In Non Orthogonal Multiple Access we user Successive Interference Cancellation at receiver side to detect the more power full signal and cancel it until reaching the desired signal based on special constellation,
why we don`t make a combined constellation at the receiver so it becomes able to cancel all undesired signals together therefore reduce the process time?
You can in principle detect all the signals, desired as well as interfering, jointly. That is essentially the best thing that you can do, but the complexity of doing so will be large. SIC has a lower complexity but performs equally well, at least in some scenarios.
This is because interference happens symbol by symbol. Adjacent symbols interfere more often. When multipath delays are comparable, symbols overlap. You generally do not expect overlap between signals which have extremely different multipath delays while passing through a wireless channel.
So, you observe the time when you are doing a decision logic on the type of bit/symbol that was sent from transmitter at the receiver and make a decision based on adjacent symbol interference. Combined interference cancellation does not make sense since the symbol times/bit times are extremely small for high bandwidth wireless channels.
Dear Prof.Bing-Yuh Lu,
Thank you. why should i check ?
if sampling rate is high will it become difficult to estimate the detection region correctly?
I think a optimal sampling frequency can reduce the noise naturally, and then we can speed up the computation of algorithms.
You can in principle detect all the signals, desired as well as interfering, jointly. That is essentially the best thing that you can do, but the complexity of doing so will be large. SIC has a lower complexity but performs equally well, at least in some scenarios.
Dear Ayman,
I don't think sampling frequency deals with your questions, while I think Dr. Björnson provided a good answer. SIC is proved to be capacity achieving, but it requires rates set a priori, Gaussian codewords, and packets with infinite length (you can find the proof in Huang, Papadias, Venkatesan, "MIMO Communications for cellular Networks", chapter 3, where a single antenna is considered). On the other hand, maximum likelihood detection can work well in some scenarios, but complexity grows exponentially.
we wrote a paper about this for random access networks. It has been accepted very recently in IEEE Trans. Comm., you can find it in the attached link.
Please let me know if the paper is of interest for you.
Regards,
Giulio
http://ieeexplore.ieee.org/document/7801023/
- Badges
- Science topic
- Similar topics
- Engineering
- Telecommunications Engineering