Design and implementation of an experimental platform for performance analysis in wireless sensor networks

Abstract: Analysis of transmission delay and throughput in wireless sensor networks (WSNs) is essential for the deployment of WSNs in real applications. Although researchers have already done a lot of theoretic analysis on transmission delay and throughput in WSNs, very few of them gave the real sensor test results. In this master thesis, an experimental platform for wireless sensor networks has been designed and implemented. The purpose of building this platform is to find out the systems and environment factors that have impacts on the transmission delay and throughput of single hop and multi-hop wireless sensor networks. With this platform, we can combine the theory analysis with real field tests and give more reliable analysis on the performance analysis problem. The platform is based on one of the most popular operating system for sensor networks –Contiki OS and the Maxfor CM 5000 sensor motes. On this experimental platform, several groups of simulation tests and field tests in single hop, multi-hop and multi-path sensor networks have been implemented. All of these tests are divided into several groups in order to test different system and environment conditions that impact the transmission delay and throughput. Through the tests, a large amount of information and results about the transmission delay and throughput of the wireless sensor networks has been obtained. To achieve the visualized result and help the analysis, the figures of transmission period time for every type of tests are drawn using Matlab. The results of simulation tests and field tests are compared. We found that the program used in COOJA simulator can be easily transplanted to the real sensor nodes. It is also figured out which test condition would have the effect on the transmission delay and throughput. Moreover, the throughput values of the wireless sensor networks under different conditions are given in the paper.