JSensor – a high performance Java simulator for sensor networks

How can we analyze the water quality of a river? How can we monitor streets, cities, provinces of a country? How can we predict and prevent problems coming from earthquakes, snowstorms, fire spreads and all kinds of natural disasters? How can we test some new drugs without using animals? These and many other problems are initially solved by using simulators. It is extremely expensive, and sometimes dangerous, to test initial ideas in a real environment. We must use simulators to achieve comfortable confidence levels before real tests. This way, we develop simulation models where entities, named sensors, collect environment data, process data, communicate, move around the space and a lot more. JSensor is a high performance sensor network simulator implemented at Computer Science Departments of both universities UFOP and UFAL, Brazil. In contrast to JSensor counterparts, it is designed to simulate millions of sensors in multicore architectures.

JSensor simulator, its guides and example models (flooding, mobile phone and air condition, precisely) – link

There is the master thesis of Matheus Leônidas where all kernel and other simulator layers are detailed and exemplified, precisely at Section 3.4 – link

Two recent papers of the project:

  • Journal of Supercomputing – “Reproducibility Model for Wireless Sensor Networks Parallel Simulations” – DOI: 10.1007/s11227-020-03298-8
  • IEEE Transactions on Parallel and Distributed Systems (TPDS) – “JSensor: A Parallel Simulator for Huge Wireless Sensor Networks Applications” – DOI:  10.1109/TPDS.2019.2908845

Mention of Honor Award for the outstanding dissertation entitled Simplicity, reproducibility and scalability for huge wireless sensor networks simulations, authored by Matheus L. Silva and Joubert Lima, presented at Concurso de Teses e Dissertações (WSCAD-CTD – 2019).


Figure from http://www.isisingenieria.com/imagenes/productos_red.jpg