Data Coding Theory/Spectrum Spreading
Spread Spectrum edit
There are multiple ways in which multiple clients can share a single transmission medium. Time-Division Multiplexing (TDM) is the system where different clients get "time slices", periods of time where one client gets to transmit and other clients cannot. Since nobody else is transmitting, each client gets to use the entire spectrum, which means high bandwidth and a fast data rate. There is also Frequency-Division Multiplexing (FDM), where each client transmits at the same time, but on different frequencies. In FDM there is less available bandwidth, but each client gets to transmit for more time.
There is one more type that we need to discuss: Code-Division Multiplexing (CDM). CDM is also known as Spread-Spectrum multiplexing. In CDM, all clients can transmit at all times on all frequencies.
Other benefits to CDM, that we will discuss later, is it's use in encryption and the inability for another client to jam or interfere with a transmission.
Why CDM? edit
CDM, and the related CDMA, are technologies that were originally designed for use in the military. Many of the techniques described earlier in this book are designed to make optimal use of a transmission medium, in terms of higher bitrate while simultaneously minimizing the necessary bandwidth. CDM, however, takes the opposite approach of maximizing bandwidth in order to pursue other valuable qualities. CDM techniques specifically widen and flatten the transmitted signal, in some instances the signal becomes so wide and so short that it actually falls below the noise floor, and is impossible to discern. In addition, CDM signals rely on a special type of code called a PN code to decipher. These two qualities together mean that CDM transmissions are virtually immune to eavesdropping or interception.
Another benefit to CDM is that since the bandwidth being used is spread out, it is nearly impossible for another person to jam the transmission with interference. Also, since the transmission energy is spread out across a wide frequency range, the data signal can be indistinguishable from the background noise, which can prevent eavesdropping.
A third benefit, as if we need a third, is that CDM transmissions have the nearly magical property that we can simultaneously transmit multiple transmissions in the same time slice, and in the same frequency band, without ISI problems. This is how modern cellphone networks operate: multiple cell phones can talk at the same time in the same frequency band.
Direct CDM edit
Let's start out with an exercise. Let's say that we want to send some digital data, for instance 1011. What we will do is combine this data with another code, called the spreading code that is 3 times as long, and travelling 3 times as fast.
But wait a minute, if the spreading code is moving 3 times as fast (3 times the bit rate), then it will require 3 times the bandwidth to transmit! Well, this is correct. CDM requires more bandwidth to transmit a single signal than either TDM or FDM, but we will see in a minute why this doesn't matter to us.
Let's say that our spreading code is as follows: 101100111000. Now, we will combine our information signal (1011) with the spreading code as follows: We will logically XOR each digit of the information signal with 3 consecutive digits of the spreading code (because the spreading code is moving 3 times as fast. We will use the "X" symbol to denote the XOR operation:
1 0 1 1 X 101 X 100 X 111 X 000 ----- ----- ----- ----- 010 100 000 111
And now our resulting code is the value 010100000111. This will be called our transmission code. Now, we can pick one of our binary transmission methods(ASK, M-ary PSK, QAM, etc...) to transmit this resultant digital signal over the channel to the receiver.
To demultiplex a CDM signal, we make use of the following mathematical rule:
a X b X b = a
Remember, the "X" symbol here is the XOR operation. It would seem that if we XOR the transmission code with the spreading code, we will get back our original information signal! The only problem is that we need to have the same exact spreading code on the receiving end, to demultiplex the signal.
Benefits of CDM edit
So what exactly is the benefit of CDM? First, let's take a look at the spreading code:
Spreading Code edit
The spreading code is selected to have a number of different properties, but the most important is that the spreading code must be a pseudo-random number (PN). PN numbers look random, but since we must be able to generate the same signal on the receiving end, the spreading code can't be completely random. Because of this property, it can be very difficult for an eavesdropper to intercept our data, because other people do not necessarily know our spreading code. Even if other people did know our spreading code, they would not necessarily know what point in the code we were at. Some spreading codes can be very long.
Increased Bitrate edit
Our spreading code is being generated at a much higher bitrate than our data code is. According to Shannon's channel capacity, for the same SNR and a larger bit rate, our signal will be spread out over a much larger bandwidth.
Now, since we know that the Spreading code looks like noise, it is safe to assume that when we XOR in our information signal, the result will look like noise as well. In fact, the transmitted signal looks so much like noise, that unless the receiver has a perfectly synchronized spreading code to use in demultiplexing, the signal is impossible to recover. This makes CDM relatively secure.
Using CDM, we can transmit multiple signals simultaneously over exactly the same frequency band. This is important because TDM allows us to use the same frequency band but it separates out time slices. FDM doesn't need time slices, but it separates out signals into different frequency bands. CDM allows us to transmit signals at exactly the same time, in exactly the same frequency band.
Now, when we receive the multiplexed signals simultaneously, each receiver can XOR with the spreading code, to get back the original signal. Remember, each transmission signal looks like noise, so the receivers will simply ignore the signals that don't correlate with the spreading code. --Mangosrgr8 (discuss • contribs) 15:33, 4 December 2013 (UTC)I think this needs expanding; how does this (non)correlation occour? Why don't other signals (including 'real' noise) affect the transmission?
direct CDM (DSS) edit
Direct CDM, also called direct spread spectrum (DSS), multiplexes data using a trick with the digital XOR operation.
Examples of Direct CDM edit
Direct CDM techniques are used in Cellular phone transmissions in North America. This is because there are far too many cellphones to give each phone its own frequency range, or its own time slice. Using CDM, all the phones can simultaneously talk over the same frequency band at the same time.
Frequency Hopping CDM (FHSS) edit
Frequency Hopping CDM, also called frequency hopping spread spectrum (FHSS), transmits on only one narrow frequency band for a very short amount of time, then "hops" to some other frequency.
Let's say we have our PN generator, which is creating a spreading code in real time. We take a certain number of bits from that PN generator at a time, and use them to pick a frequency range. We then transmit a short burst of data on that range, and then hop to the next frequency range for the next burst of data.
If we have the same PN generator on the receiving end, we can then use that spreading code to pick which frequency bands to listen to at any given time. Using frequency hopping, our signal will take more bandwidth than normal (because we are using multiple bands to send information on. Frequency Hopping has several advantages including the fact that it allows us to make better use of more bandwidth simultaneously, and—if we hop fast enough—it provides some protection against reflections ("fading") and noise sources obliterating a few of the narrow bands.
Frequency hopping has the disadvantage the receiver cannot use PSK, QPSK, or M-ary PSK techniques, because it is just too hard to synchronize the different phase angles at different times on different frequency bands. However M-ary FSK is frequently used, with great success.
Fast and Slow Hops edit
A system is called a "Slow Hopping System" if the hop rate is slower than the symbol rate (an integer number of symbols transmitted per hop). A system is called a "Fast Hopping System" if the symbol rate is slower than the hop rate (an integer number of hops per each symbol). Fast hop systems are more complicated to implement, but are more immune to channel distortion.
An orthogonal frequency division multiplexing (OFDM) system is similar to a frequency hopping system in that part of the information is transmitted at one narrowband frequency, and other information is transmitted at a different narrowband frequency. However, rather than transmitting at only one narrowband frequency at a time, an OFDM system transmits at all of its frequencies all the time. Because any one narrow band it transmitted constantly, it's fairly easy to synchronize phase angles and use the best modulation technique for that band (QPSK). Because all the frequencies are generated by one transmitter, it is possible to pack them far more tightly together without interference than a system that dedicated each frequency to a different transmitter.
Reflections ("fading") and noise sources often obliterate a few of the narrow bands. To overcome this, the bits are "coded" (COFDM). A forward error correction (Data Coding Theory/Forward Error Correction) code is spread across every channel, such that even if 2 channels are completely obliterated, only 2 bits of the code word are lost. Then the receiver uses the FEC to reconstruct the obliterated data.
In practice, practically all OFDM systems are coded orthogonal frequency division multiplexing (COFDM) systems.