Research on Autonomous Distributed Wireless Ad Hoc Networks

点击翻译按钮获取中文摘要

Autonomous distributed wireless ad hoc networks are self-configuring systems formed by co-operating nodes without any established infrastructure. A key assumption is that any two nodes can communicate directly or using multi-hop relaying. By paying attention to multiple layers, i.e., physical, media access control, network, etc., several essential properties of wireless ad hoc networks are clarified. Major research results are summarized below.1.Symbol rate control and multi-hop transmission have the same ability to expand the service area at the loss of throughput. Numerical results revealed a significant difference between symbol rate control and multi-hop transmission such that only multi-hop transmission can increase the spatial spectral efficiency.2.In distributed control wireless ad hoc networks, there might occur conflicts in route selection among multiple routes if each node seeks for better route with better throughput in a selfish manner due to mutual interference. To analyze … More this behavior, game theory is found to be effective and is applied to evaluate the achievable throughput. It is revealed that when the transmit power is low, decentralized adaptive route selection may attain throughput near the capacity.3.CSMA/CA is widely used for media access control protocol of ad hoc networks. In order to meet the demand to increase the network throughput, multi-channel and channel bonding techniques are proposed to effectively utilize multiple existing channels. Numerical and simulation results indicate that under low-traffic conditions, channel-bonding technique can achieve a low delay, while under high-traffic conditions, the multi-channel technique improve both throughput and delay.4.TCP is widely considered to be a transport protocol for ad hoc networks. However, instable link quality of wireless networks may result in lower efficiency compared to wired networks. A new decoding method which makes Tornado-type codes effective to bit errors by exploiting error packets is proposed and the advantage of the proposed method is confirmed by experiments. Less