The reason to recruit only male in this study is due to the proportion of female FFs in the capital city of Korea which is about 5%. Also, due to the Covid-19, the fire agency reduced the employees at work so it was hard to recruit a real FFs. So, this study used a convenience sample of FFs who were in the range of Korean FFs with regard to relatively low fitness levels of VO₂max (below 42 ml/kg/min) and age (20-50yrs) [4]. Total 5 male participated in this study. Participants were all students from a local university, with the following characteristics: average age (M±SD): 27.8±3.1yrs, height (M±SD): 176.2±6.2 cm, weight (M±SD): 76±10.8 kg, and VO₂max (M±SD): 37.7±2.9 ml/kg/min. All participants provided written informed consent after a detailed explanation of the aims, benefits, and risks involved in the investigation. The study was initiated after explaining the experimental process to all participants in detail and obtaining their consent for participation and inspection regarding the conduct of various tests, according to the ethical principles of the Declaration of Helsinki, approved by the ethics committee of Yonsei University (IRB approval no: CR318031).

Using a test-retest design, five male participants undertook three trials, with 1-week washouts. Participants were asked not to drink alcohol or caffeine, avoid high-intensity exercises, sauna, and were encouraged to sleep for more than 8 hours before every visits. Inclusion criteria for the study were as follows: age between 20 and 50 years; VO₂max between 28 and 42 ml/kg/min; BMI between 18.5 kg/m² and 30 kg/m²; free of major diseases (Parkinson’s disease, cardiovascular disease, diabetes, hypertension, and hyperlipidemia); able to use a treadmill and a Kinect-based mixed reality device; on no hormonal or mental therapy; not engaged in regular moderate-to-high intensity training within the past year; nonalcoholic; on no dietary supplementation; and, non-usage of certain drugs (i.e., those influencing muscle metabolism).

During screening, the standing height (to the nearest 0.1 cm) of each participant was measured without shoes using an extensometer. Body weight (nearest 0.1 kg) was measured without shoes using a digital scale (InBody 720, InBody Co., Ltd, Seoul, Korea). Next, the participants rested for 5 min in a sitting position, and blood pressure was measured for each subject using a blood pressure monitor (BPBIO320S, InBody Co, Seoul, Korea). Aerobic fitness was assessed using a gas analyzer (Quark CPET, COSMED, Rome, Italy). The Bruce protocol was used for the graded incremental treadmill test for the screening [17]. Calibration was conducted 30 min before the test with a gas mixture and a 3-L syringe. The participants were equipped with a strap (Polar Team 2, Polar Electro Oy, Kempele, Finland), and were asked to walk on a treadmill for 1 min at 6 km/h to adapt to the device and as a warm-up. A respiratory mask was strapped onto the participant’s face, and gas leakage from the masks was checked before the test. Rating of perceived exertion (RPE) was measured using the Borg scale 6-20. This was measured after each exercise (every minute). The test was stopped when the participants met the following criteria: 1) participants were unable to continue, 2) VO₂ plateau, 3) HR above 95% of HRmax, 4) RPE > 17, or 5) RER ≥ 1.10 [18]. Qualified participants were asked to visit twice. All tests were conducted at a restricted temperature and humidity in the same room.

The second and third visits were initiated after a week of washout. The participants were informed about the device before the workout, and they started the workout after they fully understood. Researchers provided login service only and comments were not allowed. Heart rate (HR) and RPE were measured during, and after the exercise session using a Polar watch (Polar Watch M600, Polar Electro Oy, Kempele, Finland) and the Borg scale. The data from the Polar watch were later downloaded using the Polar Flow website and imported as an Excel sheet for analysis.

Finger-tip lactate was measured before and after the workout using a lactate testing device (LT-1730, Lactate Pro2, Arkray, Japan). After sterilizing the finger-tip, a puncture was made using a sterile lancet. The first drop of blood was wiped away. The second drop of blood was applied to an assay strip and inserted into the lactate-testing device. The same procedure was followed for the third session.

The workout consisted of a circuit of 18 stations comprising resistance and aerobic exercises that used all major muscle groups and movements that would be involved in fire task (e.g., ladder climbing and carrying a hose) [5, 19]. Body weights were used for internal resisted exercises (e.g., side step, push-up, and arm walking) and two 5 kg dumbbells were used for each of the externally resisted exercises (i.e., shoulder press, deadlift, and dumbbell row). The weight of the dumbbell was fixed due to occupational characteristic of FFs which does not need periodized programs like elite athletes. So, we aimed to make one optimal fitness program. Participants were informed regarding KMR and its contents before the workout, and no other comments were provided during the exercise. Each exercise was performed for 60 s, and 15-s recovery periods were provided after each exercise to prepare for the next exercise station. The repetition and the time of the fitness program are shown in <Table 1>. Repetitions were counted by KMR using a motion-detecting camera that has standard movement points for counting. The main screen displays users on the screen with AR environments which shows the points to be reached for the counts and informed the three appropriate posture using AR environments (e.g., straightened upper body, straightened neck, etc.). For example, when “squatting” using KMR, floor images indicate a starting point using the AR environment. The exercise began when the user was at the correct starting point, and the main screen display the manner in which way to move using the AR. In this way the AR and users interact, the MR environment is generated. This was caused by human movement, which creates an AR environment by interacting with the pre-AR environment (e.g., ‘good’ indicated by blue lights and ‘bad’ indicated by red lights). Additionally, the MR environment had sound effects to encourage or provide information to the users <Figure 1>.

The table shows the components of the circuit-training. Every exercise time duration was 60 s, and the rest was 15 s. All times are presented in seconds.

This study used a test-retest design in which the same participants conducted the same exercise twice. Data were first viewed using Microsoft Excel, and all statistical tests were performed using IBM SPSS Statistics version 25.0 (SPSS, IBM Corporation, IL, USA). Descriptive statistical analysis (mean ± SD) was used to describe the variables (characteristics, HR, RPE, finger-tip lactate, and compliance). HR was expressed as a percentage of HRmax using Fox’s equation [20] and trimmed into 20 s each. To evaluate the reliability within a specific day (e.g., visit 2 vs. visit 3), the results from two repetitions from the individual trials were used for Cronbach’s alpha (> 0.9, excellent reliability; 0.7–0.9, high reliability; 0.5–0.7, moderate reliability; 0.5 and below, low reliability; and >0.7 was considered acceptable) [21] and Pearson correlation coefficient (r >.60 was considered acceptable). Statistical significance of all data was set at p < .05. [22, 23].

Five participants participated in this study. At visits 2 and 3, the participants performed 95.9±2.7% and 101.1±2.5% of the total recommended repetition, respectively.

Through a test-retest design, excellent reliability was found for HR and RPE using Cronbach’s alpha <Table 2>. The reliability of HR was 0.967, and that of RPE was 0.969, which is near 1. According to the “items if deleted” data, every item would have a lower reliability if deleted. However, the third visit of subject number 5 showed a higher HR value than the total Cronbach a (item if deleted = 0.973), and the second visit of subject number 2 showed a higher RPE value than the total Cronbach a (item if deleted = 0.975).

The table describes the reliability of HR and RPE evaluated by Cronbach’s alpha. HR and RPE showed high Cronbach’s alpha values of 0.967 and 0.969, respectively. “Item if deleted” shows the value of each visit and the subject’s data. A score higher than Cronbach’s alpha indicates that the data has a lower value in the total data.

Pearson’s correlation coefficients (r) for participants 1 to 5 between visits 2 and 3 for the HR are shown in <Table 3> (r = .525 - .814, p < .001). The total data of V2 and V3 has a correlation of r = .946. The range of correlations in RPE is shown in <Table 4>, with r = .718 - .958 (p < .001). The total data of V2 and V3 has a correlation of r = .948. Which is highly correlated in both HR and RPE. Additionally, correlations can be seen in <Figure 2>; the flow chart of visit 2 and visit 3.

The exercise intensities are shown in <Table 5>. The exercise intensity for RPE was in a range of “moderate-to-high intensity” (13-14: moderate, M±SD = 13.75±1.19), and that for the HR was in a range of “vigorous intensity” (70-85% HRmax: vigorous intensity, M±SD = 72.89±7.18%) [24]. Lactate showed a high value in this study (M±SD = 24.82±4.56 mmol/L). Based on these results, this fitness program was considered to be of “moderate-to-high” intensity in male participants with a VO₂max range of 28-42 ml/kg/min. However, when subject number 2 in visit 2 was excluded, the intensity for RPE changed to high intensity (14-15: hard, M±SD = 14.04±0.30).

The intensity of the exercise program is shown above by HR, RPE, and finger-tip lactate. Heartrate and RPE were tested before, during and after the exercise and lactate were tested before and after the exercise. HR was calculated as the percentage by their maximal HR. The result shows that the intensity of the fitness program is expressed as moderate to high intensity for firefighters with low fitness level (28ml/kg/min-42ml/kg/min).