We all know that unlicensed bandwidths are critically saturated. I think that the migration towards smart antenna system, Massive MIMO processing technologies, and beamforming is great for innovation. However, since the interference is still out there in any environmental condition, why Cambium cannot adopt a Code Division Multiplexing over OFDM modulation scheme?
Actually, is not that recent intuition but, orthogonal codes adopted almost everywhere in telecommunication industry is the last step for broadband wireless communications. Cellular communication technologies like 3G and LTE are already on!
By using a different orthogonal code for each SM the co-channel interference could be seen as low power noise. I think there could be many possible applications in terms of sector configuration allowing hypothetical full frequency reuse scenario. Moreover, it could be possible to deploy several PTP links for massive backhaul data transmission with limited cross-interference.
I know it should represent a great investment for research and development but wireless technology is already there!
Go ahead Cambium! All possible multiplexing technologies we know nowadays are needed!
Just to take stock of the situation, by analyzing the problem from a mathematical point of view there are several limitations but it could still appear pretty interesting.
Let us suppose to have an OFDM symbol of period T (nothing to deal with frame length, we're speaking about signal physical generation), the OFDM subcarrier will take B=1/T bandwidth inside the channel. If we adopt a code on such a signal, the spectrum is spread due to high-frequency chips.
By supposing to have 4-chip code we get 16 possible overlapping communications but (and here the point becomes critical), spread spectra (new coded subcarriers) overlap in a non-orthogonal way. Each subcarrier will be 4 times wider than original modulation scheme No way to keep all subtones for transmission.
The result is a new OFDM modulator which sacrifices/decimates subcarriers to create orthogonal coded frequency spectra. Less precise estimation of the channel putting a good word for overlapped communication and real parallel streams on the same frequency channel.
Correct me if I am wrong but wouldnt using CDM or CDMA access techniques be a step backward i remember reading something on the subject a few years ago it pretty much stated that OFDM is far superior to CDM/CDMA i believe the argument was that while CDMA was able to handle more multiplexed signals in a cell, it was able to do so at the cost of lower spectral efficiency.
also so far as I am aware True LTE technologies use OFDM in the downlink and SC-FDMA in the uplink.
dont get me wrong both techniques have there Pro's and Con's but when it comes to multipath and cochannel interference OFDM is the best choice that is if your goal is to squeeze the most bitrate out of the channel bandwidth as you can.
well those are my thoughts on the subject anyway :)
I also found some papers about such a mixed technique and you're absolutely right mentioning LTE technology. I generalized too much about that.
Yes, CDMA is absolutely worst than OFDM from a spectral efficiency point of view and for sure when you are dealing with very dramatic interfered environment, a best spectral utilization is the most important thing. This is not true in general but only when the channel status is not ideal and you have to face with a non-uniform frequency response of the channel itself. As you can check SC-FDMA is a tradeoff, each user uses a specific set of subcarriers in order to avoid mutual interference coming from
As you can check, SC-FDMA is a tradeoff, each user uses a specific set of subcarriers in order to avoid mutual interference coming from many simultaneous transmission. Basically, SC-FDMA reduces the precision of channel estimation in favor of multiple access. but they are still "speaking the same language" we cannot overlap communication avoiding causing critical interference.
You said that the key point is focused on bitrate but this is not really fundamental on the single link (SM to AP). The point becomes critical when you have to aggregate clients in very dense sectors and probably you would like to aggregate data from them.
My "proposal" would be the adoption of less fine channel estimation (eventually 1/4 of actual 1024 subcarriers in high-end products) and the implementation of a spreading code for effective multiple parallel streams from clients. I think my proposal is close to what is known as Carrier Interferometry (maybe).