How OFDM works

When high bandwidth utilization is not required, the most commonly used parallel transmission system uses frequency division multiplexing (FDM) modulation, but this method also has low channel utilization. In order to improve the channel utilization of FDM, the spectrum of each subchannel can be partially overlapped, and the receiving end uses a correlation filter to receive the corresponding subchannel signal during the code element period. As long as the other subchannel signals are orthogonal to this local correlation signal during the code element period, their influence can be eliminated. A typical set of orthogonal signals is {1, cosΩt, cos2Ωt, ... cosmΩt ... sinΩt, sin2Ωt, ... sinmΩt ...}. They form an orthogonal function set in [0, T] T = 2π/Ω. The signal sent by OFDM is modulated by such a set of orthogonal signals as subcarriers and a non-return-to-zero square wave with a code element period of T as the baseband code. The receiver demodulator also performs correlation operations with the transmitted signal in [0, T] to achieve demodulation. The modulation and demodulation principles are shown in Figures 1 and 2.

As shown in Figures 1 and 2: At the modulation end, the serial binary data to be sent is converted into M complex sequences through a data encoder (such as 16QAM), where D(m) = A(m)-jB(m). This complex sequence is converted by a serial/parallel converter to obtain M parallel codes (one frame) with a symbol period of T. The code type is a non-return-to-zero square wave. The M parallel codes are used to modulate M subcarriers to achieve frequency division multiplexing. The resulting waveform can be expressed by equation (1):

At the demodulation end, d(t) is correlated with a sine or cosine signal with a frequency of fm in [0, T] to obtain A(m) and B(m), and then the same data sequence as the transmitting end is restored after parallel/serial conversion and data decoding.