Lucas Wittmann

Advisors

Maike Stöve (M. Sc.), Rebecca Lennartz (M. Sc.),  Prof. Dr. Björn Eskofier

Duration

11 / 2023 – 05 / 2024

Abstract

Technology and sports have been intertwined already for a long time. In recent years, for example, wearables like smart watches or heart rate monitors have helped athletes understand their body and performance better, making their training regimen more efficient and individually suited to them [1]. Those insights are based on the collection of a multitude of different biosignals from the wearables, for example, inertial measurement units, global positioning system sensors, blood volume pulse and electrocardiography (ECG) [2, 3].

Adding to that, machine learning (ML) also moved into this field, giving athletes more precise reports and predictions on their training regiment and recommendations on action like training plans or nutritional recommendations [4]. Those functionalities are usually integrated into sports and fitness apps like Runtastic, Strava, Runna or Enduco, to name a few. Despite this, a significant number of users tend to discontinue using a sports app shortly after they started [5, 6]. This aligns with Robinson et al.’s [7] previous findings that about 50% of participants abandon fitness programs within the first six months. Mustafa et al. [5] suggest that it is primarily the highly motivated individuals who persist in using health apps.

For that reason, sports and fitness application developers are trying to increase the motivation of their users by adding gamification features such as badges, leaderboards or progress bars [8]. However, accurately measuring the motivation of an athlete still poses a problem. Currently, motivation is usually measured by different kinds of questionnaires like the Sports Motivation Scale [9] or the Coach-Athlete Relationship Questionnaire [10]. While valuable, this approach has certain limitations regarding the measurement. For instance, potential response biases, limitations in capturing nuances, and getting an analysis parallel to the activity. Hence, it is crucial to find a way to precisely detect motivation continuously and objectively, for instance, by using biosignals and video data.

The goal of this master thesis is to investigate the effectiveness of ML in classifying data from two distinct soccer drills. One drill incorporates gamification elements to enhance player engagement and motivation, while the other omits those elements. This work will focus on analyzing biosignals, primarily eye tracking and ECG data, to determine their relevance in assessing participants’ engagement and motivation. Following this, the thesis also aims to identify the most influential features that contribute to accurately representing the two different scenarios in the context of a soccer drill. This research has the potential to provide valuable insights into how machine learning can be used to understand motivation and its aspects better.

References

[1] Tony Luczak, Reuben F. Burch, Edwin C. Lewis, Harish Chander, and John Ball. State-ofthe-art review of athletic wearable technology: What 113 strength and conditioning coaches and athletic trainers from the USA said about technology in sports. 15(1).

[2] Ryan T. Li, Scott R. Kling, Michael J. Salata, Sean A. Cupp, Joseph Sheehan, and James E. Voos. Wearable performance devices in sports medicine. 8(1):74–78. Publisher: SAGE Publications.

[3] Gobinath Aroganam, Nadarajah Manivannan, and David Harrison. Review on wearable technology sensors used in consumer sport applications. 19(9).

[4] Ting Wang and Jinkyung Park. Design and implementation of intelligent sports training system for college students’ mental health education. 12.

[5] Abdulsalam Salihu Mustafa, Norashikin Ali, Jaspaljeet Singh Dhillon, Gamal Alkawsi, and Yahia Baashar. User engagement and abandonment of mHealth: A cross-sectional survey. 10(2):221. Number: 2 Publisher: Multidisciplinary Digital Publishing Institute.

[6] Lynn Katherine Herrmann and Jinsook Kim. The fitness of apps: a theory-based examination of mobile fitness app usage over 5 months. 3:2.

[7] Jonathan Robison and Marc A. Rogers. Adherence to exercise programmes. 17(1):39–52.

[8] Paula Bitrian, Isabel Buil, and Sara Catalan. Gamification in sport apps: the determinants of users’ motivation. 29(3):365–381. Publisher: Emerald Publishing Limited.

[9] Luc G. Pelletier, Kim M. Tuson, Michelle S. Fortier, Robert J. Vallerand, Nathalie M. Briere, and Marc R. Blais. Toward a new measure of intrinsic motivation, extrinsic motivation, and amotivation in sports: The sport motivation scale (SMS). 17(1):35–53. Publisher: Human Kinetics, Inc. Section: Journal of Sport and Exercise Psychology.

[10] Nur Amirah Zaker and Vincent Parnabas. The correlation between coach-athlete relationship and motivation among universiti teknologi MARA (UiTM) shah alam athhletes. 7(1).