Optimizing compressive sensing in the Internet of things

In order to save cost of sensors in the process of transmitting information and gathering data, Compressive Sensing(CS), as a novel and effective signal transform technology, has been used gradually in the Internet of Things(IOT). This paper presents an optimizing method of CS in real applications of IOT. Compared to the traditional CS techniques that the sparsity of the signal need to be known, this algorithm can effective solve the actual issue that a lot of signals' sparsities usually could not be known in advance. According to experiments in the cases of both random distribution and Gauss distribution of signals, our algorithm is proved to be effective. Especially, when the amount of the sampling is 30 (the dimension of the data is 256) and the sparsity is unknown, the relative error is only 1.5%. © 2012 Springer-Verlag GmbH.