Pilot based channel estimation is the best way, and easiest way as well. As long as you understand your channel, if it is time or frequency varying, then you place your pilot accordingly and use least square algorithm in the receiver.

Look at papers emphasizing on pilot placing, and you can work out the rest from there.

Ahmed,

you must know all the items I described below to begin to understand channel estimation. Good luck.

Chanel Estimation

------------------------

Physical channels

----------------------------

A channel is the point of view of the propagation channel described in OSI 1 interface : Physical layer. Channels also described in OSI 2 (Data Layer) but it’s logical channels. This second channels are impacted with message with training sequences to tets the first layer. Then repeat same actions you do in GSM. In FDD or TDD you have to make measurements on the radio channel.

Measurement data

----------------------------

•Signal to Noise Ratio (SNR).

•Bit Error Rate (BER).

•Inter symbols Interference (ISI).

•Propagation time.

•Impulse Response (IR).

What to do ?

-----------------

•Channel equalisation to adapt transmission to propagation behavior.

•Choose the mean beam on Rayleygh diffusion to decide the right direct radio ray.

•Choose what channels to delete on the transmission way.

Methods

--------------

By sending methods

------------------------------

•Training base : downlink transmitted N samples signal known ==> Estimated Inpulse Response by receptor.

•Blind : downlink transmitted N samples signal unknown.

By analysis methods

-------------------------------

Time domain :

•Least square estimation.

•Singular Value Decomposition (SVD).

Frequency domain

•Discret Fourier Transform (DFT).

•Direct Cosin Transform (DCT).

http://www.nanotechinnov.com/lte-channel-estimation

thanks to Jean-Paul Cipria for his great cooperation and his clear and deep answer

Hi,

if it could help, I found these papers really interesting:

1. C. Mehlfuhrer, M. Wrulich, J. Ikuno, B. Colom, D. Bosanska, and M. Rupp, “Simulating the long term evolution physical layer,” 17th European Signal Processing Conference (EUSIPCO), pp. 1471-1478, Aug. 2009.

2. K.I. Pedersen, G. Monghal, I.Z. Kovacs, T.E. Kolding, A. Pokhariyal, F. Frederiksen, and P. Mogensen, “Frequency domain scheduling for OFDMA with limited and noisy channel feedback,” VTC-Fall, 2007.

i think that you have to start with frame structure of LTE. number of subcarriers, slot structure would very helpful to understand the channel estimation.