The validity, reliability & sensitivity of global positioning system inertial sensors to monitor training readiness in soccer

McMillan, Paul-Patrick (2015) The validity, reliability & sensitivity of global positioning system inertial sensors to monitor training readiness in soccer. MSc(R) thesis, University of Glasgow.

Abstract

Over recent decades there has been a progressive change in the volume and frequency of soccer matches at the elite level. Although the relative duration of each match has remained constant, the number of matches played per competitive season and the intensity of match play has increased. With this progressive change in volume of soccer competition from season-toseason at the elite level there now exists a rationale for monitoring player’s athletic readiness, with the aim to prevent conditions such as non-functional overreaching (NFOR) and overtraining syndrome (OTS).

Although the literature provides evidence for the utilisation of heart rate (HR) measures, subjective and psychometric questionnaires, hormonal and biochemical markers as well as neuromuscular markers to determine athletic readiness to train and/or compete, each has its respective limitations and all are subject to the extensive requirement of time. However, with technological advances it is now possible to break down temporal patterns of match play and training load to provide a better understanding of the overall physiological demands imposed upon elite soccer players. In light of this, the purpose of this thesis was to consider the current methods available to monitor athletic readiness, and subsequently to investigate the validity, reliability and sensitivity of utilizing a wearable global positioning system (GPS) based inertial measurement unit (IMU) to monitor vertical jump performance. This system could potentially provide an instantaneous marker of an athlete’s ballistic performance capacity and therefore aid in the prescription of appropriate loading or recovery strategies.

Three independent studies were conducted to investigate the validity, reliability and sensitivity of the IMU. In study 1, participants underwent a 5 -10 minute standard warm-up involving jogging locomotion at an exercise intensity eliciting 40-60% of maximum HR The participants were then required to perform in total 9 jumps on a Just Jump Contact Mat (JJCM) whilst wearing a Catapult MinimaxX 10Hz S4 GPS device placed on the trunk at the level of thoracic vertebrae 2-3 (T2-3) using a manufacturer issued harness. The JJCM acted as a gold standard reference. The 9 jumps, 3 countermovement jumps to a pre-determined depth (CMJPDD), 3 countermovement jumps to a self-selected depth (3 CMJSSD) & 3 squat jumps (SJ) were divided into three bouts. All jumps were performed with the hands positioned on the pelvis and the trunk upright.

In study 2, participants underwent the same warm up as in study 1 and then performed 3 jumps whilst wearing a Catapult MinimaxX 10Hz S4 GPS device placed on the trunk at the level of T2-3 using a manufacturer issued harness. Each participant performed a single CMJPDD, a single CMJSSD and a single SJ. Each participant conducted >2 of 3 jump testing protocols within 7 days. In the event that they did not meet this requirement they were subsequently excluded from analysis. All participants performed their jump protocol at the same time in order to mimic the proposed use for these investigations in the field. Inter-day reliability of jumps was then taken as a marker of system reliability. In study 3, each participant was required to perform a test battery on two occasions in a repeated-measures protocol, both before and after a short 6-week preparatory period leading into the 2014-2015 youth professional domestic soccer season. The test battery consisted of body mass measurement, 5m-sprint, 10m-sprint, 20m-sprint, free jump (FJ) and CMJPDD performance measures.

The results of study 1 suggest that the use of Catapult accelerometer power derived measures are not a valid indicator for vertical jump performance. On the contrary, jump height (H) as well as flight time (FT) appear to provide an accurate and valid measure of vertical jump performance when employing the techniques put forth in this manuscript; FTJJCM (s) was not significantly different from derived IMU FTP-T (s), for CMJSSD (0.634 + 0.046 vs. 0.636 + 0.042, p>0.05), CMJPDD (0.625 + 0.045 vs. 0.624 + 0.043, p>0.05) and SJ (0.616 + 0.049 vs. 0.612 + 0.004, p>0.05), respectively. HeightJJCM (cm) was not significantly different from derived IMU height based on FTP-T (cm), for CMJSSD (50.40 + 7.39 vs. 49.86 + 6.71, p>0.05), CMJPDD (48.60 + 7.18 vs. 47.93 + 6.73, p>0.05) and SJ (47.16 + 7.63 vs. 46.14 + 7.31, p>0.05), respectively (Fig 6).

Study 2 suggests that one jump condition has sufficient inter-day reliability with regards to FT and H estimation; the lowest mean coefficient of variation (CV) that we found was for height peak-to-peak (HP-P) for the CMJPDD condition (CV = 6.46%), with the other conditions up to two times higher and thus not supported in reference to the literature. Although this marker is not valid as shown in study 1 it is the only reliable measure from these investigations and therefore we would suggest that it warrants further research. Furthermore, as the HP-P estimation is shown to be reliable and the height estimation is based on flight-time, we would also suggest that further research for the utilisation of flight-time as a reliable measure when utilising flight-time peak-to-peak (FTP-P) is warranted to. FTP-T & HP-T for CMJPDD, CMJSSD & SJ were not shown to be reliable.

Study 3 revealed that in testing for a change in performance, there was no significant difference between before and after measures for 5m-sprint time (s) (1.08 + 0.05 vs. 1.08 + 0.05, p=1), 10m-sprint time (s) (1.83 + 0.07 vs. 1.81 + 0.06, p=0.211), 20-m sprint time (s) (3.08 + 0.07 vs. 3.06 + 0.08, p=0.355) and FJ (cm) (24.75 + 3.27 vs. 24.55 + 2.68, p=0.512), although there was a statistical difference for CMJPDD (s) (0.67 + 0.03 vs. 0.64 + 0.03, p=0.003). Accordingly, as there was no change in performance shown it was not possible to determine the sensitivity of the Catapult IMU as a means to monitor athletic readiness. Therefore we would advise using caution when implementing the system for the purpose of monitoring vertical jump performance, and suggest a change of 2 standard deviations (SD) as a benchmark for reason to question readiness to train until further more conclusive research can be conducted.

In conclusion, it is clear from these results that IMU flight-time and height measures provide valid markers of vertical jump performance when compared to the gold standard, when employing the techniques put forth in this manuscript; FTP-T & HP-T for CMJPDD, CMJSSD & SJ. However, none of these measures were shown to be reliable, and FTP-P & HP-P for CMJPDD were shown to be the only reliable measures. Additionally, it was not possible to accurately determine the true sensitivity of the IMU system to minute changes in training status for the CMJPDD, as there was no change shown in the performance markers tested in study 3. Future research should therefore aim to determine the true sensitivity of the Catapult IMU for estimating vertical jump performance, while reinvestigating the sensitivity and reliability of the system.

• Item Type: Thesis (MSc(R))
• Qualification Level: Masters
• Keywords: GPS, inertial sensors, soccer, training readiness, monitoring
• Subjects: Q Science > QP Physiology
• Colleges/Schools: College of Medical Veterinary and Life Sciences > School of Life Sciences > Life Sciences Human Biology/Sports Science
• Supervisor's Name: MacFarlane, Dr. Niall
• Date of Award: 2015
• Depositing User: Mr Paul Patrick McMillan
• Unique ID: glathesis:2015-6505
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 21 Jul 2015 14:35
• Last Modified: 22 Jul 2015 13:55
• URI: https://theses.gla.ac.uk/id/eprint/6505