In order to improve the planning decisions of the urban environment and consider the well-being of the residents, an objective assessment of mental stress in the urban environment is needed. The aim of this study was to investigate the changes in the photoplethysmographic (PPG) and second derivative of PPG (SDPPG) signal parameters related to arterial stiffness during the mental stress assessment. The study was carried out on 15 female and 13 male subjects between the age of 25 and 60 years. The finger PPG signal was recorded with a transmission mode sensor during the subject's inactive state, followed by an arithmetic stress test. The PPG signals were post-processed and the statistically significant increases in an augmentation index related parameter PPGAI, in ‘aging index’ AGI, and in the SDPPG amplitude ratios b/a and decreases in d/a and the slope of the ascending waveform front due to induced mental stress were found ( $p<0.05$ ). The results indicate that the arterial stiffness was increased during the stress test. In summary, the PPG waveform parameters related to arterial stiffness could be considered for the mental stress assessment.

This paper proposes a novel method for physical fatigue assessment that can be applied in wearable systems, by utilizing a set of real-time measurable cardiovascular parameters. Daylength measurements, including a morning test set, physical exercise during the day, and an afternoon test set were conducted on 16 healthy subjects (8 female and 8 male). To analyze cardiovascular parameters for physical fatigue assessment, electrocardiography, pulse wave and blood pressure were measured during the test sets. The fatigue assessment questionnaire score, reaction time, countermovement jump height and hand grip strength were also measured and used as reference parameters. This study demonstrates that (i) the compiled test battery can selectively assess the rested vs. physically-fatigued states; (ii) the obtained linear support-vector machine, trained using the heart rate variability based parameter (F-score 0.842, accuracy 0.813) and pulse arrival time based parameter (F-score 0.875, accuracy 0.875) shows a promising ability to classify between the physically mildly fatigued and significantly fatigued states. Despite the somewhat limited study group size, the results of the study are unique and provide a significant advancement on the existing physical fatigue assessment methods towards a personalized and continuous real-time fatigue monitoring system with wearable sensors.

The aim of this study was to evaluate how heart rate variability (HRV) and reaction time differ when measured in the morning and in the evening of one physically exhausting day in order to explore if these parameters would be sufficient for estimating physical fatigue accumulated during one day. Five different experiment days with fixed schedules were conducted which consisted of measurement set in the morning, physical exercise during the day and measurement set in the evening. The total average reaction time was lower in the morning (228 ± 18 ms) compared to the values measured in the evening (257 ± 22 ms). Both assessed HRV parameters SDNN and RMSSD showed a tendency to decrease during the day (total averages were respectively 61 ± 6 ms and 40 ± 6 ms in the morning vs 37 ± 4 ms and 24 ± 3 ms in the evening). This decrease was more prominent in the heart rate recovery phase compared to the resting heart rate. The results of this study give promising results for new methods for estimating daily physical fatigue and could be a basis for multiple future studies.

The aim of this study was to develop an optimized physical activity classifier for real-time wearable systems with the focus on reducing the requirements on device power consumption and memory buffer. Classification parameters evaluated in this study were the sampling frequency of the acceleration signal, window length of the classification fragment, and the number of classification features, found with different feature selection methods. For parameter evaluation, a decision tree classifier was created based on the acceleration signals recorded during tests, where 25 healthy test subjects performed various physical activities. Overall average F1-score achieved in this study was about 0.90. Similar F1-scores were achieved with the evaluated window lengths of 5 s (0.92 ± 0.02) and 3 s (0.91 ± 0.02), while classification performance with 1 s were lower (0.87 ± 0.02). Tested sampling frequencies of 50 Hz, 25 Hz, and 13 Hz had similar results with most classified activity types, with an exception of outdoor cycling, where differences were significant. Using forward sequential feature selection enabled the decreasing of the number of features from initial 110 features to about 12 features without lowering the classification performance. The results of this study have been used for developing more efficient real-time physical activity classifiers.

Human activity measurement and classification has been hot research topic for several years. Most of the solutions are based on the mobile phones, however there are also some wearable device implementations that have very specific functionality. The aim of this paper is to propose a human activity recognition and fall detection solution that provides extra safety for people working in challenging conditions. The system is integrated with the monitoring solution that provides real-time information about all workers and raises automatically an alarm in case of accidents or abnormal conditions.

The aim of this study was to compare and evaluate multiple predictive equations for basal metabolic rate in order to choose the most suitable one for energy expenditure models. Eight different predictive equations were compared to each other using regression analysis and with the results of indirect calorimetry tests with 25 participants. Mifflin-St Jeor, Livingston-Kohlstadt and Henry-Rees predictive equations performed better than other formulas with Mifflin-St Jeor having the lowest RMSE of 175 kcal/day compared to the results of indirect calorimetry. The results of this study can be used to develop more accurate energy expenditure models.

Human activity recognition using wearable sensors and classification methods provides valuable information for the assessment of user’s physical activity levels and for the development of more precise energy expenditure models, which can be used to proactively prevent cardiovascular diseases and obesity. The aim of this study was to evaluate how maritime environment and sea waves affect the performance of modern physical activity recognition methods, which has not yet been investigated. Two similar test suits were conducted on land and on a small yacht where subjects performed various activities, which were grouped into five different activity types of static, transitions, walking, running and jumping. Average activity type classification sensitivity with a decision tree classifier trained using land-based signals from one tri-axial accelerometer placed on lower back and leave-one-subject-out cross-validation scheme was 0.95 ± 0.01 while classifying the activities performed on land, but decreased to 0.81 ± 0.17 while classifying the activities on sea. An additional component produced by sea waves with a frequency of 0.3–0.8 Hz and a peak-to-peak amplitude of 2 m/s2 was noted in sea-based signals. Additional filtration methods were developed with the aim to remove the effect of sea waves using the least amount of computational power in order to create a suitable solution for real-time activity classification. The results of this study can be used to develop more precise physical activity classification methods in maritime areas or other locations where background affects the accelerometer signals.

The aim of this study was to evaluate how the physical activity classification window length, accelerometer sampling frequency and the number of correlating features affect the classifier performance. It is important to study the effect of these elements in order to reduce the computational power and memory buffers needed for wearable systems, where classification is done in real-time. Three different window lengths (5 s, 3 s, 1s), sampling frequencies (50 Hz, 25 Hz, 13 Hz) and two feature sets (110 and 43 features) were tested and evaluated in this study. As a result, it was found that the classifier performed similarly with the window lengths of 5 s and 3 s and the sampling rates of 50 Hz and 25 Hz, but the results with the window length of 1 s or the sampling rate of 13 Hz were lower. No noticeable difference in classifier sensitivity was found by decreasing the number of features based on correlation.