System Level Research On LTE/LTE-A Key Technologies

Long Term Evolution (LTE, also as 3.9G Radio network) is widely recognized as evolution of Universal Telecommunication Radio Access Network (UTRAN) and it adopts OFDM and SC-FDMA in downlink and uplink respectively. By enhancing 3G air interface, LTE improves the coverage and cell capacity while reducing packet transmission latency. Besides, LTE is also considered as the main stream technology envolved from 3G to 4G and commercial system has been deployed in some countries successfully. Based on the success of LTE, inorder to ful-fill the requirement of IMT-Advanced,3GPP starts LTE-Advanced (LTE-A) which is a smooth evolution from LTE. By introducing a series of key technologies:Cooperative Multiple point transmission (CoMP), Carrier Aggrevagation, Relay, and enhanced MIMO techniques, LTE-A can further improve spectrum efficiency and provides more flexible network configuration. This paper mainly studied CoMP and reference symbol design under enhanced MIMO technology.CoMP is about the coordination of signals transmitting and receving between mutliple points. Existed academical study shows that CoMP can significantly improve link level transmission signaling quality. However, it requires corresponding Radio Resource Management algorithm, inter-cell channel estimation and feedback schemes to get higher system level spectrum efficiency. Now, not many studies have taken realistic environments and LTE backward compatibility into account. This paper builds mathmatic models for CoMP in system level simulator. It concludes that CoMP is very sensitive to realistic environment.This paper has thress main contributions as following:(1) Build a complete system level simulator, propose new mathmatic models or improve the existed models for wireless system. Firstly this paper improved the "Wrap Around" model by proposing a computer searching based seamless simulation area generation scheme which generates seamless simulation area automatically. Secondly, we analyzed the RLC protocol established general RLP mathmatic model to describe the system performance impact from RLC retransmission. Finally, this paper studied LTE Uplink feedback channel assignment scheme, and proposed the empty feedback channel analysis model and give the optimized assignement scheme targeting at minimal system overhead. Simulator covers the main components for LTE system, and we further benchmark the simulation resutls with other company.(2) Targeting at the two main techniques of CoMP:Coordinated Beamforming and Joint process. This paper designed simple scheduler structure with backwards compatibility:keeping the original hardware structure unchanged. Then we designed geometry based transmission set selection scheme and we found out that intra-site CoMP can bring benefit to majority of users, thus it becomes the main study direction. After that, this paper designed the basic CoMP system model, which can illustrate basic performance gains under either ideal or realistic conditions. Finally this paper concludes that CoMP gives good performance gain under ideal assumptions.(3) Based on the systme level simulator and the basic CoMP performance results, this paper further studied several key components which might impact the CoMP performance under realistic environment. For inter-cell channel estimation error, this paper studied the CSI-RS muting, proposed that muting FR=3 and further designed more compact muting signaling method which can reduce half of the signaling overhead. For the X2 latency perspective, this paper concluded that more than 20ms latency will make CoMP of no gain at all, which actually make the inter-site CoMP impossible. And for the feedback perspective, we concluded that CoMP is very sensitive to quantized channel state information feedback and doesn't have any gain in such case.Based all study results, this paper gives the overview suggestion:CoMP can provide obvious performance gain under ideal assumptions. However, under realistic environment, the performance gain is quite moderate, further enhancement is needed. As a result, this paper proposes that it's too early to start CoMP standardization work right now.