Application of fuzzy inference to CDMA soft handoff in mobile communication systems

The adaptation of soft handoff thresholds in direct sequence code division multiple access mobile communication systems using fuzzy inference is described. The soft handoff thresholds obtained from the proposed method are considered to depend on two inputs: the number of remaining traffic channels and the number of base stations connected to each mobile station. By evaluation of performance indicators, the fuzzy inference tends to give higher performance than the conventional soft handoff especially in trunkresource efficiency, call blocking, and carried traffic. Among fuzzy inference schemes considered, fuzzy inference with defuzzification using weighted average formula and fuzzification using bell membership function of the number of connected base stations and triangular membership function of the number of remaining traffic channels is the most compromised scheme when call blocking is not concerned. When call blocking is the serious indicator, fuzzy inference with defuzzification using weighted average formula and fuzzification using triangular membership function of both inputs should be selected. Regarding computational complexity, fuzzy inference with defuzzification using weighted average formula is the simplest.