The objective of this research is to identify the passive response of the rider’s body to translational and rotational random perturbations. A custom made bicycle mock-up equipped with a system of sensors has been developed, capable of measuring the linear accelerations, angular velocity and the rider’s force responses in all translational axes of all bicycle interfaces. The bicycle mock-up is driven by a hexapod that generates coloured noise perturbations in the range 0–10 Hz. Twenty four healthy male adults participated in this study and gave informed consent according to the guidelines of the ethical committee of Delft University of Technology. The responses of all subjects are represented in the frequency domain by means of frequency response functions. More specific, the interaction of the rider’s body at the seat, foot pegs and handlebars are expressed in terms of apparent mass and as seat-to-sternum transmissibility functions (STS). The apparent mass and STS transfer functions for the surge and heave motion suggest a simple underlying passive response system. For surge, a clear resonance peak was found at 2 Hz for nearly all interfaces and directions, whereas, for the heave motion a clear resonance peak at 5 Hz was found for the seat and handlebars and a resonance peak at 6 Hz for the foot pegs. The apparent mass of the pitch and yaw motion also suggest, to a certain extend, simple passive dynamics after 1 Hz characterised by resonance peaks at 1.8 and 2.3 Hz, respectively. Only the corresponding yaw STS transfer function showed a resonance peak around 2.3 Hz, as well. The sway and roll motion do not suggest simple passive dynamics showing similar trends in apparent mass characterised by an ever-decreasing gain and no resonance peaks. Finally, the surge, heave, pitch and yaw apparent mass transfer functions suggest that higher body mass in general yields higher peak magnitude and lower resonance frequency. This effect was most apparent for the surge motion.

Experimental data were obtained from riding a steer-by-wire bicycle on the open road while perturbing balance with impulsive forces at the seat post (lateral pertubations) as well as perturbing balance with impulsive torques at the steering assembly (steering pertubations). The experiments were conducted at 2.6–5.6 m/s covering both the stable and the unstable forward speed range. For the lateral pertubation experiments two conditions were explored; normal steering and reduced torque feedback steering. Three metrics are used to assess the effect of torque feedback on rider steer control and balance. Results failed to indicate any statistically significant difference between experimental conditions. Bicycle and rider mechanics have been modeled using the Whipple bicycle model extended with the rider inertia. A rider control model is developed that incorporates all of human's sensory pathways and includes a strategy to compensate for sensory dead time. The identified rider control parameters, stabilize the system and mimic realistic rider control behavior. From the results the importance of the torque feedback pathway is strongly indicated. Finally for the steering pertubations the rider control model is modified to account for the cocontraction mechanism. The model manages to approximate the rider measured response and simultaneously captures the significance of the intrinsic response. A high level of intersubject variability is exhibited. The hypothesis that this variability is in fact due to the modulation of admittance in the shoulder joint is strongly suggested.