R/convolveUnitStep.R

In TIMP: Fitting Separable Nonlinear Models in Spectroscopy and Microscopy

convolveUnitStep <- function(k, x, irfpar) {
  m <- matrix(0, nrow = length(x), ncol = length(k))

  # save a little time by storing the unitstep results
  step1 <- matrix(0, nrow = length(x), ncol = length(irfpar))
  for (j in 1:length(irfpar)) {
    step1[, j] <- unitstep(x - irfpar[j])
  }
  for (i in 1:length(k)) {
    dec <- exp(k[i] * x)
    for (j in 1:length(irfpar)) {
      s1 <- if (j %% 2 == 0) -1 else 1
      s2 <- -1 * s1
      decj <- exp(k[i] * irfpar[j])

      m[, i] <- m[, i] +
        ((s1 * dec + s2 * decj) * step1[, j]) +
        ((-1 + decj) * unitstep(-irfpar[j]))
    }
    m[, i] <- m[, i] * exp(-k[i] * x) * (1 / k[i])
  }
  m
}