An agent-based scheme for interference cancellation in a DS/CDMA system

DS/CDMA has been proposed as a multiple access scheme suitable for future mobile communication systems. CDMA with a conventional matched filter detector is affected by the near-far problem where there are high-powered users. To support the high quality of service in a fading environment, many mechanisms using multi-user detection schemes are proposed. Based on these existing mechanisms and a distributed artificial intelligence technique, we propose a distributed multi-agent scheme to assist CDMA system managers with interference cancellation. The scheme consists of a collection of problem-solving agents, and there are three key modules including knowledge source, blackboard system, and control engine which were built in the scheme. Performance comparisons between our scheme and other heuristics were derived from random traffic patterns. Extensive simulations show that the proposed scheme significantly reduced the bit-error-rate (BER) in CDMA networks and dramatically improves the connection blocking performance