CNN-based Speed Detection Algorithm for Walking and Running using Wrist-worn Wearable Sensors

Wrist-worn activity trackers have experienced a tremendous growth lately and studies on the accuracy of mainstream trackers used by older adults are needed. This study explores the performance of six trackers (Fitbit Charge2, Garmin VivoSmart HR+, Philips Health Watch, Withings Pulse Ox, ActiGraph GT9X-BT, Omron HJ-72OITC) for estimating: steps, travelled distance, and heart-rate measurements for a cohort of older adults. Eighteen older adults completed a structured protocol involving walking tasks, simulated household activities, and sedentary activities. Less standardized activities were also included, such as: dusting, using a walking aid, or playing cards, in order to simulate real-life scenarios. Wrist-mounted and chest/waist-mounted devices were used. Gold-standards included treadmill, ECG-based chest strap, direct observation or video recording according to the activity and parameter. Every tracker showed a decreasing accuracy with slower walking speed, which resulted in a significant step under-counting. A large mean absolute percentage error (MAPE) was found for every monitor at slower walking speeds with the lowest reported MAPE at 2 km/h being 7.78%, increasing to 20.88% at 1.5 km/h, and 44.53% at 1 km/h. During household activities, the MAPE climbing up/down-stairs ranged from 8.38–19.3% and 10.06–19.01% (dominant and non-dominant arm), respectively. Waist-worn devices showed a more uniform performance. However, unstructured activities (e.g. dusting, playing cards), and using a walking aid represent a challenge for all wrist-worn trackers as evidenced by large MAPE (> 57.66% for dusting, > 67.32% when using a walking aid). Poor performance in travelled distance estimation was also evident during walking at low speeds and climbing up/down-stairs (MAPE > 71.44% and > 48.3%, respectively). Regarding heart-rate measurement, there was no significant difference (p-values > 0.05) in accuracy between trackers placed on the dominant or non-dominant arm. Concordant with existing literature, while the mean error was limited (between -3.57 bpm and 4.21 bpm), a single heart-rate measurement could be underestimated up to 30 beats-per-minute.This study showed a number of limitations of consumer-level wrist-based activity trackers for older adults. Therefore caution is required when used, in healthcare or in research settings, to measure activity in older adults.

BackgroundFew studies have investigated the validity of mainstream wrist-based activity trackers in healthy older adults in real life, as opposed to laboratory settings.ObjectiveThis study explored the performance of two wrist-worn trackers (Fitbit Charge 2 and Garmin vivosmart HR+) in estimating steps, energy expenditure, moderate-to-vigorous physical activity (MVPA) levels, and sleep parameters (total sleep time [TST] and wake after sleep onset [WASO]) against gold-standard technologies in a cohort of healthy older adults in a free-living environment.MethodsOverall, 20 participants (>65 years) took part in the study. The devices were worn by the participants for 24 hours, and the results were compared against validated technology (ActiGraph and New-Lifestyles NL-2000i). Mean error, mean percentage error (MPE), mean absolute percentage error (MAPE), intraclass correlation (ICC), and Bland-Altman plots were computed for all the parameters considered.ResultsFor step counting, all trackers were highly correlated with one another (ICCs>0.89). Although the Fitbit tended to overcount steps (MPE=12.36%), the Garmin and ActiGraph undercounted (MPE 9.36% and 11.53%, respectively). The Garmin had poor ICC values when energy expenditure was compared against the criterion. The Fitbit had moderate-to-good ICCs in comparison to the other activity trackers, and showed the best results (MAPE=12.25%), although it underestimated calories burned. For MVPA levels estimation, the wristband trackers were highly correlated (ICC=0.96); however, they were moderately correlated against the criterion and they overestimated MVPA activity minutes. For the sleep parameters, the ICCs were poor for all cases, except when comparing the Fitbit with the criterion, which showed moderate agreement. The TST was slightly overestimated with the Fitbit, although it provided good results with an average MAPE equal to 10.13%. Conversely, WASO estimation was poorer and was overestimated by the Fitbit but underestimated by the Garmin. Again, the Fitbit was the most accurate, with an average MAPE of 49.7%.ConclusionsThe tested well-known devices could be adopted to estimate steps, energy expenditure, and sleep duration with an acceptable level of accuracy in the population of interest, although clinicians should be cautious in considering other parameters for clinical and research purposes.

BackgroundMeasuring physical activity with commercially available activity trackers is gaining popularity. People with a chronic disease can especially benefit from knowledge about their physical activity pattern in everyday life since sufficient physical activity can contribute to wellbeing and quality of life. However, no validity data are available for this population during activities of daily living.ObjectiveThe aim of this study was to investigate the validity of 9 commercially available activity trackers for measuring step count during activities of daily living in people with a chronic disease receiving physiotherapy.MethodsThe selected activity trackers were Accupedo (Corusen LLC), Activ8 (Remedy Distribution Ltd), Digi-Walker CW-700 (Yamax), Fitbit Flex (Fitbit inc), Lumoback (Lumo Bodytech), Moves (ProtoGeo Oy), Fitbit One (Fitbit inc), UP24 (Jawbone), and Walking Style X (Omron Healthcare Europe BV). In total, 130 persons with chronic diseases performed standardized activity protocols based on activities of daily living that were recorded on video camera and analyzed for step count (gold standard). The validity of the trackers’ step count was assessed by correlation coefficients, t tests, scatterplots, and Bland-Altman plots.ResultsThe correlations between the number of steps counted by the activity trackers and the gold standard were low (range: –.02 to .33). For all activity trackers except for Fitbit One, a significant systematic difference with the gold standard was found for step count. Plots showed a wide range in scores for all activity trackers; Activ8 showed an average overestimation and the other 8 trackers showed underestimations.ConclusionsThis study showed that the validity of 9 commercially available activity trackers is low measuring steps while individuals with chronic diseases receiving physiotherapy engage in activities of daily living.

This study presents a deep learning model devoted to the analysis of swimming using a single Inertial Measurement Unit (IMU) attached to the sacrum. Gyroscope and accelerometer data were collected from 35 swimmers with various expertise levels during a protocol including the four swimming techniques. The proposed methodology took high inter- and intra-swimmer variability into account and was set up for the purpose of predicting eight swimming classes (the four swimming techniques, rest, wallpush, underwater, and turns) at four swimming velocities ranging from low to maximal. The overall F1-score of classification reached 0.96 with a temporal precision of 0.02 s. Lap times were directly computed from the classifier thanks to a high temporal precision and validated against a video gold standard. The mean absolute percentage error (MAPE) for this model against the video was 1.15%, 1%, and 4.07%, respectively, for starting lap times, middle lap times, and ending lap times. This model is a first step toward a powerful training assistant able to analyze swimmers with various levels of expertise in the context of in situ training monitoring.

Obesity prevalence continues to increase globally, with increased physical activity considered as the most effective intervention. Fitness activity tracker have been shown to assist people increase their physical activity level, with increasing number of features and affordability over the past decade. An affordable fitness tracker, Archon Alive 001 was compared to Actigraph wGT3x-BT, with participants walking on treadmill at three, four, five and eight kilometers per hour (kph) and hand tally used as criterion measurement. Mean Absolute Percentage Error (MAPE) of Archon and Actigraph relative to hand tally at varying speeds, and related bivariate correlation were calculated, along with multivariate analysis of variance to assess the effect of speed and participant demographics on tracker MAPE. Polar OH1 heart rate monitor synchronized with PNOĒ metabolic analyzer was used to monitor heart rate and calorie expenditure. MAPE and correlation between PNOĒ and Archon calorie expenditure was calculated. Bland-Altman analysis and Intraclass Correlation Coefficient (ICC) was used to compare heart rate measurement between Polar and Archon. Compared to hand tally calculations, Archon step count was accurate, with average MAPE of 3.46%, and average correlation of r = 0.986, p < 0.001, while Actigraph average MAPE was 31.46%, with average correlation of r = 0.513, p = 0.001 at various speed. Age and speed had small effect on Archon’s accuracy. Sex had small effect, while speed had a large effect on Actigraph accuracy. Archon heart rate had Intraclass Correlation Coefficient (ICC) of 75.8%, with mean difference of -3.33 beats per minute (bpm) and 95% Limit of Agreement at -31.55 bpm to 24.90 bpm relative to Polar OH1. Archon total calorie expenditure output had 29.3% MAPE, and correlation r = 0.629, p < 0.001 relative to PNOĒ. Archon Alive demonstrated high accuracy in step counting, along with acceptable reliability and accuracy for heart rate and calorie expenditure measurements, making it sufficient for monitoring exercise intensity. Heart rate and calorie expenditure outputs were insufficiently accurate for clinical assessment.

BackgroundAortic valve stenosis affects approximately half of people aged ≥85 years, and the recommended surgical treatment for older patients is transcatheter aortic valve implantation (TAVI). Despite strong evidence for its advantages, low attendance rate in cardiac rehabilitation is observed among patients after TAVI. Cardiac telerehabilitation (CTR) has proven comparable with center-based rehabilitation; however, no study has investigated CTR targeting patients after TAVI. On the basis of participatory design, an exercise-based CTR program (TeleTAVI) was developed, which included a web-based session with a cardiac nurse, a tablet containing an informative website, an activity tracker, and supervised home-based exercise sessions that follow the national recommendations for cardiac rehabilitation.ObjectiveThis study aims to explore patients’ and health professionals’ experiences with using health technologies and participating in the exercise-based CTR program, TeleTAVI.MethodsThis study is a part of a feasibility study and will only report patients’ and health professionals’ experiences of being a part of TeleTAVI. A total of 11 qualitative interviews were conducted using a semistructured interview guide (n=7, 64% patients and n=4, 36% health professionals). Patient interviews were conducted after 8 weeks of participation in TeleTAVI, and interviews with health professionals were conducted after the end of the program. The analysis was conducted as inductive content analysis to create a condensed meaning presented as themes.ResultsReticence toward using the website was evident with reduced curiosity to explore it, and reduced benefit from using the activity tracker was observed, as the patients’ technical competencies were challenged. This was also found when using the tablet for web-based training sessions, leading to patients feeling worried before the training, as they anticipated technical problems. Disadvantages of the TeleTAVI program were technical problems and inability to use hands-on guidance with the patients. However, both physiotherapists and patients reported a feeling of improvement in patients’ physical fitness. The home training created a feeling of safety, supported adherence, and made individualization possible, which the patients valued. A good relationship and continuity in the contact with health professionals seemed very important for the patients and affected their positive attitude toward the program.ConclusionsThe home-based nature of the TeleTAVI program seems to provide the opportunity to support individualization, autonomy, independence, and adherence to physical training in addition to improvement in physical capability in older patients. Despite technological challenges, basing the relationship between the health professionals and patients on continuity may be beneficial for patients. Prehabilitation may also be considered, as it may create familiarity toward technology and adherence to the training.

BackgroundWearable activity trackers have become a viable business opportunity. Nevertheless, research has raised concerns over their potentially detrimental effects on wellbeing. For example, a recent study found that while counting steps with a pedometer increased steps taken throughout the day, at the same time it decreased the enjoyment people derived from walking. This poses a serious threat to the incorporation of healthy routines into everyday life. Most studies aim at proving the effectiveness of activity trackers. In contrast, a wellbeing-oriented perspective calls for a deeper understanding of how trackers create and mediate meaningful experiences in everyday life.MethodsWe present a study of real life experiences with three wearable activity trackers: Fitbit, Jawbone Up and Nike + Fuelband. Using need fulfillment as a theoretical lens, we study recent, memorable experiences submitted by 133 users of activity trackers.ResultsWe reveal a two-dimensional structure of users’ experience driven by the needs of physical thriving or relatedness. Our qualitative findings further show a nuanced picture of the adoption of activity trackers and their impact on wellbeing. For instance, while reflection about own exercising practices lost its relevance over time, users continued to wear the tracker to document and collect their runs. More than just supporting behavioral change, we find trackers to provide multiple psychological benefits. For instance, they enhance feelings of autonomy as people gain more control about their exercising regime. Others experience relatedness, when family members purchase a tracker for relatives and join them in their efforts towards a better, healthier self.ConclusionsThe study highlights that activity trackers can be more than “tools” to change behavior. Through incorporation in daily life, they offer new social experiences, new ways of boosting our self-esteem and getting closer to our ideal selves.

The purposes are to analyze the safety and real-time performance of intelligent transportation under big data and solve urban transportation safety problems. The existing problems of complex traffic data collection and single factor analysis are analyzed. Guangzhou target road section and its historical traffic flow are taken as the data source, the big data analytics (BDA) platform is constructed after preprocessing, and then the deep belief network (DBN) is introduced to build the intelligent transportation model based on BDA. Finally, the model is simulated to analyze its performance. The results reveal that the Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Mean Absolute Percentage Error (MAPE) of the proposed algorithm are 4.42%, 6.21%, and 8.03%, respectively compared to Recurrent Neural Network (RNN), Convolutional Neural Network (CNN), Neuro fuzzy c-means model (FCM), Deep Neural Network (DNN), and neural network (NN). The mean absolute error (MAE) and mean absolute percentage error (MAPE) are 4.42%, 6.21% and 8.03% respectively. The prediction accuracy is significantly better than that of other algorithms. The analysis of Accuracy, Precision, Recall and F1 suggests that the safety prediction accuracy of the proposed algorithm can reach up to 88.57%, which is at least 3.19% higher than that of other algorithms. Meantime, the proposed algorithm can effectively inhibit the spread of congestion, and achieve the effect of timely evacuation for traffic congestion. Regarding the data transmission performance, the delay is less than 40ms and the packet loss rate is less than 0.1, which is significantly better than other model algorithms in real-time performance. Hence, the constructed intelligent transportation algorithm based on BDA has high precision and excellent congestion evacuation performance, which can provide an experimental basis for constructing intelligent and safe transportation in the future.

Predicting discharge coefficient of triangular labyrinth weir using Support Vector Regression, Support Vector Regression-firefly, Response Surface Methodology and Principal Component Analysis

Reporting adherence, validity and physical activity measures of wearable activity trackers in medical research: A systematic review

PURPOSE: In the last decade or so, a plethora of fitness tracking devices have come to market. With this, use of these devices has increased exponentially. Among these devices’ many functions is the ability to monitor heart rate (HR). The purpose of this investigation was to determine if HR measured by the Suunto Spartan Sport watch was statistically comparable to that of our criterion, the Polar H7 HR monitor. METHODS: Twenty-one participants (male n = 11, female n = 10, age = 31±2 yrs [mean±SE], ht = 173±2 cm, wt = 76±3 kg) completed a two-mile trail run which included elevation changes of 48 m (McCullough Hills Trail, Henderson, NV), 55m (Three Peaks Trail, Cedar City, UT) or 104m (Bristlecone Trail, Mt Charleston, NV). Heart rate was obtained from the Suunto Spartan Sport watch with accompanying heart strap and the Polar H7 chest-worn heart rate monitor as the criterion reference. Validity was determined using the Mean Absolute Percent Error (MAPE), Bland-Altman analysis with accompanying bias and Limits of Agreement (LoA), and single measures Intraclass Correlations (ICC). A Pearson Product Moment Correlation Coefficient was used to determine the relationship between the validity measures with significance accepted at p<0.05. RESULTS: During uphill running, the MAPE was 1.46%, and the lower and upper LoA were -11.53 and 12.34, respectively. The single measures ICC was 0.95 with a 95% CI of 0.953 to 0.956 (F(14834, 14834) = 43.15, p < 0.000). During downhill running, the MAPE was 2.18% with the lower and upper LoA at -15.76 and 14.15, respectively. The single measures ICC was .94 with a 95% CI of 0.946 to 0.951 (F(9075, 9075) = 38.45, p < 0.000). CONCLUSIONS: These results demonstrate a there was a very low percentage of error for the Suunto Spartan Sport watch in recording HR for both uphill and downhill running. This suggests that The Suunto Spartan Sport might be a valid and reliable option for consumers wishing to monitor HR during outdoor trail running activities.

Background: Running is a popular form of exercise, and its physiological effects are strongly modulated by speed. Accelerometry-based activity monitors are commonly used to measure physical activity in research, but no method exists to estimate running speed from only accelerometer data. Methods: Using three cohorts totaling 72 subjects performing treadmill and outdoor running, we developed linear, ridge, and gradient-boosted tree regression models to estimate running speed from raw accelerometer data from waist- or wrist-worn devices. To assess model performance in a real-world scenario, we deployed the best-performing model to data from 16 additional runners completing a 13-week training program while equipped with waist-worn accelerometers and commercially available foot pods. Results: Linear, ridge, and boosted tree models estimated speed with 12.0%, 11.6%, and 11.2% mean absolute percentage error, respectively, using waist-worn accelerometer data. Errors were greater using wrist-worn data, with linear, ridge, and boosted tree models achieving 13.8%, 14.0%, and 12.8% error. Across 663 free-living runs, speed was significantly associated with run duration (p = .009) and perceived run intensity (p = .008). Speed was nonsignificantly associated with fatigue (p = .07). Estimated speeds differed from foot pod measurements by 7.25%; associations and statistical significance were similar when speed was assessed via accelerometry versus via foot pod. Conclusion: Raw accelerometry data can be used to estimate running speed in free-living data with sufficient accuracy to detect associations with important measures of health and performance. Our approach is most useful in studies where research grade accelerometry is preferable to traditional global positioning system or foot pod-based measurements, such as in large-scale observational studies on physical activity.

When measuring maximum strength, a high accuracy and precision is required to monitor the training adaptations. Based on available reliability parameters, the literature suggests the replacement of the one repetition maximum (1RM) by isometric testing to save testing time. However, from a statistical point of view, correlation coefficients do not provide the required information when aiming to replace one test by another. Therefore, the literature suggests the inclusion of the mean absolute error (MAE), the mean absolute percentage error (MAPE) for agreement analysis. Consequently, to check the replaceability of 1RM testing methods, the current study examined the agreement of isometric and dynamic testing methods in the squat and the isometric mid-thigh pull. While in accordance with the literature, correlations were classified high r = 0.638–0.828 and ICC = 0.630–0.828, the agreement analysis provided MAEs of 175.75–444.17 N and MAPEs of 16.16–57.71% indicating an intolerable high measurement error between isometric and dynamic testing conditions in the squat and isometric mid-thigh pull. In contrast to previous studies, using MAE, MAPE supplemented by CCC and BA analysis highlights the poor agreement between the included strength tests. The recommendation to replace 1RM testing with isometric testing routines in the squat does not provide suitable concordance and is not recommended.

Ubiquitous technology, e.g., smartphones or tablets, has created a continuously available digital world, drastically changing our feeling of being in the here and now – named presence. We thus increasingly shift between the real and the digital world, ranging from losing awareness of real surroundings to cutting out the digital world to truly be in the real one. In this work, we aim to explore the middle ground in between. We move beyond classic VR research on presence and look at everyday ubiquitous technology and its influence on presence in the real and digital world. By means of a focus group (N=6) and a subsequent online survey (N=36), we gathered individual notions of presence as well as experiences and situations in which people move between the worlds. We discuss the need to further explore presence and its dynamics across the real and digital worlds in everyday life.

The objective of this study was to investigate the development and tracking of nutritional status, body composition and physical fitness, and the longitudinal relationship of changes in nutritional status, and body composition with changes in physical fitness over a 1-year period of follow-up. Studied were 380 boys and 322 girls aged 7-14 years from the Ellisras Longitudinal Study. Boys and girls were divided into two groups of pre-adolescence (<11 years) and adolescence (>10 years). High tracking coefficients (>8) were found for nutritional status, body mass index, and fat-free mass, while low tracking coefficients (<4) were found for the sum of skinfolds, fat mass, arm muscle area, and central fat distribution. Moderate and low tracking coefficients were found for the physical fitness items. Longitudinal regression analyses showed that physical fitness performances that require a high energy flux over a short period of time are affected by muscle wasting, whereas having a low body weight appears to be important for a good performance on other fitness items in these malnourished rural South African children.