An automatic cycling performance classifier system based on the crank arm force measurement data

This paper describes the development of an automatic cycling performance measurement system with a Fuzzy Logic Controller (FLC), using Mamdani Inference method, to classify the performance of the cyclist. From data of the average power, its standard deviation and the effective force bilateral asymmetry index, a score that represents the cyclist performance is determined. Data are acquired using an experimental crank arm load cell force platform developed with built-in strain gages and conditioning circuit that measure the force that is applied to the bicycle pedal during cycling with a linearity error under 0.6%. A randomized block experiment design was performed with 15 cyclists of 29 +/- 5 years with a body mass of 73 +/- 9kg and a height of 1.78 +/- 0.07m. The average power reached by the subjects was 137.63 +/- 59.6W; the mean bilateral asymmetry index, considering all trials, was 67.01 +/- 6.23%. The volunteers cycling performance scores were then determined using the developed FLC; the mean score was 25.4% +/- 16.9%. ANOVA showed that the subject causes significant variation on the performance score.