Linear Identification of a Combined System of an Iso-power Controlled Bicycle Ergometer and a Human

The purpose of this study was to identify a bicycle-human system during moderate exercise. Four male subjects performed pedalling exercises on an electrically controlled ergometer which had a feedback loop to maintain the power value given by the other computer. Work load was varied for 9.6 sec as a pseudorandom binary sequence (dt = 0.05 sec, 6 bit, 3 cycle) which enable us to calculate the frequency response and the transfer function of the system, after dead time correction. Three trials were performed with a set load of 40-80, 80-120, and 40-120 W, respectively. The transfer function could be described as the fourth-order rational function when calculated by least squares on time domain as follows: G(jω) = [0.46–0.0030(jω) + 0.00098(jω)² – 0.0000024(jω)³ + 0.00000052(jω)⁴]/[1.0 + 0.01(jω) + 0.0050(jω)² + 0.000013(jω)³ + 0.0000031(jω)⁴] The parameters of these transfer functions had no significant differences among the subjects nor the work load settings (p > 0.05). The characteristics of Bode diagrams showed the resonance on the average to be approximately 2.2 and 6.4 Hz. The resonance of 6.4 Hz was unstable. It is necessary for identification of a human body system to design a control system that tracks rapid changes of work load.