Research On Beam Management Of 60GHz Millimeter Wave WIFI Distributed Network

With the increase of the number of mobile devices and the rapid development of various multimedia applications,mobile communication systems will inevitably require higher transmission rates,greater system capacity,and lower system latency in the future.However,the existing low-frequency system cannot meet such requirements,so the industry has turned its attention to the unlicensed 60 GHz mmWave(Millimeter Wave)band.Because with a very large system bandwidth,this band can easily provide data rates of up to gigabits per second.However,due to the high frequency of the mmWave and the great path loss,the high-gain directional beam is actually used to compensate the loss,so an efficient BF(beamforming)training technique is needed.In addition,directional links mean smaller interference range,and spatial reusing can be achieved by a proper scheduling method,thereby the system throughput can be further improved.The mmWave distributed network is a mesh network scenario proposed by the IEEE 802.11 ay working group.It enables APs(Access Points)communicate with each other directly and adopts a time division duplex mode.Thus the related mechanisms in the existing standard are no longer appropriate.Therefore this thesis is to study the BF training and slot scheduling in the 60 GHz mmWave distributed network.This thesis first introduces the channel access method and beam training scheme of millimeter wave WLAN(Wireless Local Area Network),and analyzes and calculates the training time.It introduces a architecture of the millimeter wave WiFi distributed network.Because the scene characteristics are different from that of traditional millimeter wave network,the existed standards have designed a new beam training mechanism and time slot allocation mechanism.Both of them are introduced in this thesis.Then,this thesis proposes an STA(Station)-to-multi-AP BF training mechanism for dynamic scenarios of mmWave WiFi distributed network.The training process is designed in detail,and the training time and bit overhead are theoretically analyzed and simulated.The simulation proves that compared with the original one-to-one individual training and group training,the proposed scheme can greatly reduce the training overhead and time.Furthermore,we assume new STAs appear randomly with a certain probability to do the training with APs and then transmit data in the actual scene.The simulation proves that system throughput can be effectively improved with STA-to-multi-AP training mechanism.Besides,this thesis proposes a time slot scheduling scheme based on interference coordination.After collecting related information,the controller calculates the interference and distributes the results to all APs.When allocating all the links of an AP to each time slot,the interference between the coexisting links is calculated to ensure that after adding the link to the slot the total number of data packets that can be transmitted by all the links in the time slot is the largest.Compared with the existing three-stages allocation scheme,the proposed solution can not only save the signaling and time overhead of negotiation between APs,but also improve the network throughput significantly and reduce the total transmission time.