Motion analysis of kyudo shooting: Shoji 2013 study

This short conference paper describes the first step of a larger project: the authors want to develop a way of teaching kyudo that takes account of both the bow and the archer's body, and they start with the bow on its own. A carbon-fibre yumi with a hemp string was clamped in a vice by the grip, drawn and released, and the movement of its two tips was filmed and compared with a computer model.

Before release the upper tip had moved about 240 mm and the lower tip about 180 mm. The model gave 230 mm at the top and 220 at the bottom (in the text the authors put both at about 220): for the upper tip the calculation is close to the measurement, for the lower tip less so. After release the two tips vibrate with similar amplitudes; the upper tip lags the lower one in phase, and after 3 seconds all movement had stopped. In the calculation, by contrast, the lower tip starts with large oscillations that die away, while the upper tip moves little but keeps oscillating for longer. In the comparison the phase of the upper tip matches the measurement reasonably well, though its first oscillations do not; for the lower tip the initial amplitude follows a similar pattern but the phase differs. The authors conclude that a model made of rigid bodies and springs gives reasonable results, but that the segmentation and the spring constants still need refining.

The grip was fitted into a rectangular aluminium tube (120 by 35 by 20 mm, 3 mm wall) held in a vice. With the bow braced, the string sat 150 mm from the grip; it was pulled back parallel to its rest position with a horizontal force of 300 N and then released. The string was drawn to 700 mm from the grip; in the section on the calculation the draw is 700 mm from the string's rest position, which is 850 mm from the grip. The tips were filmed with a video camera whose frame rate is not given, and no arrow is mentioned anywhere in the text.

The model was built in RecurDyn, a programme for analysing the motion of linked bodies: the bow is divided into 18 rigid elements and the string into 3, joined by rotational springs, with a different cross-section for each element. The inputs are the spring constant, damping and density; in the calculation the string was drawn following the same procedure as in the test.

This is a five-page symposium paper, not peer reviewed, on a single carbon-fibre bow. It does not say how many times the test was repeated (the authors report good repeatability without showing data) or the camera's frame rate, which matters for reading fast vibrations; in the copy read for this record the values in the table of model constants are hard to make out. The bow was clamped in a vice, so the archer's hand is missing, and the hand is exactly what the project eventually wants to study; the authors also admit they did not examine the contact between bow and mount in detail.

The text never mentions an arrow, so the release appears to have been a dry release, which is not the same as a shot. The test draw is described as 700 mm from the grip and the calculated draw as 700 mm from the string's rest position, which is 850 mm from the grip: it is not clear whether the two were identical, and this could account for part of the gap between measured and calculated deflections. The paper is in Japanese, with an English abstract.

There is nothing here for an archer to take to the shooting line. The study shows that the asymmetric yumi does not bend equally above and below the grip (tip displacements of about 240 mm against 180 mm) and that, with the bow clamped in a vice, the two tips vibrate with different phases for about 3 seconds after release. How this movement changes when the bow is in an archer's hand, with an arrow on the string, is the question the authors leave to later work. It is a piece of modelling useful to people who study the Japanese bow, not a finding about technique.