Problem in Channel Estimation using Least Square Approach

I have pilots which are either +1 or -1, (bpsk modulated). I am using OFDM
. After passing through rayleigh channel (used the function rayleighchan)
and than adding noise, i am trying to estimate the channel. Simplest way is
the least squared approach , do (Y/X) to get H_ls (least squared impulse
response in freq domain).
note : this is done after doing fft in ofdm.

The problem is how do i know what the pilot was? Its either +1 or -1, but
how do i know which one was actually sent. since the gains are complex, +1
sent could be received as -1 and -1 could be received as +1.

OR,
How do i simulate the least squared approach ? What should I take for X ?

shekar1989 <yogicyogesh@gmail.com> wrote:
> The problem is how do i know what the pilot was?

The point of a pilot is that you know what it is. "It's either +1 or -1"?
That's true for any BPSK-modulated signal!

Cheers,
Clemens

On Jun 30, 4:33&#4294967295;pm, Clemens Buchacher <dri...@aon.at> wrote:
> shekar1989 <yogicyog...@gmail.com> wrote:
> > The problem is how do i know what the pilot was?
>
> The point of a pilot is that you know what it is. "It's either +1 or -1"?
> That's true for any BPSK-modulated signal!
>
> Cheers,
> Clemens

I think the OP simply did not conceive that it is possible to
detect f a pilot signal if it has been subjected to a random phase
rotation ...

Julius