move.HSMM.mllk: Compute negative log likelihood of HSMM
In benaug/move.HMM: Fit HMM and HSMM animal movement models

This function, modified from Langrock et al. (2012), computes the negative log likelihood of the hidden Markov model.

1 2	move.HSMM.mllk(parvect, obs, PDFs, CDFs, skeleton, inv.transforms, nstates, m1, ini, useRcpp = FALSE)

`parvect`	The vector of parameters to be estimated
`obs`	A n x ndist matrix of data. If ndist=1, obs must be a n x 1 matrix.
`PDFs`	A list of PDFs for the dwell time and ndist observation distributions.
`CDFs`	A list of CDFs for the dwell time and ndist observation distributions.
`skeleton`	A list with the original parameter structure used to reassemble parvect
`inv.transforms`	A list of inverse transformations used to transform parvect back to the original scale
`nstates`	Number of hidden states
`m1`	a vector of length nstates that specifies how many states will be used to approximate each state of the HSMM (see Langrock and Zuchinni 2011)
`ini`	numeric value that specifies how the initial state distribution is calculated. 0 sets the initial distribution to the stationary distribution. If this matrix is not invertible, 1 sets the initial distribution for each state within each state agreggate to 1/m(state).
`useRcpp`	Logical indicating whether or not to use Rcpp.