User-centred design customisation of rugby wheelchairs based on the Taguchi method

Athlete's biomechanic propulsion is affected by the design of the wheelchair. It is possible to improve an individual athlete's performance by varying the key design parameters of the wheelchair to accommodate the athlete's preferences and requirements. This is particularly relevant in elite Paralympics sports. In order to customise the mechanical design of the wheelchair in an effective and informed manner and for the individual user and performance requirements; it is essential that the potential design space encompassing all possible design variations is narrowed down to a near optimal set of design parameters for the individual athlete. In this paper, we approach this problem using the Taguchi method. Primarily, this paper aims to show the considerations involved in applying the Taguchi method to an application involving repetitive tests with human subjects; it also provides an analysis for individual results for the group of athletes tested. A group of five elite wheelchair rugby athletes performed five sets of the 14m sprint test using a purpose built adjustable wheelchair on a wheelchair ergometer. Four design factors (A: Wheel diameter, B: Camber angle, C: Seat height and D: Camber bar depth) were each tested at different levels according to the standard L9 orthogonal array. Performance variables of mean acceleration (axp), push velocity (Vp), cumulative velocity (Vc) and times (tp) of the push phase and recovery time (tr) were analysed. The Taguchi method is applied to determine the contribution of each factor to particular performance outputs. The results are used here to define the optimum design space for wheelchairs used by each athlete respectively. A minimum correlation coefficient r of 0.294 was considered for assessment of significant factor effect on all performance variables. Significant factors levels and their possible combinations are used here to predict optimum performance. Near optimum combinations of design parameters are selected based on a defined performance parameter ranking and mechanistic and practical inferences on the true value of the effect statistic between factor levels. The best wheelchair design combinations are selected and assessed against athlete's feedback on comfort for final selection of the preferred design. By using the approach presented in this paper it is possible to customise the design of the wheelchair for particular performance requirements and athlete attributes