A Research On The Rate Adaptive Algorithm Of Microwave Communication Link

Rate adaptation algorithm is widely used because of the unreliability of the wireless link in microwave communications. It will lead to a poor throughput if the MAC layer choose a lower data rate, otherwise it will make a high BER(Bit Error Rate) if the choosen data rate is too high. The rate adaptation algorithm works in MAC layer, and determines data rate and transmission mode through calculating and predicting the real-time channel quality. While in PHY layer, the different MCSs(Modulation and Coding Schemes) indicate different data rates, and the MIMO system which contains spatial diversity and spatial multiplexing decides the transmission mode. The purpose of rate adaptation algorithm is adapting to the transmission condition of the wireless link, which leads to a significant improvement of the radio resource utilization and the performance of the wireless network.The rate adaptation algorithms are divided into closed-loop rate adaptation algorithms and open-loop rate adaptation algorithms based on whether there is a feedback. Each of them has its own characteristics, the open-loop rate adaptation algorithms without changing the protocol are simple; the closed-loop rate adaptation algorithms estimate the channel more accurately. Even if in the same category, the different rate adaptation algorithms performance is different, there is no algorithm can perfectly adapt to all scenarios.This paper presents a new rate adaptation retransmission algorithm called AMRAL(Adaptive Multi-Rate and Aggregation-Level), which is compatible with AMPDU under IEEE802.11 n. The algorithm mainly includes four mechanisms: the prediction of channel quality, the adjustment of rate, the adjustment of Aggregation-Level, the A-RTS/CTS. AMRAL gets the distribution of lost packets in AMPDU by analyzing the Bitmap to estimate the change of channel quality, then adjusts the data rate with the AggregationLevel based on the expected trend of channel quality. AMRAL is special because it firstly proposes to adjust the aggregation length of AMPDU, which was not considered before but was essential. AMRAL also proposes to predict the trend of channel change with the Bitmap analysis, which can make the rate adaptation algorithm more reasonable.We have implemented AMRAL in NS-3, as well as Atheros and Mi RA. Atheros and Mi RA are rate adaptation algorithms which have been applied to hardware driver actually and used widely. In terms of the algorithm’s convergence, simulation results show that AMRAL could get to the appropriate data rate faster than Atheros and Mi RA. In terms of the adaptability to the channel, AMRAL also performs well, especially in the scenarios that contain long distance between the nodes. In terms of network throughput, simulation results show that AMRAL could improve the utilization of the channel effectively within transmission time.