The paper deals with two-dimensional （2D） channel estimation of Orthogonal Frequency Division Multiplexing （OFDM） system ill slow fading wireless channel. We concentrate on two channel estimation schemes： Least S...The paper deals with two-dimensional （2D） channel estimation of Orthogonal Frequency Division Multiplexing （OFDM） system ill slow fading wireless channel. We concentrate on two channel estimation schemes： Least Square （LS）＋Weighted BiLinear （WBL） and LS＋Linear Minimum Mean-Squared Error （LMMSE） where the first method is proposed in this paper. After theory analysis and simulation in Typical Urban （TU） channel, we find that LS＋LMMSE achieves the optimal perform- anee by exploiting prior knowledge of channel whereas LS＋WBL, without requiring channel knowl- edge and with only half of the computational amount of LS＋LMMSE, approaches LS＋LMMSE in Bit Error Ratio （BER） performance when the distance of two adjoining pilot symbols along frequency direction is sufficiently small. This makes LS＋WBL very suitable for wideband wireless applications.展开更多
This article proposes a new lack-of-test based on the weighted ratio of residuals and variances for partially linear regression models. The large and small sampling properties of the proposed test are established. The...This article proposes a new lack-of-test based on the weighted ratio of residuals and variances for partially linear regression models. The large and small sampling properties of the proposed test are established. The testing procedure is illustrated via several examples. Simulation studies show that the testing procedures are powerful even in small samples. An application of the test to a real data set is presented.展开更多
基金Supported by the National Natural Science Foundation of China (No: 60496311).
文摘The paper deals with two-dimensional （2D） channel estimation of Orthogonal Frequency Division Multiplexing （OFDM） system ill slow fading wireless channel. We concentrate on two channel estimation schemes： Least Square （LS）＋Weighted BiLinear （WBL） and LS＋Linear Minimum Mean-Squared Error （LMMSE） where the first method is proposed in this paper. After theory analysis and simulation in Typical Urban （TU） channel, we find that LS＋LMMSE achieves the optimal perform- anee by exploiting prior knowledge of channel whereas LS＋WBL, without requiring channel knowl- edge and with only half of the computational amount of LS＋LMMSE, approaches LS＋LMMSE in Bit Error Ratio （BER） performance when the distance of two adjoining pilot symbols along frequency direction is sufficiently small. This makes LS＋WBL very suitable for wideband wireless applications.
基金partially supported by the Major Project of Humanities Social Science Foundation of Ministry of Education under Grant No.08JJD910247Key Project of Chinese Ministry of Education under Grant No. 108120+3 种基金National Natural Science Foundation of China under Grant No.11271368Beijing Natural Science Foundation under Grant No.1102021the Fundamental Research Funds for the Central Universitiesthe Research Funds of Renmin University of China under Grant Nos.10XNL018 and 11XNH107
文摘This article proposes a new lack-of-test based on the weighted ratio of residuals and variances for partially linear regression models. The large and small sampling properties of the proposed test are established. The testing procedure is illustrated via several examples. Simulation studies show that the testing procedures are powerful even in small samples. An application of the test to a real data set is presented.
