Energy-Efficient Digital Signal Processing for Fiber-Optic Communication Systems

Doktorsavhandling, 2019

Fiber-optic communication is a keystone of our global communication networks, and is vital to support the modern digital society. In these systems, we use lasers to transmit our data, and we detect the changes in the light at the opposite end of the fiber. This allows us to transmit and receive data very fast and efficiently. However, as the data travels through the transmitter, the fiber, and the receiver, it gets smeared out and misshaped, and noise is added. Due to this, it is hard to figure out what was originally transmitted, and we need to use electronic circuits that attempt to fix the signal.

Since we want to send our information very fast, we also need to be able to correct these issues in the received signal incredibly fast. The electronic circuits apply math to try to twist the data back into its original shape, as well as try to use some added data to figure out where noise may have caused mistakes. However, the math is difficult to process and the circuit can require quite a bit of electrical power. This is an issue, as there are many of these links and the systems are rather densely packed. Power and energy efficiency is thus a very important concern in order to allow for transmission of more data in a sustainable manner.

In this thesis, we investigate ways to design fast and efficient electronic circuits that are good at correcting the errors and distortion of the signal that is transmitted, while still using a low amount of energy in the processing.