BPSK_mod_and_demod.py

import numpy as np
import matplotlib.pyplot as plt

sample_time = 20
samples = 2**16 * sample_time
symbol_count = sample_time * 200
symbol_length = samples // symbol_count
frequency = 10000
time = np.linspace(0, sample_time, samples)

# 1 is 1 and -1 is 0
data = symbols = np.random.choice([-1,1], size=symbol_count)

# Now generate a sin wave
carrier_wave = np.sin(2 * np.pi * frequency * time)

# Spread the signal
symbols = np.repeat(symbols, symbol_length)

symbols = np.pad(symbols, (0, max(0, len(carrier_wave) - len(symbols))), 'constant')
# Now multiply
signal = carrier_wave * symbols

fig, (ax1, ax2) = plt.subplots(2)
ax1.plot(time, signal)
ax1.plot(time, symbols)

signal += np.random.normal(0, 5, len(signal))
# Decode
demodded = signal * np.sin(2 * np.pi * frequency * time)

linear = np.split(demodded[:symbol_length * symbol_count], symbol_count)
decoded = np.sum(linear, axis=1) / symbol_length * 2
decoded[decoded > 0] = 1
decoded[decoded < 0] = -1
decoded_extended = np.repeat(decoded, symbol_length)

correct = np.sum(np.equal(decoded, data))
error_rate = (symbol_count - correct) / symbol_count * 100
print("Error rate: " + str(error_rate) + "%")
# Integrate among the symbol lenght
ax2.plot(time, symbols)
ax2.plot(time[:symbol_length * symbol_count], decoded_extended)


plt.show()