Real-Time Monitoring of the Heart Rate Response to Power Output for Cyclists

Every cyclist responds individually to a certain training intensity. This response can influence the effectiveness of predefined training schedules. In this study, real-time measurements of physiological responses that cyclist already use (e.g. heart rate, power measurements), were used in an algorithm that could accurately monitor different individual heart rate responses to training intensities. The developed algorithm used data-based model structures to estimate the relation between heart rate and training intensity recursively. Two different model structures were tested, namely an ARX model structure and an output error model structure (OE). The average RT2 of the simulated heart rates with the two different model structures were 0.86 ± 0.15 and 0.89 ± 0.13 respectively. This algorithm allows accurate modeling of the heart rate response to training intensity in real-time. In future research, such algorithms could be implemented in a body sensor network for training optimization. The real-time estimated model characteristics could then be used to calculate the needed training intensity during training in order to acquire the optimal training effect for an individual at any given moment.