Successful Reboot of High-Performance Sporting Activities by Japanese National Women's Handball Team in Tokyo, 2020 during the COVID-19 Pandemic: An Initiative Using the Japan Sports-Cyber Physical System (JS-CPS) of the Sports Research Innovation Project (SRIP)

Abstract

The COVID-19 pandemic has negatively impacted sporting activities across the world. However, practical training strategies for athletes to reduce the risk of infection during the pandemic have not been definitively studied. The purpose of this report was to provide an overview of the challenges we encountered during the reboot of high-performance sporting activities of the Japanese national handball team during the 3rd wave of the COVID-19 pandemic in Tokyo, Japan. Twenty-nine Japanese national women's handball players and 24 staff participated in the study. To initiate the reboot of their first training camp after COVID-19 stay-home social policy, we conducted: web-based health-monitoring, SARS-CoV-2 screening with polymerase chain reaction (PCR) tests, real-time automated quantitative monitoring of social distancing on court using a moving image-based artificial intelligence (AI) algorithm, physical intensity evaluation with wearable heart rate (HR) and acceleration sensors, and a self-reported online questionnaire. The training camp was conducted successfully with no COVID-19 infections. The web-based health monitoring and the frequent PCR testing with short turnaround times contributed remarkably to early detection of athletes' health problems and to risk screening. During handball, AI-based on-court social-distance monitoring revealed key time-dependent spatial metrics to define player-to-player proximity. This information facilitated appropriate on- and off-game distancing behavior for teammates. Athletes regularly achieved around 80% of maximum HR during training, indicating anticipated improvements in achieving their physical intensities. Self-reported questionnaires related to the COVID management in the training camp revealed a sense of security among the athletes that allowed them to focus singularly on their training. The challenges discussed herein provided us considerable knowledge about creating and managing a safe environment for high-performing athletes in the COVID-19 pandemic via the Japan Sports-Cyber Physical System (JS-CPS) of the Sports Research Innovation Project (SRIP, Japan Sports Agency, Tokyo, Japan). This report is envisioned to provide informed decisions to coaches, trainers, policymakers from the sports federations in creating targeted, infection-free, sporting and training environments.

Keywords: COVID-19; Japanese athletes; Tokyo 2020; handball; heart rate; infection control; performance enhancement; physical activity; proximity; stay active.

Figure 1

Real-time automated quantitative monitoring of social distancing with a motion image-based AI system, shown on Ubuntu 18.04.5 LST. This system identified participant positions using a function on OpenPose [17]. The distance between the players was calibrated based on known court geometry.

Figure 2

Body temperature chart line–scatter plot for all athletes. The darker grey shaded area is the camp period (15 days); thick vertical lines denote days on which PCR tests were performed; and the asterisk denotes the single febrile athlete recorded on day 3.

Figure 3

Results of online questionnaire regarding perception of PCR tests during the camp.

Figure 4

Heat map of area-specific proximity density on the handball court during training matches. The color map represents a 0 to 100% scaled value from areas where no proximity IDs were observed (0%, white) to areas where the most frequent proximity IDs were observed (100%, dark red) during the time window of interest.

Figure 5

Density plot of the duration for each proximity ID during match time and stretching time.

Figure 6

Transition of high-proximity distribution during stretching, before and after intervention. The color map represents a 0 to 100% scaled value where no proximity IDs observed (0%, white) to where the most frequently proximity IDs observed (100%, dark red) during the time window of interest.