Merging time and date data

Hello,

I hope somebody can help me with this. I've uploaded a sample of my dataset below. The table to the left is the original data, the table to the right shows the way I would like to transform it.

As you can see on column A, date and time measures have been taken every 80 seconds with associated values for columns B to D. I would like to transform my dataset in the following way:

Intead of having date and time measures every 80 seconds, I would like to have them every 240 seconds (4 minutes). That means rows should be merged by three for the whole dataset. As for the values in columns B to D, that would imply summing them every three rows to a single row. I have left the formulas in the cells for the table on the right hand side for better understanding.

Is there a way ro automat this process for a whole dataset with thousands of rows?

Any input is appreciated!

Thanks for your reply @Special-K, but formula on G2 isn't working. Note that, for each time, its associated ACTIVITY_X, ACTIITY_Y and ACTIVITY_Z measures are calculated from an interval corresponding to: [previous time + 1 seconds to actual time].

Example: For the second line with time 6/19/18 10:41:20 values for ACTIVITY_X, ACTIITY_Y and ACTIVITY_Z correspond to a sampling interval of [6/19/18 10:40:01 to 6/19/18 10:41:20].

The first row of the output table will have 0 for ACTIVITY_X, ACTIITY_Y and ACTIVITY_Z values since I'm missing the values from the two previous time intervals, but that's no problem.