Exploiting independence to cancel interferences due to adjacent cells in a DS-CDMA cellular system

The use of spreading codes in a communication system provides inherent interference cancellation in the form of processing gain. This is not sufficient at times, since the quality of communications can degrade severely when the interfering signal has not been canceled effectively. Hence, additional interference mitigation schemes are necessary. In a cellular network employing direct sequence code division multiple access (DS-CDMA) system, interferences originate from various sources, including the neighboring cells. This is called adjacent cell interferences. Since this interference is independent of the source-of-interest, this paper uses this fact by applying independent component analysis (ICA) to cancel this interference. ICA is used as a filter to eliminate this interference after which any conventional detection scheme can be used. Numerical experiments indicate the performance gains achieved when using this structure. Additionally, we have simulated a soft lower bound that demonstrates the loss of performance due to this interference.