Cricket requires an umpire to signal different events in the game to a distant scorer/bookkeeper. The
signals are communicated with motions of the hands. For example, No-Ball is
signaled by touching each shoulder with the opposite hand, and TV-Replay, a request
for an off-field review of the video of a play, is signaled by miming the outline of a
TV screen.

The dataset introduced in Ko et al. (2005) consists of four umpires performing 
twelve signals, each with ten repetitions. The data, recorded at a frequency of 184Hz,
was collected by placing accelerometers on the wrists of the umpires. Each accelerometer
has three synchronous measures for three axes (X, Y and Z). Thus, we have a
six-dimensional problem from the two accelerometers. 60% of the data is used for training and use the rest as testing data. The series are 1197 long.

Relevant Publications
[1] Ko MH, West G, Venkatesh S, Kumar M (2005) Online context recognition in multisensor systems using
dynamic time warping. In: Proceedings of the IEEE international conference on intelligent sensors,
sensor networks and information processing (ISSNIP), pp 283288
[2] M. Shokoohi-Yekta, B. HuHongxia, J. Wang and E Keogh (2017) Generalizing DTW to the multi-dimensional case requires an adaptive approach.
Data Mining and Knowledge Discovery (31) 1, pp 131