Unobtrusive Sensors Research Articles

Ambient and Wearable Sensor Technologies for Energy Expenditure Quantification of Ageing Adults.

COVID-19 has affected daily life in unprecedented ways, with dramatic changes in mental health, sleep time and level of physical activity. These changes have been especially relevant in the elderly population, with important health-related consequences. In this work, two different sensor technologies were used to quantify the energy expenditure of ageing adults. To this end, a technological platform based on Raspberry Pi 4, as an elaboration unit, was designed and implemented. It integrates an ambient sensor node, a wearable sensor node and a coordinator node that uses the information provided by the two sensor technologies in a combined manner. Ambient and wearable sensors are used for the real-time recognition of four human postures (standing, sitting, bending and lying down), walking activity and for energy expenditure quantification. An important first aim of this work was to realize a platform with a high level of user acceptability. In fact, through the use of two unobtrusive sensors and a low-cost processing unit, the solution is easily accessible and usable in the domestic environment; moreover, it is versatile since it can be used by end-users who accept being monitored by a specific sensor. Another added value of the platform is the ability to abstract from sensing technologies, as the use of human posture and walking activity for energy expenditure quantification enables the integration of a wide set of devices, provided that they can reproduce the same set of features. The obtained results showed the ability of the proposed platform to automatically quantify energy expenditure, both with each sensing technology and with the combined version. Specifically, for posture and walking activity classification, an average accuracy of 93.8% and 93.3% was obtained, respectively, with the wearable and ambient sensor, whereas an improvement of approximately 4% was reached using data fusion. Consequently, the estimated energy expenditure quantification always had a relative error of less than 3.2% for each end-user involved in the experimentation stage, classifying the high level information (postures and walking activities) with the combined version of the platform, justifying the proposed overall architecture from a hardware and software point of view.

An IoT-based interoperable architecture for wireless biomonitoring of patients with sensor patches

The alliance between the Internet of Things (IoT) and healthcare has the potential to improve healthcare assistance at different stages of care through distributed vital sign sensing, paving the way for domiciliary hospitalization. In this work, we propose an innovative design for an IoT-based interoperable healthcare system to wirelessly monitor and classify patient status. To support our research, we identify gaps, and discuss standards, protocols and technologies based on works that use relevant IoT applications in healthcare. The proposed architecture is centered on several low-energy unobtrusive sensors attached to the patients’ bodies, as well as their beds, which encompass data acquisition nodes linked to a smart gateway that aggregates data. The smart gateway is integrated with an existing hospital information system through the exchange of Electronic Health Records (EHR), making relevant patient data easily available to health professionals on systems which are familiar to them. A use case scenario is presented in order to fulfill functional and non-functional requirements and provide a better understanding of connection and communication between the distinct entities of the proposed architecture, which is based on Bluetooth Low Energy (BLE) technology at the data acquisition level, the Message Queuing Telemetry Transport (MQTT) protocol at the internal level, and on the Fast Healthcare Interoperability Resources (FHIR) standard at the higher level.

Variability of the Center of Mass in Trained Triathletes in Running After Cycling: A Preliminary Study Conducted in a Real-Life Setting.

While the sport of short-distance (Sprint) triathlon provides an opportunity to research the effect of the center of mass (CoM) when cycling and running, much remains to be done. The literature has failed to consistently or adequately report how changes to hand position influence subsequent running as inferred by the magnitude of CoM acceleration. The demands of cycle training in a drops and aerodynamic position followed by running remain unquantified in Sprint Distance triathlon. Thus, far data collected indicate that the cycle to run transition (T2) is important for overall race success. While many age-groupers participate in Sprint Distance triathlon, the lack of T2 based research make comparisons between cycle hand position and ensuing running difficult. The motion of the human body when cycling and running in triathlon can be described by the motion of its CoM in a local coordinate system. Unobtrusive wearable sensors have proven to be an informative resource to monitor the magnitude of CoM accelerations in running. However, the extent to which they are used in cycling is unclear. Therefore, the aim of the present study was to analyse the temporal magnitudes of CoM acceleration when cycling position and cadence is changed and to analyse these effects on running after cycling. Ten recreational triathletes completed two 20 km cycling trials at varied cadence in a drops position (parts of the handlebars that curve outward, CycleDrops) and an aerodynamic position (arms bent, forearms parallel to the ground, CycleAero) immediately followed by a 5 km run at self-selected pace. Torso kinematics by way of CoM acceleration magnitude were captured in a typical training setting using a triaxial accelerometer. CoM acceleration was quantified in m/s2 and variability was measured by the coefficient of variation (CV) and root mean square (RMS). Results from CycleAero indicated that acceleration of the CoM in longitudinal (CV = 1%) and mediolateral directions (CV = 3%) was significantly reduced (p < 0.001) compared to CycleDrops. As for rate of perceived exertion (RPE), a significant difference was observed with triathletes reporting higher values in CycleAero alongside a greater CoM acceleration magnitude in the anteroposterior direction. The CoM varied significantly from RunAero with less longitudinal (CV = 0.2, p < 0.001) and mediolateral acceleration observed (CV = 7.5%, p < 0.001) compared to RunDrops. Although greater longitudinal acceleration was observed in the initial 1 km epoch of RunAero, triathletes then seemingly adjusted their CoM trajectory to record lower magnitudes until completion of the 5 km run, completing the run quicker compared to RunDrops (22.56 min1 ± 0.2, 23.34 min1 ± 0.5, p < 0.001, CV = 1.3%). Coaches may look to use triaxial accelerometers to monitor performance in both cycling and running after cycling.

Unobtrusive Measurement of Physiological Features Under Simulated and Real Driving Conditions

Objective: For driver state estimation, physiological features might be promising input parameters. As cable-bound sensing of these parameters is impractical for ubiquitous monitoring, the measurement certainly has to be based on unobtrusive and contact-free technologies. In this work, unobtrusive methods for heart rate (HR) and respiration rate (RR) monitoring, including a hybrid imaging approach, are evaluated under simulated and real driving conditions. Methods: The feasability of unobtrusive methods was tested by comparing measurements from unobtrusive sensors to reference sensors. Under laboratory conditions, magnetic induction and photoplethysmography, both integrated into the seat belt, and hybrid imaging, combining visual and thermal imaging, were evaluated for RR sensing. In real driving, creating an urban and a rural scenario, sensing of RR by hybrid imaging and sensing of HR by a seat-integrated capacitive ECG were evaluated. Results: Under laboratory conditions, a reliable RR detection was possibly using all three sensor technologies. In real-world driving, a reliable HR and RR detection was possible during the rural scenario. In the urban scenario, only the RR detection was feasible. Due to motion artifacts, the capacitive ECG was disturbed and the HR detection impaired. Conclusion: The evaluated unobtrusive measurement systems can monitor physiological parameters during e.g. long-time driving on highways, but may not yet be feasible for monitoring during agile inner-city driving situations, due to motion artifacts. Therefore, future work should focus on artifact reduction. Significance: Physiological features might be used as input parameters for driver state estimation systems. This work presents unobtrusive sensing methods for these parameters.

Detection of health deterioration in a COVID-19 patient at home: the potential of ambient sensor systems

The COVID-19 pandemic created increased interest in monitoring patients at home to allow timely recognition of health deteriorations. Hospital care is particularly demanding in these patients because of the necessity for isolation to avoid further spread of the disease. Therefore, home care is a preferred treatment setting for these patients. This is, to our knowledge, the first report indicating the potential of an affordable, contactless, and unobtrusive ambient sensor system for the detection of signs of health deterioration in a patient with COVID-19 by a caregiver from a distance. Prospective data acquisition and correlation of the data with clinical events were obtained from an 81-year-old senior with COVID-19 before and, in particular, over a period of 10 days prior to hospitalization. Clinical signs included weakness, increased respiration rate, sleep disturbances, and confusion. The visualization of a combination of this information on a dedicated dashboard allowed the caregiver to recognize a serious health deterioration that required a lifesaving hospitalization. The potential of such ambient sensor systems to detect signs of serious health deterioration in patients with COVID-19 opens new opportunities for use in asymptomatic or oligosymptomatic patients who live alone and are sent back to their homes for isolation in quarantine after diagnosis.

Design of Care Decision Support System Based on Home-Based Behavior of Elderly: A Design Science Study

With the development of Information Technology and Internet of Things, using unobtrusive sensors to monitor the home-based behavior of the elderly, and assisting the care givers to make care decisions based on this data plays an important role in ensuring the health and safety of the elderly living alone. Adopting the Design Science Approach, this study designs, implements, and evaluates a care decision support system based on home-based behavior of elderly. This system preprocesses the behavior data collected by sensors and divides it into Instantaneous Behavior data and Continuous Behavior data. Adopting Multivariate Gaussian Model and Topic Model, this system automatically provides the visualized results of overall analysis, baseline analysis, and long-term analysis. It can assist caregivers in finding early signs threatening elderly’s health and safety, and making care decisions. Three caregivers with more than 1-year relevant experience participates in the evaluation, and the results indicate that the system designed in this paper has more support effectiveness. This system provides a more effective tool of supporting caregivers making decisions for elderly living alone.

System Architecture of Unobtrusive Sensors for Supporting Home Care and Independet Living

The paper deals with an implementation of unobtrusive sensors installed in home environment for continuous monitoring of functional-health parameters of the observed persons. A multi-tier architecture links sensory devices through sensor-data concentrators to a home server. Automated sensory measurements are supported by a concept of sensoractivated events, event-driven data transmission and processing by a dedicated application interface. Its logic and data structures are revealed. Examples of three typical execution scenarios are given and a short description depicts clinical installation of proposed system for testing purposes.

High-Efficiency Multi-Sensor System for Chair Usage Detection.

Recognizing Activities of Daily Living (ADL) or detecting falls in domestic environments require monitoring the movements and positions of a person. Several approaches use wearable devices or cameras, especially for fall detection, but they are considered intrusive by many users. To support such activities in an unobtrusive way, ambient-based solutions are available (e.g., based on PIRs, contact sensors, etc.). In this paper, we focus on the problem of sitting detection exploiting only unobtrusive sensors. In fact, sitting detection can be useful to understand the position of the user in many activities of the daily routines. While identifying sitting/lying on a sofa or bed is reasonably simple with pressure sensors, detecting whether a person is sitting on a chair is an open problem due to the natural chair position volatility. This paper proposes a reliable, not invasive and energetically sustainable system that can be used on chairs already present in the home. In particular, the proposed solution fuses the data of an accelerometer and a capacitive coupling sensor to understand if a person is sitting or not, discriminating the case of objects left on the chair. The results obtained in a real environment setting show an accuracy of 98.6% and a precision of 95%.

Leveraging Unsupervised Machine Learning to Discover Patterns in Linguistic Health Summaries for Eldercare.

The Center for Eldercare and Rehabilitation Technology, at University of Missouri, has researched the use of smart, unobtrusive sensors for older adult residents' health monitoring and alerting in aging-in-place communities for many years. Sensors placed in the apartments of older adult residents generate a deluge of daily data that is automatically aggregated, analyzed, and summarized to aid in health awareness, clinical care, and research for healthy aging. When anomalies or concerning trends are detected within the data, the sensor information is converted into linguistic health messages using fuzzy computational techniques, so as to make it understandable to the clinicians. Sensor data are analyzed at the individual level, therefore, through this study we aim to discover various combinations of patterns of anomalies happening together and recurrently in the older adult's population using these text summaries. Leveraging various computational text data processing techniques, we are able to extract relevant analytical features from the health messages. These features are transformed into a transactional encoding, then processed with frequent pattern mining techniques for association rule discovery. At individual level analysis, resident ID 3027 was considered as an exemplar to describe the analysis. Seven combinations of anomalies/rules/associations were discovered in this resident, out of which rule group three showed an increased recurrence during the COVID lockdown of facility. At the population level, a total of 38 associations were discovered that highlight the health patterns, and we continue to explore the health conditions associated with them. Ultimately, our goal is to correlate the combinations of anomalies with certain health conditions, which can then be leveraged for predictive analytics and preventative care. This will improve the current clinical care systems for older adult residents in smart sensor, aging-in-place communities.

Energy-Positive Activity Recognition - From Kinetic Energy Harvesting to Smart Self-Sustainable Wearable Devices.

Wearable, intelligent, and unobtrusive sensor nodes that monitor the human body and the surrounding environment have the potential to create valuable data for preventive human-centric ubiquitous healthcare. To attain this vision of unobtrusiveness, the smart devices have to gather and analyze data over long periods of time without the need for battery recharging or replacement. This article presents a software-configurable kinetic energy harvesting and power management circuit that enables self-sustainable wearable devices. By exploiting the kinetic transducer as an energy source and an activity sensor simultaneously, the proposed circuit provides highly efficient context-aware control features. Its mixed-signal nano-power context awareness allows reaching energy neutrality even in energy-drought periods, thus significantly relaxing the energy storage requirements. Furthermore, the asynchronous sensing approach also doubles as a coarse-grained human activity recognition frontend. Experimental results, using commercial micro-kinetic generators, demonstrate the flexibility and potential of this approach: the circuit achieves a quiescent current of 57 nA and a maximum load current of 300 mA, delivered with a harvesting efficiency of 79%. Based on empirically collected motion data, the system achieves an energy surplus of over 232 mJ per day in a wrist-worn application while executing activity recognition at an accuracy of 89% and a latency of 60 s.

Contactless Gait Assessment in Home-like Environments.

Gait analysis is an important part of assessments for a variety of health conditions, specifically neurodegenerative diseases. Currently, most methods for gait assessment are based on manual scoring of certain tasks or restrictive technologies. We present an unobtrusive sensor system based on light detection and ranging sensor technology for use in home-like environments. In our evaluation, we compared six different gait parameters, based on recordings from 25 different people performing eight different walks each, resulting in 200 unique measurements. We compared the proposed sensor system against two state-of-the art technologies, a pressure mat and a set of inertial measurement unit sensors. In addition to test usability and long-term measurement, multi-hour recordings were conducted. Our evaluation showed very high correlation () with the gold standards across all assessed gait parameters except for cycle time (). Similarly, the coefficient of determination was high () for all gait parameters except cycle time. The highest correlation was achieved for stride length and velocity (). Furthermore, the multi-hour recordings did not show the systematic drift of measurements over time. Overall, the unobtrusive gait measurement system allows for contactless, highly accurate long- and short-term assessments of gait in home-like environments.

Sleep Diagnostics for Home Monitoring of Sleep Apnea Patients.

Objectives: Sleep time information is essential for monitoring of obstructive sleep apnea (OSA), as the severity assessment depends on the number of breathing disturbances per hour of sleep. However, clinical procedures for sleep monitoring rely on numerous uncomfortable sensors, which could affect sleeping patterns. Therefore, an automated method to identify sleep intervals from unobtrusive data is required. However, most unobtrusive sensors suffer from data loss and sensitivity to movement artifacts. Thus, current sleep detection methods are inadequate, as these require long intervals of good quality. Moreover, sleep monitoring of OSA patients is often less reliable due to heart rate disturbances, movement and sleep fragmentation. The primary aim was to develop a sleep-wake classifier for sleep time estimation of suspected OSA patients, based on single short-term segments of their cardiac and respiratory signals. The secondary aim was to define metrics to detect OSA patients directly from their predicted sleep-wake pattern and prioritize them for clinical diagnosis.Methods: This study used a dataset of 183 suspected OSA patients, of which 36 test subjects. First, a convolutional neural network was designed for sleep-wake classification based on healthier patients (AHI < 10). It employed single 30 s epochs of electrocardiograms and respiratory inductance plethysmograms. Sleep information and Total Sleep Time (TST) was derived for all patients using the short-term segments. Next, OSA patients were detected based on the average confidence of sleep predictions and the percentage of sleep-wake transitions in the predicted sleep architecture.Results: Sleep-wake classification on healthy, mild and moderate patients resulted in moderate κ scores of 0.51, 0.49, and 0.48, respectively. However, TST estimates decreased in accuracy with increasing AHI. Nevertheless, severe patients were detected with a sensitivity of 78% and specificity of 89%, and prioritized for clinical diagnosis. As such, their inaccurate TST estimate becomes irrelevant. Excluding detected OSA patients resulted in an overall estimated TST with a mean bias error of 21.9 (± 55.7) min and Pearson correlation of 0.74 to the reference.Conclusion: The presented framework offered a realistic tool for unobtrusive sleep monitoring of suspected OSA patients. Moreover, it enabled fast prioritization of severe patients for clinical diagnosis.

Monitoring In House Patients During Pandemic using Internet of Things

There is a worldwide issue that has increased the number of patients at hospitals, especially as a result of the pandemic. It is difficult for the survivors to attend routine checkups after their surgeries. Our model is helpful to solve the problem. Internet of things (IoT) arrives as a modern technical model, presenting communication and scalability, to clean this challenge. The wireless body space community is gaining quality for IoT related valuable resource packages as wearable devices enter the market. The aim of this paper is to expand the network by incorporating wearable and unobtrusive sensors to track far-flung patients after their surgeries and to receive direct assistance from doctors during an emergency.

Unobtrusive Sensors for the Assessment of Older Adult's Frailty: A Scoping Review.

Ubiquity (devices becoming part of the context) and transparency (devices not interfering with daily activities) are very significant in healthcare monitoring applications for elders. The present study undertakes a scoping review to map the literature on sensor-based unobtrusive monitoring of older adults’ frailty. We aim to determine what types of devices comply with unobtrusiveness requirements, which frailty markers have been unobtrusively assessed, which unsupervised devices have been tested, the relationships between sensor outcomes and frailty markers, and which devices can assess multiple markers. SCOPUS, PUBMED, and Web of Science were used to identify papers published 2010–2020. We selected 67 documents involving non-hospitalized older adults (65+ y.o.) and assessing frailty level or some specific frailty-marker with some sensor. Among the nine types of body worn sensors, only inertial measurement units (IMUs) on the waist and wrist-worn sensors comply with ubiquity. The former can transparently assess all variables but weight loss. Wrist-worn devices have not been tested in unsupervised conditions. Unsupervised presence detectors can predict frailty, slowness, performance, and physical activity. Waist IMUs and presence detectors are the most promising candidates for unobtrusive and unsupervised monitoring of frailty. Further research is necessary to give specific predictions of frailty level with unsupervised waist IMUs.

Unobtrusive Continuous Stress Detection in Knowledge Work—Statistical Analysis on User Acceptance

Modern knowledge work is highly intense and demanding, exposing workers to long-term psychosocial stress. In order to address the problem, stress detection technologies have been developed, enabling the continuous assessment of personal stress based on multimodal sensor data. However, stakeholders lack insights into how employees perceive different monitoring technologies and whether they are willing to share stress-indicative data in order to sustain well-being at the individual, team, and organizational levels in the knowledge work context. To fill this research gap, we developed a theoretical model for knowledge workers’ interest in sharing their stress-indicative data collected with unobtrusive sensors and examined it empirically using structural equation modeling (SEM) with a survey of 181 European knowledge workers. The results did not show statistically significant privacy concerns regarding environmental sensors such as air quality, sound level, and motion sensors. On the other hand, concerns about more privacy-sensitive methods such as tracking personal device usage patterns did not prevent user acceptance nor intent to share data. Overall, knowledge workers were highly interested in employing stress monitoring technologies to measure their stress levels and receive information about their personal well-being. The results validate the willingness to accept the unobtrusive, continuous stress detection in the context of knowledge work.

Unobtrusive Health Monitoring in Private Spaces: The Smart Home.

With the advances in sensor technology, big data, and artificial intelligence, unobtrusive in-home health monitoring has been a research focus for decades. Following up our research on smart vehicles, within the framework of unobtrusive health monitoring in private spaces, this work attempts to provide a guide to current sensor technology for unobtrusive in-home monitoring by a literature review of the state of the art and to answer, in particular, the questions: (1) What types of sensors can be used for unobtrusive in-home health data acquisition? (2) Where should the sensors be placed? (3) What data can be monitored in a smart home? (4) How can the obtained data support the monitoring functions? We conducted a retrospective literature review and summarized the state-of-the-art research on leveraging sensor technology for unobtrusive in-home health monitoring. For structured analysis, we developed a four-category terminology (location, unobtrusive sensor, data, and monitoring functions). We acquired 912 unique articles from four relevant databases (ACM Digital Lib, IEEE Xplore, PubMed, and Scopus) and screened them for relevance, resulting in papers analyzed in a structured manner using the terminology. The results delivered 25 types of sensors (motion sensor, contact sensor, pressure sensor, electrical current sensor, etc.) that can be deployed within rooms, static facilities, or electric appliances in an ambient way. While behavioral data (e.g., presence (), time spent on activities ()) can be acquired effortlessly, physiological parameters (e.g., heart rate, respiratory rate) are measurable on a limited scale (). Behavioral data contribute to functional monitoring. Emergency monitoring can be built up on behavioral and environmental data. Acquired physiological parameters allow reasonable monitoring of physiological functions to a limited extent. Environmental data and behavioral data also detect safety and security abnormalities. Social interaction monitoring relies mainly on direct monitoring of tools of communication (smartphone; computer). In summary, convincing proof of a clear effect of these monitoring functions on clinical outcome with a large sample size and long-term monitoring is still lacking.

An IoT-Based Healthcare Platform for Patients in ICU Beds During the COVID-19 Outbreak.

There is a global concern with the escalating number of patients at hospitals caused mainly by population aging, chronic diseases, and recently by the COVID-19 outbreak. To smooth this challenge, IoT emerges as an encouraging paradigm because it provides the scalability required for this purpose, supporting continuous and reliable health monitoring on a global scale. Based on this context, an IoT-based healthcare platform to provide remote monitoring for patients in a critical situation was proposed in the authors’ previous works. Therefore, this paper aims to extend the platform by integrating wearable and unobtrusive sensors to monitor patients with coronavirus disease. Furthermore, we report a real deployment of our approach in an intensive care unit for COVID-19 patients in Brazil.

Forcecardiography: A Novel Technique to Measure Heart Mechanical Vibrations onto the Chest Wall.

This paper presents forcecardiography (FCG), a novel technique to measure local, cardiac-induced vibrations onto the chest wall. Since the 19th century, several techniques have been proposed to detect the mechanical vibrations caused by cardiovascular activity, the great part of which was abandoned due to the cumbersome instrumentation involved. The recent availability of unobtrusive sensors rejuvenated the research field with the most currently established technique being seismocardiography (SCG). SCG is performed by placing accelerometers onto the subject’s chest and provides information on major events of the cardiac cycle. The proposed FCG measures the cardiac-induced vibrations via force sensors placed onto the subject’s chest and provides signals with a richer informational content as compared to SCG. The two techniques were compared by analysing simultaneous recordings acquired by means of a force sensor, an accelerometer and an electrocardiograph (ECG). The force sensor and the accelerometer were rigidly fixed to each other and fastened onto the xiphoid process with a belt. The high-frequency (HF) components of FCG and SCG were highly comparable (r > 0.95) although lagged. The lag was estimated by cross-correlation and resulted in about tens of milliseconds. An additional, large, low-frequency (LF) component, associated with ventricular volume variations, was observed in FCG, while not being visible in SCG. The encouraging results of this feasibility study suggest that FCG is not only able to acquire similar information as SCG, but it also provides additional information on ventricular contraction. Further analyses are foreseen to confirm the advantages of FCG as a technique to improve the scope and significance of pervasive cardiac monitoring.

Personalized Pervasive Health

The articles in this special section focus on personalized pervasive health. For over more than two decades, mobile, wearable, and ambient sensor and interaction devices have grown into today’s plethora of computing platforms and tools for pervasive health. Pervasive computing is now assimilating into medicine, from disease risk prevention to diagnostics, and from treatment support to chronic-care management. Unlike more traditional medical lab technology, pervasive computing systems can be continuously available, using unobtrusive sensors, actuators, computing, and interactive interfaces to liberalize access to information on basic body functions and processes in everyday life, e.g., physical activity and vital signs. Pervasive computing platforms now natively provide practitioners and researchers with a set of functionalities that allow for the development of applications and systems that are able to make sense of complex situations, e.g., daily routines, mood, and the user’s physiology. With this potential at hand, pervasive computing systems will play a central role as enabling technology for healthcare.

Respiratory Rate Estimation using PPG: A Deep Learning Approach.

Respiratory rate (RR) is an important vital sign marker of health, and it is often neglected due to a lack of unobtrusive sensors for objective and convenient measurement. The respiratory modulations present in simple photoplethysmogram (PPG) have been useful to derive RR using signal processing, waveform fiducial markers, and hand-crafted rules. An end- to-end deep learning approach based on residual network (ResNet) architecture is proposed to estimate RR using PPG. This approach takes time-series PPG data as input, learns the rules through the training process that involved an additional synthetic PPG dataset generated to overcome the insufficient data problem of deep learning, and provides RR estimation as outputs. The inclusion of a synthetic dataset for training improved the performance of the deep learning model by 34%. The final mean absolute error performance of the deep learning approach for RR estimation was 2.5±0.6 brpm using 5-fold cross-validation in two widely used public PPG datasets (n=95) with reliable RR references. The deep learning model achieved comparable performance to that of a classical method, which was also implemented for comparison. With large real-world data and reference ground truth, deep learning can be valuable for RR or other vital sign monitoring using PPG and other physiological signals.