LTE-A Downlink Channel Estimation And DSP Implementation

Because of its many advantages in low-latency, high-speed, high-capacity andeasy cloth mesh, LTE technology becomes the mainstream of the next-generationmobile communication standard. In order to meet the demand for high-end users, the3GPP begins deployment of an enhanced version of LTE, named LTE-A standard. Onthe basis of the LTE, LTE-A adds CoMP, enhanced MIMO, carrier aggregation andrelay technology. LTE-A can support a maximum100M bandwidth. Its downlink peakrate is up to1Gbit/s and uplink peak rate is up to500Mbit/s. LTE-A has excellentproperties. Therefore, the study of the LTE-A has very great significance.Wireless channels are complex, so channel estimation is very important. As theone of the key technologies of wireless communication, channel estimation techniquecan overcome time-varying fading characteristics of wireless channel and be againstthe inter-symbol interference. The main work and achievements of the thesis displaysas followed:This thesis first analyzes the wireless channel transmission characteristics and itsimpact. Based on the channel modeling, four test channels are studied, then it buildssoftware simulation platforms of the channels.Secondly, based on the baseband transmission of the LTE-A system model, thisthesis makes a brief exposition of the generation and mapping method of pilot andchannel estimation algorithms at pilot positions and non-pilot positions.Simultaneously, this thesis analyzes the principle of each algorithm and simulates thesealgorithms in the four test environments.Last, combined system performance and the complexity of the DSPimplementation, this thesis selects LS estimated channel estimation and first-orderlinear interpolation as channel estimation scheme of hardware. Then the channelestimation scheme is achieved and tested on the platform of Freescale MSC8156AMC.It introduces the process and main functions of the hardware implementation, datacalibration and complex multiplication on the fixed-point DSP implementations. Itindicates that the EPA channel is significantly better than TU6channel. Hardware results are presence of about1.0dB loss compared with MATLAB simulation results.Then based on the efficiency of the implementation of the function, the DSPimplementation channel estimation module is assessed. The result shows that performefficient channel estimation module is about0.29808ms. Far less than1ms, the channelestimation scheme meets LTE-A system requirements.