Swimming with the X6-2mini
It seems most of the world's deadliest creatures reside in Australia. This means that Australians need to keep on top of their game. A user of the X6-2mini developed an analysis method for improving his swim technique. No doubt this keeps him a few strokes ahead of the crocs, the sharks, the Blue Ring Octopus, the Scorpion Fish, the Stone Fish, the Box Jellyfish....well, you get the idea.
Thanks to Simon of Sydney, Autralia for the following report.
Prepare and mount the X6-2Mini
The X6 is prepared for data recording in the normal way. Once it is recording it is placed in a balloon with desiccant to absorb condensation and mounted on the body using elastic athletic supports. It has been mounted on back, lower forearm and back of hand without hindrance or problems. Starting the device, making it water and condensation proof, and attaching to the body takes about 2 minutes.
Data recording is continuous for the duration of the training session. As most data during swimming is rhythmic and the device is not often vertical, place marks in the data stream can be recorded with a number of distinct vertical arm pumps.
Once the training session is over the data is downloaded to computer and uploaded into the R open software for data analysis.
Swimming is a combination of power and hydrodynamic technique. The power must be effectively applied to the water and the bodies&rsquo interaction with the water must not unduly retard forward progress. Data analysis consists of charting the accelerations and the calculation of some useful statistics and displays. The displays assist in relating changes in stroke style with measures of performance enabling an iterative stroke correction based on detailed knowledge of precise arm, hand or torso accelerations and orientation against performance statistics.
Retrieve data for a single lap.
The start of the lap is the point at which sagittal acceleration (long axis of X6) is at a horizontal maximum and occurs when the feet leave wall after push off. For the last lap the finish is when the saggital data goes deliberately vertical.
Using graph interaction in R or in software provided by GCDC the lap start and finish and the data between is subset out and made into a file or R data object.
A lap is 50 meters. Calculate lap time T and average velocity V using the time recorded by the X6.
Each lap is split into stroke and glide phases. The glide phase includes the push off and the glide to the wall prior to turning for the next lap. The time spent in the stroke phase is the time between first and last stroke as recognised by data analysis discussed below. It is assumed that the velocity during stroke and glide phases are the on average same and equal to the average speed V.
Analyse the X6 acceleration data
Analysis identifies repetitive stroke cycles in X6 data using techniques from the chaotic time series analysis to identify limit cycles. X6 data is embedded (or lagged) with itself for approximately one cycle or 20 records if sampled at 20 hertz. A singular value decomposition of the embedded data enables a filtering of noise by back calculation using around 20 significant eigenvectors. Greater or lesser smoothing can be achieved with more or less eigenvectors depending on needs and asthetics of visual display. Back calculation using the most significant eigenvector displays the principal mode cycle and a simple peak detection using a span around one to two cycles identifies each stroke. The principal mode cycle is shown in grey at the top of the figure with the cycle location identified with a red dot.
Count the number of strokes taken per lap N. Calculate distance travelled in stroke phase Ds and glide phases Dg by separating 50m into stroke and glide phases using time spent in each. (Ds = 50*Ts/T and Dg = 50*(T-Ts)/T). Ts time in stroke phase. T is total time for 50m. Time in glide phase is T-Ts. Calculate time taken and distance travelled for each stroke.
Charting of Results
Statistics of stroke length, time, and count per lap with lap times as shown in the chart can be used to judge performance. Stroke morphology from the chart can be compared with statistics to improve technique.
Stroke morphology can be read from the graphical display. This includes filtered acceleration traces from each axis (red, green, blue) of the X6-2, total acceleration (yellow) and the principal mode (grey). Also shown is a notional display of the orientation of the X6. The arm pointed up the pool places the X6 in the flat position with acceleration due to gravity show in as a maximal -1 in the green trace. As the arm cycles downwards at the beginning of the stroke cycle the long saggital axis of the device (blue) becomes increasingly negative and the arm also makes lateral (red) movements as the body rolls with breathing. These movements can be encoded into a graphic which can show fore-arm orientation as it may appear looking along the axis of the swimmer from behind. This is shown in the graphic at the bottom of the chart. Colour coding in this chart reflects the forward acceleration component. Blue through red shows strong forward to backward acceleration. Luckily there is no red!
Note the stroke change at 30m. The swimmer caught up on the person in front and changed stroke to slow down.
All calculations and diagrams use the R open software. R Development Core Team (2009). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL http://www.R-project.org.