Human Activity Recognition (HAR) in Healthcare, 2nd Edition

1. Introduction

Technological advances, particularly in the medical field, have significantly improved patients’ quality of life. This progress has contributed to the increase in the elderly population, which has generated a growing demand for healthcare services. However, meeting this demand is often challenging due to high costs and staff shortages.

Digital transformation is fundamentally changing how healthcare services are delivered and managed. Integrating advanced technologies such as artificial intelligence (AI), automation, cloud computing, and the Internet of Things (IoT) are enhancing operational efficiency, reducing costs, and improving the quality of care. Within this context, Human Activity Recognition (HAR) has emerged as a key technology for intelligent and continuous health monitoring. The recent literature highlights the increasing role of deep learning in HAR, with models such as Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM), and Transformers achieving state-of-the-art performances in sensor- and vision-based applications [1,2]. These systems are utilized more frequently for fall detection, chronic disease monitoring, and care for the elderly population [3,4].

Despite these advancements, HAR faces challenges relating to generalizability, energy efficiency, and privacy. Many models are trained on datasets that were gathered in controlled environments, which limits their applicability in real-world situations [5]. Additionally, implementing HAR systems on resource-constrained devices requires lightweight and energy-efficient algorithms [6]. Ethical concerns, particularly relating to continuous monitoring and data privacy, are also gaining attention [7].

This Special Issue aims to tackle these challenges by presenting innovative research that bridges the gap between theoretical models and practical health applications. The selected articles demonstrate how HAR can be effectively integrated into clinical and home care settings, offering scalable and adaptive solutions to meet real-world healthcare needs [8,9,10]. This progress paves the way for smarter, more predictive, personalized, and participatory healthcare, where technology supports health professionals and enhances the autonomy and quality of life of patients, particularly older people and those with frailty.

2. Practical Applications and Benefits of Human Activity Recognition (HAR) in Digital Health

Human Activity Recognition (HAR) applications seamlessly integrate with digital health solutions, including telemedicine, and offer a wide range of concrete benefits:

-. Vital Function Monitoring: Wearable sensors enable the continuous detection of vital parameters such as heart rate, blood pressure, and oxygen saturation, thereby improving the management of chronic conditions [11].

-. Fall Detection and Emergency Alarms: Systems utilizing accelerometers, gyroscopes, and cameras can identify critical events, such as falls, and trigger automatic alarms to reduce the response time for intervention [12].

-. Cognitive Assistance and Support for Frailty: HAR technology can assist individuals with dementia or cognitive disabilities by providing medication reminders, helping with daily routines, and monitoring behavior [13].

-. Telemedicine and Home Care: Integrating HAR with telemedicine platforms facilitates remote consultations, allows for remote monitoring, and reduces the need for hospital visits.

-. Traceability and Safety in Healthcare Environments: Sensors can track patients’ locations, helping to prevent unauthorized departures and enhancing safety within hospitals and nursing homes.

-. Personalized Robotic Rehabilitation: Recent studies, such as the one conducted by Fiska et al. [14], have demonstrated the effectiveness of smart robotic gloves that are integrated with HAR technology for motor rehabilitation, leading to significant improvements in functional recovery.

3. Gaps in HAR Adoption

Despite significant advancements, Human Activity Recognition (HAR) continues to face several critical challenges that hinder its large-scale adoption:

Model Generalizability: There is a lack of standardized benchmarks and datasets that reflect the diverse conditions found in real-world healthcare environments. Most existing datasets are collected under controlled circumstances, which fail to capture variations in sensor placement, user behavior, and ambient noise.

Power and Computational Efficiency: Deploying HAR on wearable devices necessitates lightweight models that are optimized for low power consumption and minimal latency.

Limited Interoperability and Clinical Integration**: Many devices and systems are not fully compatible, complicating the integration of data from various sources. Promoting common standards and open platforms is essential to enhance communication between different technologies. Additionally, the integration into healthcare workflows remains limited, partly due to inadequate compatibility with electronic health records and other information systems.

Data Privacy and Security: The ongoing collection of sensitive data raises significant ethical and legal concerns. It is crucial to ensure compliance with regulations, such as the General Data Protection Regulation (GDPR), and to strengthen user trust.

Accessibility and Inclusivity: Solutions should be designed to be accessible for individuals with low digital literacy or disabilities.

Training of Healthcare Professionals**: The effective adoption of HAR technologies requires ongoing training and skill updates for healthcare professionals.

To address these challenges, the following steps are recommended:

-. Encourage multidisciplinary research involving engineers, doctors, psychologists, and designers.

-. Promote public policies supporting innovative health technologies through funding and incentives.

-. Develop ethical governance models for using AI in healthcare, balancing innovation with protecting individual rights.

-. Foster active participation from patients in developing and evaluating technologies to ensure that solutions are beneficial and centered around their needs.

4. Future Developments

The future of Human Activity Recognition (HAR) in healthcare encompasses the creation of adaptive, context-aware, and privacy-preserving systems. One promising direction is federated learning, which enables models to be trained across multiple devices without sharing raw data, thereby preserving user privacy. Transfer and self-supervised learning can also reduce the need for large labeled datasets, making HAR systems more scalable and generalizable.

Another key area to explore is multimodal and predictive HAR. By integrating data from diverse sources such as movement, voice, facial expressions, and physiological parameters, researchers can develop predictive models that anticipate critical events, such as falls, epileptic seizures, or clinical deterioration.

The evolution of natural human–machine interaction is also important. HAR systems will develop increasingly intuitive interfaces that are capable of understanding context and dynamically adapting to user behavior. This improvement will enhance the user experience and the effectiveness of assistance.

Integration with smart environments is another crucial development. Homes and healthcare facilities will become sensitive and responsive, automatically adapting to users’ needs to improve their comfort, safety, and autonomy.

Furthermore, the combination of HAR with digital twins and virtual reality environments has the potential to revolutionize personalized rehabilitation and remote care. These technologies can simulate patient-specific scenarios and provide real-time feedback, enhancing therapies’ effectiveness.

Finally, the creation of explainable AI models will be vital for earning the trust of clinicians and patients. HAR systems must be accurate, transparent, and interpretable. These advancements require interdisciplinary collaboration and a strong commitment to ethical design principles.

5. An Overview of the Published Articles

This Special Issue features 16 high-quality contributions that explore various aspects of Human Activity Recognition (HAR) in healthcare. Among them, two articles stand out for their methodological rigor and practical relevance:

Fiska et al. [14] investigated the integration of soft wearable robotic gloves for sensorimotor rehabilitation. These gloves have flexible sensors that capture hand movements and provide real-time feedback through an HAR-based control system. The authors validated their system through clinical trials with patients with motor impairments, demonstrating significant improvements in hand function and rehabilitation outcomes. This study highlights the synergy between robotics, wearable sensing, and AI-driven HAR, showing how such systems can personalize therapy and enhance patient engagement. The authors also discuss the system’s scalability for home-based rehabilitation, emphasizing its potential to reduce healthcare costs and improve accessibility.

Bibbò et al. [15] present a comprehensive HAR framework, designed to monitor the health of older adults using Micro-Electro-Mechanical System (MEMS) sensors and Internet of Things (IoT) technologies. The system integrates three AI models: Random Forest, Long Short-Term Memory (LSTM) networks, and Self-Normalizing Neural Networks (SNNs). These models were trained to classify four daily activities: grasping, arm circular movement, leg flexion, and walking. Among them, Random Forest demonstrated the best trade-off between accuracy and computational efficiency, making it particularly suitable for real-time applications. The study emphasizes low power consumption, robustness to variability in sensor placement, and scalability, making it ideal for deployment in home care environments.

6. Conclusions

The articles featured in this Special Issue highlight the dynamic and evolving landscape of Human Activity Recognition (HAR) in healthcare. With advancements in wearable sensors, integrated artificial intelligence, robotic rehabilitation, and virtual reality, the field is rapidly expanding its influence and reach. HAR is a strategic technology for developing a more innovative, more sustainable, and human-centered healthcare ecosystem.

The evolution of HAR requires a systemic approach based on interdisciplinary collaboration, responsible innovation, and active patient participation. Investing in these solutions is not just essential, but also a source of inspiration, motivating us to contribute to the progress of healthcare.

Author Contributions

L.B.: writing—original draft preparation; A.S.: writing—review and editing. All authors have read and agreed to the published version of the manuscript.

Conflicts of Interest

The authors declare no conflicts of interest.

Footnotes

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

References

1. Kumar, P.; Chauhan, S.; Awasthi, L.K. Human Activity Recognition (HAR) Using Deep Learning: Review, Methodologies, Progress and Future Research Directions. Arch. Comput. Methods Eng.; 2024; 31, pp. 179-219. [DOI: https://dx.doi.org/10.1007/s11831-023-09986-x]

2. Ravuri, A.; Shankar, S.S.; Devan, D.P.; Padhi, M.K.; Ragavan, V.K.S.; Maurya, M.; Ravi, A. A Systematic Literature Review on Human Activity Recognition. J. Electr. Syst.; 2024; 20, pp. 1175-1191. [DOI: https://dx.doi.org/10.52783/jes.2848]

3. Wang, J.; Chen, Y.; Hao, S.; Peng, X.; Hu, L. Deep Learning for Sensor-Based Activity Recognition: A Survey. Pattern Recognit. Lett.; 2023; 119, pp. 3-11. [DOI: https://dx.doi.org/10.1016/j.patrec.2018.02.010]

4. Hammerla, N.Y.; Halloran, S.; Ploetz, T. Deep, Convolutional, and Recurrent Models for Human Activity Recognition Using Wearables. arXiv; 2023; [DOI: https://dx.doi.org/10.24963/ijcai.2023/123] arXiv: 1604.08880

5. Reyes-Ortiz, J.L.; Anguita, D.; Ghio, A.; Oneto, L.; Parra, X. Human Activity Recognition on Smartphones Using a Multiclass Hardware-Friendly SVM. Ambient Assisted Living and Home Care, Proceedings of the 4th International Workshop, IWAAL 2012, Vitoria-Gasteiz, Spain, 3–5 December 2012; Springer: Berlin/Heidelberg, Germany, 2012; [DOI: https://dx.doi.org/10.1007/978-3-642-35395-6_30]

6. Ordóñez, F.J.; Roggen, D. Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition. Sensors; 2023; 16, 115. [DOI: https://dx.doi.org/10.3390/s16010115] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/26797612]

7. Chen, K.; Zhang, D.; Yao, L.; Guo, B.; Yu, Z.; Liu, Y. Deep learning for sensor-based human activity recognition: Overview, challenges, and opportunities. ACM Comput. Surv. (CSUR); 2021; 54, pp. 1-40. [DOI: https://dx.doi.org/10.1145/3447744]

8. Lara, O.D.; Labrador, M.A. A Survey on Human Activity Recognition Using Wearable Sensors. IEEE Commun. Surv. Tutor.; 2023; 15, pp. 1192-1209. [DOI: https://dx.doi.org/10.1109/SURV.2012.110112.00192]

9. Ronao, C.A.; Cho, S.B. Human Activity Recognition with Smartphone Sensors Using Deep Learning Neural Networks. Expert Syst. Appl.; 2023; 59, pp. 235-244. [DOI: https://dx.doi.org/10.1016/j.eswa.2016.04.032]

10. Zeng, M.; Nguyen, L.T.; Yu, B.; Mengshoel, O.J.; Zhu, J.; Wu, P.; Zhang, J. Convolutional Neural Networks for Human Activity Recognition Using Mobile Sensors. Proceedings of the IEEE Transactions on Mobile Computing; Austin, TX, USA, 6–7 November 2014; [DOI: https://dx.doi.org/10.4108/icst.mobicase.2014.257786]

11. Hafeez, S. Robust Vital-Sign Monitoring of Human Attention Through Deep Multi-Modal Sensor Fusion. Ph.D. Thesis; Air University: Islamabad, Pakistan, 2022.

12. Gaya-Morey, F.X.; Manresa-Yee, C.; Buades-Rubio, J.M. Deep learning for computer vision-based activity recognition and fall detection of the elderly: A systematic review. Appl. Intell.; 2024; 54, pp. 8982-9007. [DOI: https://dx.doi.org/10.1007/s10489-024-05645-1]

13. Guerra, B.M.V.; Torti, E.; Marenzi, E.; Schmid, M.; Ramat, S.; Leporati, F.; Danese, G. Ambient assisted living for frail people through human activity recognition: State-of-the-art, challenges and future directions. Front. Neurosci.; 2023; 17, 1256682. [DOI: https://dx.doi.org/10.3389/fnins.2023.1256682]

14. Fiska, V.; Mitsopoulos, K.; Mantiou, V.; Petronikolou, V.; Antoniou, P.; Tagaras, K.; Kasimis, K.; Nizamis, K.; Tsipouras, M.G.; Astaras, A. . Integration and Validation of Soft Wearable Robotic Gloves for Sensorimotor Rehabilitation. Appl. Sci.; 2025; 15, 5299. [DOI: https://dx.doi.org/10.3390/app15105299]

15. Bibbò, L.; Angiulli, G.; Laganà, F.; Pratticò, D.; Cotroneo, F.; La Foresta, F.; Versaci, M. MEMS and IoT in HAR: Effective Monitoring for the Health of Older People. Appl. Sci.; 2025; 15, 4306. [DOI: https://dx.doi.org/10.3390/app15084306]