Kyudo draw, a computer model: Ohishi 2011 study

This is a four-page paper for a Japanese mechanical engineering symposium, and it contains no measurements on archers. The authors build a computer model that combines the Japanese kyudo bow with the archer's two arms, seen from the front, to study hikiwake, the phase in which the bow is opened by drawing the hands apart. The result is that the model can reproduce an opening movement: the figures show the bow bending and the arms spreading at successive instants.

The title speaks of optimising the movement, but no optimisation is done yet. The torques applied at the joints are chosen arbitrarily, and finding the best ones with a multi-objective genetic algorithm is left for future work. The stated aim is a tool for later comparing the different ways the kyudo schools open the bow, which are currently judged by experience rather than by numbers.

The bow is modelled as a beam undergoing large deformation, in two segments of different length to respect the asymmetry of the yumi: 1.32 m above the grip and 0.82 m below, 2.14 m in all, rigidly joined at the grip, which is a freely rotating pivot. The string is treated as massless and inextensible, and the mass of bow and arrow is concentrated at the grip. The arms are four rigid links, forearms and upper arms of 0.36 m each, with masses taken from body segment coefficients for Japanese athletes for a hypothetical archer 1.8 m tall and weighing 77 kg.

The shape of the bow is found with the Newton-Raphson method and the movement of the arms with the equations of motion, solved with the Runge-Kutta method in small time steps. The torques at the elbows and shoulders follow a formula with parameters set by the authors, and the simulation starts with the bow already raised and the hands in their starting position.

There is no comparison with a real archer: the model is not checked against measurements of movement, force or bow shape, and its parameters, from the stiffness of the bow to the joint torques, are set by hand. The figures show the opening taking less than a tenth of a second, far shorter than a real opening, which suggests the chosen torques serve to show that the model moves rather than to reproduce an actual technique.

The model is planar, so it cannot capture rotations of the torso or bow out of the frontal plane, and the hands, wrists and rest of the body are missing. This is preliminary work: the optimal movement promised in the title is not there.

For anyone who shoots, kyudo or Olympic recurve, this work changes nothing: it measures no archers and does not get as far as pointing to a better way of opening the bow. It is of interest to those following the modelling of shooting, as one of the few attempts to simulate the bending of the Japanese bow and the movement of the arms together during the opening, whereas the earlier models of the Japanese bow the authors cite had concentrated mainly on the release and the work of the bow hand.

In the registry it counts as a sign of a line of research, not as evidence on how the bow is opened.