Measuring arrow impact points: Yu 2012 study

The authors, engineers at Pusan National University, propose a system for measuring automatically where an arrow lands, intended for quality control in manufacturing. They start from a simple idea: if the same arrow is shot several times by a machine under the same conditions, rotating the nock each time, how close the impact points are shows how well the arrow is made; and analysing holes in a paper sheet by hand is imprecise, means changing the sheet often and makes it hard to put the group into numbers. The system detects the arrow, travelling at about 275 km/h, as it passes through a 65 by 65 cm frame placed about a metre in front of the target, and calculates its coordinates. As an example the paper compares several different arrows, each shot three times with the nock rotated by 120 degrees: the arrow with the lowest dispersion index (0.2707, number 2) is judged the best.

Two line lasers are mounted on the frame, each facing a row of photodiodes on the opposite side. When the arrow passes, its shadow covers at most four photodiodes; from the voltages of these photodiodes a Mamdani fuzzy inference system, with 33 hand-written rules and five levels of shading, estimates the centre of the shadow. From the two centres, using similar triangles, the system finds the point where the arrow crosses the frame, without needing to know the shaft diameter. The arrow is launched by a machine about one metre above the ground, with an adjustable angle, about 39 metres from the target. As a quality index the authors use the distance of the three impacts from their mean point: where the mean point lies does not matter, since the machine can be zeroed, only how tightly the shots are grouped.

The paper reports no accuracy check: it does not compare the system's coordinates with holes in a paper target, nor does it give the error in millimetres, so it is not known how reliable the method is. The arrow example is minimal: three shots per arrow, a number of arrows that can only be read off the figure, no information on model, spine or weight, and a dispersion index given without units. Three impacts are too few to estimate an arrow's dispersion, and the test was not repeated. The text is in Korean, with the abstract and captions in English.

For archers the study matters only indirectly. It concerns how a manufacturer might measure the consistency of arrows leaving the factory, with a machine in place of the archer, and it follows the same logic as the checks many archers make by numbering their arrows and comparing where they land: an arrow that changes its point of impact when rotated on the nock differs from the others in some way (spine, straightness, weight or fletching). Here, though, the measurement is presented as an engineering idea and has not been validated, so it says neither which arrow is better nor by how much.