I’m trying to understand what exactly the “start study” and “end study” nodes mean. In warmup dataset 1, in what sense does a “start study” have the “chemical stationkeeping”? Is it just assumed that if the start nodes have a type (CK/EK/HK) that it’s already stationkeeping? If so does that mean in dataset 1, that it’s initially drifting at timestep 0, then at the end of the dataset it starts stationkeeping? How do we know it’s stationkeeping if we don’t see data after it?
I must have some fundamental misunderstanding of how nodes/modes work, I can’t make any sense of this warmup data.
Your initial assumption was correct. The NK, CK, HK, and EK labels do refer to the mode that begins at the corresponding node, and SS nodes provide a way to indicate the first behavioral mode without assuming that the behavior begins (IK or ID) at the first time index of the study period.
Based on your question about initiating station-keeping at the end of the study period, I believe you are working with an outdated version of the training labels. The first version of the file did not have the correct time indices for IK nodes. This error has since been fixed, so you can download the current training labels file to view the correct indices.