In a "perfect sensor world" a forest fire will be fighted with the newest techniques. If stationary sensors that are placed throughout an area with a high risk for forest fires measure a certain event that indicates a fire, this information will be transferred through a sensor network. The local firefighters are alarmed by a monitoring system and can immediately deploy their air-based mobile sensors to the area of interest to confirm the threat. The sensors follow a certain strategy to maximize their area coverage and observe the most important regions and communicate within their limited range to agree their actions. The police starts their helicopters to stream a video to the emergency management center and a simulation of the fire's future direction, that is based on weather sensors, allows to warn civil protection organisations hours in advance.
This is only one and only a rough case where a sensor network with in-situ sensors, that manages everything from data acquisition to service delivery, helps a certain user to deal with threatening situations. Other possible scenarios might be air or water pollution monitoring in which sensor networks are helpful.
Architecture and Development
Sensor World is a distributed sensor network simulation environment for above mentioned or similar scenarios. It simulates phenomena
detected by sensor networks, physical communication
between sensors and different sensor
behaviour. The distribution makes it possible to add further components and to run the simulation on separate machines. Futhermore the architecture offers the exchangeabiltity of simulation components by using interfaces which define how to integrate different simulation models for phenomenons, sensors or communication models. In order to link the different simulation components a communication layer has to be established to allow message transfer between each other. This is done by one important core fragment, the VirtualCommunicationLayer
Overall the architecture comprises of the following 4 components, which build together the simulation environment and are described more precisely:
is the connection between the others. "Virtual", because it simulates a communication that will, in reality, be built via many different communication methods and is heavily influenced by the environment of the real sensors and phenomenons. We decided to use a messaging system that is based on the Java Messaging System JMS. The concrete implementation at the moment is MantaRay
Within the project, a stand-alone prototype of a messaging system was implemented. A basic visualization is created after the a run of the program from the logging files and is encapsulated in the
of the basic simulation framework, we started to connect Sensor World with other SensorWeb
projects developed at Ifgi.
Sensor World Repository and how to start the simulation