Two neuro-fuzzy control schemes for a traffic-regulating buffer in wireless technology

Bluetooth wireless operates in 2.4-GHz Industrial Scientific and Medicine (ISM) frequency, which may interfere with other devices functioning within the same frequency band, such as WiFi. Furthermore, Moving Picture Expert Group (MPEG) variable bit rate (VBR) video demands larger and more stable bandwidth and may cause data loss and time delay as a result of the high variation in bit rate in Bluetooth channels with limited bandwidth. To address these issues, two new neuro-fuzzy schemes are developed to control the input and output of a buffer referred to here as the traffic-regulating buffer. Regarding the input of this buffer, a rule-based fuzzy scheme is introduced and supervised by a neural network technique as a master controller to monitor the arrival rate to the buffer. The output of the traffic-regulating buffer is observed by another rule-based fuzzy scheme and is supervised by a second neural network to monitor the departure rate. Computer simulation results demonstrate that the two proposed neuro-fuzzy models reduce standard deviation and excessive data loss, and they also show an improved picture quality as compared with conventional MPEG VBR video over a Bluetooth channel.