Pornchai ChanyagornHarold H. SzuHongye WangMahidol UniversityGeorge Washington UniversityNational Institute of Mental Health2018-06-212018-06-212005-12-01Proceedings of the International Joint Conference on Neural Networks. Vol.3, (2005), 1735-17392-s2.0-33750112289https://repository.li.mahidol.ac.th/handle/123456789/16466This paper presents the application of collective swarm intelligence in solving problems of non-multiplexing communication in powerline sensor network. Based on collective intelligence behavior in nature, we design a system that allows local sensors to collect local information acquired from the powerline channel. As the information become available, the sensor can adaptively adjust its parameters in order to achieve a common goal of non-multiplexing sensor communication throughout entire network. The non-multiplexing sensor communications is implemented using underdetermined Blind Source Separation (uBSS) based on the finite alphabet property of the data. The details of hardware implementation are provided as well as the result from the test-bed network. © 2005 IEEE.Mahidol UniversityComputer ScienceConference PaperSCOPUS10.1109/IJCNN.2005.1556142