Polar Codes

The establishment of information theory as an independent research field dates back to the pioneering works of Claude Shannon and others. While Shannon's noisy coding theorem (1948) not only established the theoretical limits on the rate of error-free information transmission, it also showed that, in principle, those limits could be approached by using channel coding. Although the noisy coding theorem asserted the existence of capacity-achieving codes, it did not provide a constructive approach on designing such coding schemes. More than fifty years later, in 2008 Erdal Arikan was the first to publish a coding scheme that provably achieves channel capacity on symmetric binary-input memoryless channels. Based on a phenomenon called channel polarization he discovered, Arikan coined the name polar codes for this linear coding scheme. In his landmark paper Channel polarization: A method for constructing capacity-achieving codes for symmetric binary-input memoryless channels, he not only established the asymptotic performance results, but also proposed a decoder of quasi-linear complexity.

Our current focus

Albeit their remarkable theoretical properties, most prominently their asymptotic performance, ongoing research focuses on improving the finite length performance as well as on evaluating and devising corresponding decoding schemes. Furthermore, establishing connections between polar codes and other closely related linear codes is of paramount interest.

Contact

Christopher Schnelling