No abstract

A prototype system for monitoring, transmitting and processing performance data in sports for the purpose of providing feedback has been developed. During training, athletes are equipped with a mobile device and wireless sensors using the ANT protocol in order to acquire biomechanical, physiological and other sports specific parameters. The measured data is buffered locally and forwarded via the Internet to a server. The server provides experts (coaches, biomechanists, sports medicine specialists etc.) with remote data access, analysis and (partly automated) feedback routines. In this way, experts are able to analyze the athlete’s performance and return individual feedback messages from remote locations.

Detection of neuro-muscular fatigue in strength training is difficult, due to missing criterion measures and the complexity of fatigue. Thus, a variety of methods are used to determine fatigue. The aim of this study was to use a principal component analysis (PCA) on a multifactorial data-set based on kinematic measurements to determine fatigue. Twenty participants (strength training experienced, 60% male) executed 3 sets of 3 exercises with 50 (12 repetitions), 75 (12 repetitions) and 100%-12 RM (RM). Data were collected with a 3D accelerometer and analysed by a newly developed algorithm to evaluate parameters for each repetition. A PCA with six variables was carried out on the results. A fatigue factor was computed based on the loadings on the first component. One-way ANOVA with Bonferroni post hoc analysis was calculated to test for differences between the intensity levels. All six input variables had high loadings on the first component. The ANOVA showed a significant difference between intensities (p < 0.001). Post-hoc analysis revealed a difference between 100% and the lower intensities (p < 0.05) and no difference between 50 and 75%-12RM. Based on these results, it is possible to distinguish between fatigued and non-fatigued sets of strength training.

No abstract

Accurately determining resistance-training parameters is crucial to gain knowledge about the training process and to evaluate training interventions. To current knowledge, no method exists to automatically detect a series of features in a repetition training session using one three-dimensional accelerometer. In this study, a specific algorithm was used to detect the number of repetitions and different time features. Features determined by the acceleration algorithm were compared to a reference system using a linear wire encoder to detect movements. A total of 50 healthy participants were randomly assigned to three different groups (maximal strength, hypertrophy, and muscular endurance) and executed three different resistance-training exercises (bench press, leg press, and trunk flexion). Results of both measurement systems were compared for agreement using Bland -Altman plots, regarding repetition numbers, TUT (time under tension) in concentric, eccentric, and isometric contraction forms and total TUT (sum of all contraction forms) as well as break between repetitions. Both methods showed high agreement in repetition count (mean error 20.2^0.6 repetitions). Time features were detected with less agreement, with 10.0% disagreement for TUT in first phase, 1.1% disagreement for second phase, and 56% disagreement in isometric contraction. However, it was possible to detect a series of time-based movement features, enhancing the possibility to objectively record different parameters of a resistance training session. This will improve research in resistance training and also bring advantages in the training process for coach and athlete.