Research On Cross-Layer Model For Wireless Packet Communication Systems

Global mobile communications and the rise of the Internet, making broadband, personal, and packet have become the trend of the times. How to provide more abundant and reliable multimedia services such as: data services, streaming media, video and audio broadcasting, VoIP, and so on, in the broadband wireless mobile networks is become research focus in wireless mobile communications areas. Our work is supported by the National High Technology Research and Development Program of China under Grant No. 2001AA123014, No.2003AA12331005 and National Science Foundation of China under Grant No. 60496310. Regarding to some issues in wireless packet communications systems, such as multi-carrier multi-service QoS guaranteed, multi-antenna of the multi-carrier multi-service QoS guaranteed, multi-user multi-service QoS guaranteed and support of wireless mobile efficient TCP / IP protocol stack structure, we make an in-depth studies. In this dissertation, there are five parts.In first part, we analyse the user evaluation system, extracting four evaluation indexes for the communications process of observation and experience on the user side. And in accordance with the user experience on the services differences, a four evaluation functions are proposed. User evaluation function maximization is a mixed process of evaluating varies traffics, and it is the general guidelines for cross-layer optimization on parameters and structures adjustment.In part II, we analyze the single-user multi-carrier multi-service system of cross-layer optimization problem. In determining such issues constraints, generally the service QoS constraints and resource constraints, maximizing the user evaluation function is the optimal objective. the allocation of resources derived algorithm. According to the differences of air interface resources, the best channel allocated to the most emergent packet algorithm is the optimal utilization of resources. According to the algorithm, the system schedules the packet to vary traffics based on user evaluation function,. It eventually allows user to maximize evaluation, and also protects the service QoS.In part III, we analyze the single-user multi-carrier multi-antenna multi-service system of cross-layer optimization problem regard to an OSTBC-OFDM system as a instance. Based on analyzing the relationship between OSTBC diversity gain and channel marix F norm, ans also the analyzing the latency characters, we design a kind of weight measurement for traffic packet. According to the weight of traffic packet, the requirement of each packet PSR is derived. Followed by the criteria in the second part, considering the channel matrix gain and the requirement of the packet PSR, the system can schedule the resource with maximizting the user evaluation function and also with serices QoS guaranteed.In part IV, we extend the theory of the third part, with considering multi-user scenario and defining the air interface resource as an multi-dimensional spaces without constraining on time, frequency and space. We design a two-stage scheduling structure under the restraints of the control strategy aiming to maximize the multi-user average evaluation function which named MUER criteria. The first-stage is user packet scheduling. The second-stage is inter-user scheduling. Control strategy affects multi-user evaluation function. It makes inter-user scheduling biasess when there is the competition existed among users. It can protect the users'balance and fairness with maximzing the users'average evaluation function and also with service QoS guaranteed.In part V, we analyze the structure of the traditional TCP/IP protocol stack, and propose an effective cross-layer air interface protocol stack for supporting mobile wireless communicating. The protocol stack for cross-layer ideas, put forward innovative mobile links (ML) and the concept of mirror TCP (mTCP), making the data link layer and the TCP/IP layer close coordination in air interface. With the PID designed, by increasing a few cable network's load as the cost, it avoids the HA to the FA tunnel transmission, reduce the delay on mobile IP layer. By ML signaling, mTCP maintains the MN's TCP windows state. It help the mTCP filters most of reverse TCP-ACK messages. And also, when TCP need retransmission, mTCP can retransmit the related packet but not the integrated TCP packet by cooperation with ML layer. It can save a volume of air interface resource.Part VI is the full text of summary and prospects. On the future developments in the field of cross-layer, we show a number of constructive ideas and methods, such as user strategy, the cross-layer optimization under partial channel information and heterogeneous network of cross-layer optimization.