decode_signal: Determine which symbol is being transmitted.

In tom-christie/transmit: Transmitting information using Poisson processes and Bayesian inference

Determine which symbol is being transmitted.

Description

Decoder takes a list of spike times for each Poisson process. It knows that one of them is firing at a higher rate, and tries to determine which it is by comparing likelihoods of observations

Usage

decode_signal(codebook, signal_plus_noise, signal_power, noise_power,
  prior_distribution = c(), time_interval_in_seconds = 0.1,
  entropy_threshold = 0.1, return_entropy_trace = FALSE,
  return_posterior_trace = FALSE,
  return_posterior_at_stop_time = FALSE)

Arguments

codebook	constructed using construct_codebook function
signal_plus_noise	signal to be decoded, output by add_channel_noise
signal_power	spikes per second allocated to signal process
noise_power	spikes per second allocated to every process (including signal process)
prior_distribution	initial guess at the distribution of messages. Influences decoding times.
time_interval_in_seconds	the interval at which to perform inference
entropy_threshold	transmission stops after entropy reaches this threshold
return_entropy_trace	return posterior entropy for each time step
return_posterior_trace	return trace of posterior values for every time step
return_posterior_at_stop_time	returns posterior values at stop time

Value

list of data frames. Will list decoded symbol, stop_time (in units of interval), entropy threshold, posterior at stop time, and potentially other time series as specified in the function call

tom-christie/transmit documentation built on July 19, 2023, 12:52 p.m.