On Improving Bit Error Probability of QPSK and 4-Level Amplitude Modulation Systems by Convolutional Coding

We consider Amplitude Modulation (AM) systems, PhaseShift-Keying (PSK) and Amplitude-Phase-Keying (APK) systems in which each symbol can take 2
ⁿ
values. After a systematic search of
(n, n - 1)
binary convolutional codes with large values of free distance and minimal number of code words at that distance, simulations have been performed with an 8-symbol modulation system (
n = 3
). For a bit error probability of less
2 \times 10^{-4}
, the saving in average power signal-signal-to-noise ratio over that required for an uncoded system with the same channel symbol rate was more than 2 dB with PSK and APK, and 1.5 dB with AM.