entities.
- Development of a Task Assignment Scheme
Task assignment is of particular importance for designing the
distributed event processing framework for the WSAN.
We consider several designing issues. The first issue is regarding the way of
describing the task since the task description affects the complexity of the
WSAN. The second issue is regarding how to validate the task to be assigned toeach sensor, which is a critical problem for deploying
sensor networks. The third issue is regarding how to change the assigned
task into a new task. Although the processing of a long-standing query (or
task) in the WSAN takes a long time, some tasks should be
changed according to application requirements.
- Development of a Load-balancing Scheme
We consider long-lived queries in large-scale wireless sensor
networks with multiple gateways. In such an environment, sensory data are continuously delivered through the gateways therefore,
it is important to prolong the overall system lifetime by distributing the
sensory data through the multiple gateways. In this study, we design a task
assignment scheme that enables load balancing among the multiple gateways. In
addition, balanced energy consumption at the sensor nodes around the gateways is investigated.
- We identify key requirements for the distributed event processing framework in WSANs
from realistic use cases, and these requirements can be used as guidelines for
designing new event processing techniques.
- We propose a task assignment scheme in WSANs
which is meeting the given application requirements such as long
standing query and maintenance awareness. Furthermore, the proposed data
aggregation scheme seeks to satisfy the latency and reliability requirements. A
load-balancing scheme is also introduced for WSANs with multiple gateways, which improves system
availability as well as system lifetime.
- We implement a WSAN prototype to evaluate the proposed data
aggregation and task assignment schemes. Specifically, we will build up a
multi-hop WSAN that consists of Crossbow Mote sensor nodes.
Results
¡¤
International Journal
Jaekyu Cho, Yoonbo Shim, Sangheon Pack,
Taekyoung Kwon, Yanghee Choi, and Sooyeon Kim, ¡° SARIF:
A Novel Framework for Integrating Wireless Sensor and RFID Networks,¡± IEEE
Wireless Communications, December 2007.
¡¤
Technical Report
Sangheon Pack, Jongsu Lee, Taekyoung Kwon, Yanghee Choi, and Sooyeon Kim, ¡°A Dynamic Load Balancing Scheme in Multi-Sink
Wireless Sensor Networks.¡±
Related links
Last Modified: Tuesday, Dec 4, 2007