Polar codes attracted a lot of attention since they were introduced by Arkan in 2008 . They are the rst practical codes that are proven to achieve the channel capacity at in nite length. The eld of polar coding is an active eld of research, and one of its main topics is the code construction. The construction of polar codes involves nding the set of the most unreliable bit positions, usually called the Frozen set. The complement set is then used to transport information bits. In this thesis, we look at polar codes operating over the Additive White Gaussian Noise (AWGN) channel and we consider two topics: their construction, and their performance against Turbo and Low-Density Parity-Check (LDPC) codes. In the rst part, we start with some construction algorithms that are frequently used, and then we show a new construction algorithm and demonstrate its performance. For the second part, we review the state-ofthe- art turbo and LDPC architectures and we benchmark their performance against polar codes.