Personal Data Protection in Smart Home Activity Monitoring for Digital Health: A Case Study

Claudio Bettini, Azin Moradbeikie, Gabriele Civitarese · 2025

Read Personal Data Protection in Smart Home Activity Monitoring for Digital Health: A Case Study by Claudio Bettini, Azin Moradbeikie, Gabriele Civitarese (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-13864.txt.

Introduction

Claudio Bettini, Azin Moradbeikie, and Gabriele Civitarese wrote this 2025 computing paper on privacy in smart-home health monitoring. Sensors in a home can recognize daily activities, and shifts in those routines can flag early cognitive decline. But placing such a system in an older person's home raises hard privacy and ethics questions.

The authors report their field experience building and deploying one such system under privacy-by-design rules. They describe the choices made from the ground up: what data is gathered, where it is processed, and what the residents and clinicians can see. The case study shows the tensions that appear once a lab design meets a real household. Their lessons aim at researchers who want health insight without turning a home into a surveillance site.

Catalog record

CREATORClaudio Bettini, Azin Moradbeikie, Gabriele Civitarese
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE29 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-13864

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Personal Data Protection in Smart Home Activity Monitoring for Digital Health: A Case Study Claudio Bettini, Azin Moradbeikie, Gabriele Civitarese Abstract Researchers in pervasive computing have worked for decades on sensor-based human activity recognition (HAR). Among the digital health applications, the recognition of activities of daily living (ADL) in smart home environments enables the identification of behavioral changes that clinicians consider as a digital bio-marker of early stages of cognitive decline. The real deployment of sensor-based HAR systems in the homes of elderly subjects poses several challenges, with privacy and ethical concerns being major ones. This paper reports our experience applying privacy by design principles to develop and deploy one of these systems. This paper has been accepted and presented at the TELMED ’25 workshop. It will appear in the 2025 IEEE International Conference on Pervasive Computing and Communications Workshops and other Affiliated Events (PerCom Workshops) proceedings. I Introduction Human Activity Recognition (HAR) is a very active research area, with significant progress partly due to new sensing technologies and new AI methods [ 1 , 2 ] . Sensor-based HAR enables several healthcare applications, including the early detection of cognitive decline. Using remote, unobtrusive, and continuous monitoring to diagnose early stages of cognitive decline, its phases, and its possible progression to dementia through the analysis of Activities of Daily Living (ADLs) is a challenge of clinical interest [ 3 ] . Indeed, several projects have investigated the sensor-based recognition of human activities with real deployments in smart home targeting the specific application of monitoring cognitive decline as revealed by behavioral changes of the monitored subjects [ 4 , 5 , 6 ] . The lack of a reliable privacy-preserving infrastructure is, among other challenges, one of the reasons why there have not been large-scale deployments of these systems. Privacy is by itself a research area in mobile and pervasive systems [ 7 ] . Real systems should be built with privacy by design principles, adopting end-to-end state-of-the-art solutions, complying with the new regulations, and reassuring the data subjects. In this paper, we report our experience in designing and deploying a privacy-aware infrastructure as part of a new project in this area. While the number of monitored homes is again pretty limited, this is a pilot for a possibly large-scale regional or national deployment. II The SERENADE project The SERENADE project aims to develop a telemedicine solution for the detection, through intelligent analysis of long-term sensor data, of behavioral changes suggesting the evolution of Mild Cognitive Impairment (MCI). MCI is a pre-stage of dementia where the subject, even if presenting cognitive dysfunctions on one or multiple domains, still keeps their daily functionality. For some individuals, MCI may progress to dementia. The clinical goal of this project is to identify unobtrusive digital biomarkers to monitor functional and symptomatic changes in MCI patients living alone, in order to capture the early transition from MCI to dementia. The project team includes experts in sensor-based activity recognition, clinicians (neurologists and neuropsychologists), and technicians managing the sensing infrastructure. This multidisciplinary team identified the key behaviors that can be monitor through sensors: nutrition, personal hygiene, sleep, therapy adherence, mobility, and cognition. Specifically, SERENADE leverages unobtrusive sensing devices such as environmental sensors (e.g., magnetic, PIR, and plug sensors), a smart sleep analyzer, a tablet to administer periodical cognitive tests, and a smartwatch. Activities of Daily Living (ADLs), like cooking and eating, are recognized using XAI methods on the stream of sensor data. Detected ADLs are then analyzed to infer long-term behavioral changes possibly linked to cognitive decline. These changes and corresponding explanations are inspected by the clinicians via a dedicated web dashboard during the clinical visits to support their diagnosis. SERENADE considers two patient cohorts: one with MCI due to neurodegeneration, likely progressing to cognitive decline, and another with MCI not due to neurodegeneration, typically stable. Overall, we plan to recruit 15 subjects for each group. At the time of writing, we deployed our system in the home of 18 patients. III Privacy Model In this section, we describe the privacy model by identifying the role of each entity in the data workflow, specifying the type of data exposed to each entity, and analyzing the possible threats. In the following, we consider a privacy threat the acquisition of identified sensitive data by unauthorized entities. In addition to usual personal data, we focus on the protection of sensitive data related to the health of the subjects as well as to their activities and habits in daily life, when at home or outside. By identified we mean that the data is associated with the identity of a specific subject. III-A Data subjects, Data controller and data processors III-A 1 Data subject In Serenade, the data subjects are the patients selected by the hospital, that will be monitored during the pilot. In the following, we will refer to them simply as subjects . III-A 2 Data controller The data controller is the hospital (HOS) that organizes the recruitment of the participating subjects among its patients. III-A 3 Data processor The data processors in Serenade are the following: The main data processor is our lab 1 1 1 Formally is Università degli Studi di Milano. (UniMi) with a team of computer scientists that designs algorithms, performs data analysis, and offers a dashboard to visualize results for clinicians. A second data processor is the Infrastructure Installation Team (IIT) which is in charge of installing the necessary devices (mini-PC, tablet, smartwatch, sensor devices, etc.) in the subjects’ homes and intervening during the study to fix possible malfunctions. A third data processor is the Infrastructure Monitoring Team (IMT) which configures the devices to be installed and monitors the infrastructure to ensure that sensor data is correctly acquired and stored as well as transmitted to the central repository for future analysis. IMT should not be able to analyze data. They are just responsible for making sure of the correct configuration and operation of devices. A fourth data processor is the COD20 Telemedicine technical team (TelMed) which will provide secure storage of the data, authentication and access control mechanisms, and a platform for interfacing with legacy systems. Other (external) data processors are: • the company that manufactures the smartwatch and sleep analyzer systems and that receives through a smartphone app, stores, and preprocesses in its cloud the raw data obtained by these devices. • the company that manufactures the smart toothbrush and that receives through a smartphone app, stores and preprocesses in its cloud the raw data obtained by this device. • the company that acquires and processes the outdoor movements of the subject by its geolocation services activated on the subject’s smartphone. III-B Data control Data processors have different requirements in terms of data access: • UniMi needs to receive all sensing data, but it does not need to know the subject’s identity. However, to correctly interpret sensor data, it may need some general information about the subject and the home environment. For example, it should know about the positioning of sensors in the apartment, if there are any pets, and if other individuals are living with or regularly visiting the subject. • The IIT, to perform installation and maintenance interventions, needs to know the subject’s

Read the complete text