Affordance-based control of braking in cycling: Experience reveals differences in the style of control

We investigated whether in an in-situ collision avoidance experiment cyclists regulate braking by adopting an affordance-based control strategy. Within an affordance-based control strategy for braking, deceleration is controlled relative to the maximum achievable deceleration rather than by nulling out deviations from ideal deceleration, and potentially allowing for different braking styles. Twenty active- and eighteen inactive-cyclists were asked to cycle on a straight path in an indoor gym and to stop as close as possible in front of a stationary obstacle. Maximum achievable deceleration was manipulated by loading the bike: no-load, load-5 kg, and load-10 kg. Two approach distances were used to vary cycling speed. Participants in both groups stopped farther from the obstacle when approaching with long- than short-initial distance conditions. No systematic effects of loading on braking performance and control were found across the two groups. However, both groups did increase the magnitude of brake adjustments as ideal deceleration increased and got closer to the action boundary, even when current deceleration approached the ideal deceleration. This indicates that participants adopted an affordance-based control strategy for braking. Two braking styles were identified: an aggressive style, characterized by a late braking onset and a high, steep peak in ideal deceleration, and a conservative style, characterized by an early braking onset and gradual, linear increase in ideal deceleration. The aggressive braking style was more prevalent among the active-cyclists. We suggest that the braking styles emerge from differences in calibration between information and action. The novelty of our work lies in confirming that cyclists adopt an affordance-based control strategy in an in-situ experiment and in demonstrating and explicating how affordance-based control can incorporate the emergence of different styles of braking.