For older people living alone, a fall or sudden medical emergency can go unnoticed for hours. Researchers at Nottingham Trent University, working with British technology company Integrated System Technologies, have developed an AI-powered monitoring system designed to detect potentially dangerous situations without intrusive video surveillance. Combining radar, thermal sensors and smart plugs, the technology tracks movements, body postures and everyday activities while seeking to preserve residents’ privacy. The research points towards a new approach to home safety – but important questions remain about reliability, data protection and real-world deployment.

Helping older people remain independent without compromising their safety is an increasingly pressing challenge for healthcare and social care systems. Conventional personal alarm services often depend on users wearing a device or actively requesting assistance – measures that may prove ineffective if someone loses consciousness or cannot reach an alarm button.

The system developed by Nottingham Trent University and Integrated System Technologies (IST) aims to overcome this limitation by identifying unusual patterns of activity automatically.

Rather than recording video footage, the technology combines data from several sensors to build a picture of daily routines. Artificial intelligence is intended to identify significant deviations from established behavioural patterns, potentially signalling that a resident requires assistance.

If a suspected emergency is detected, the system could alert healthcare professionals, carers or family members, enabling them to assess the situation and arrange an appropriate response.

Combining Three Technologies to Monitor Daily Activities

The system brings together three complementary sensing technologies.

Millimetre-wave radar detects movement and changes in position within a room. Low-resolution thermal sensors provide additional information about body posture, helping distinguish between activities such as standing, walking, sitting and lying down.

Smart plugs contribute a further layer of information by monitoring electricity consumption from connected household appliances. Their use can indicate routine activities, such as preparing a meal or switching on a kettle.

By combining these signals, the AI system is intended to establish patterns of normal behaviour and detect potentially significant changes. Prolonged inactivity, unusual movement patterns or the absence of familiar daily routines could indicate that something is wrong.

Privacy protection is central to the design. Unlike conventional surveillance cameras, the thermal sensors produce low-resolution representations rather than identifiable images. The proposed approach also avoids requiring residents to wear monitoring devices.

This could make the technology less intrusive than conventional camera-based monitoring while allowing continuous observation of relevant activities.

Laboratory Testing Demonstrates the Potential of Sensor Fusion

Researchers evaluated the technology in a simulated domestic environment, testing six different room configurations and varying furniture arrangements and patterns of movement.

The experiments demonstrated the complementary capabilities of the sensing technologies. Radar proved useful for detecting spatial movement, while thermal sensing offered more detailed information about body posture. Smart plugs provided additional evidence of appliance use and everyday activities.

The findings, published in the scientific journal Sensors, support the potential of combining several non-visual sensing technologies to monitor activities of daily living.

However, the distinction between detecting unusual behaviour and reliably identifying a medical emergency remains critical.

The research examined movement recognition, posture classification and appliance use under controlled conditions. It did not establish that the system can reliably diagnose or detect specific medical emergencies such as heart attacks or strokes.

Further testing in occupied homes will therefore be necessary before its performance in real-life care settings can be assessed.

Privacy and Alarm Reliability Remain Critical

From a security perspective, the technology addresses two potentially conflicting requirements: maintaining awareness of a resident’s safety while preserving privacy within their home.

Removing conventional cameras reduces the collection of identifiable visual information, but it does not eliminate data protection concerns.

Movement patterns, appliance usage and daily routines can reveal sensitive information about a person’s lifestyle, habits and potentially their health.

Any future deployment would consequently require clear safeguards governing data collection, processing, storage and access. Secure communications would also be essential to protect monitoring information and ensure that alerts reach authorised recipients.

Equally important is the reliability of emergency detection.

A practical system must distinguish genuine signs of distress from harmless changes in behaviour. Too many false alarms could overwhelm carers or emergency response services, while missed incidents could leave vulnerable residents without assistance.

System designers must also consider power failures, sensor malfunctions and communication outages. These risks become particularly significant when residents or care providers begin relying on automated monitoring as part of their safety arrangements.

The effectiveness of such a system will ultimately depend not only on accurate detection but also on clearly defined escalation procedures and a dependable response network.

Potential Applications in Assisted Living

If successfully validated, the technology could complement existing personal emergency response systems, telecare services and assisted-living arrangements.

For older people who live alone, unobtrusive monitoring may offer an additional layer of protection without requiring constant interaction with technical equipment.

Care providers could potentially use the information to identify changes in daily activity that warrant attention, while emergency alerts could enable earlier intervention when residents are unable to request help themselves.

Nevertheless, several obstacles remain before the system can move beyond experimental development. These include demonstrating reliable performance in different homes, establishing acceptable false-alarm rates, securing residents’ informed consent and integrating alerts into established care and emergency response procedures.

From Intelligent Monitoring to Dependable Protection

The Nottingham Trent University project illustrates how radar, thermal sensing and artificial intelligence could be combined to support a less intrusive approach to residential safety monitoring.

Its potential extends beyond detecting movement. By recognising changes in everyday behaviour, future systems may help identify situations requiring intervention before they develop into more serious emergencies.

Yet technological capability alone is not sufficient to establish a dependable safety solution.

For AI-based home monitoring to become a credible component of assisted living, developers must demonstrate that the technology works reliably under real-world conditions, protects sensitive personal information and connects suspected emergencies to an effective human response.

The central challenge is not simply to recognise when something unusual happens. It is to ensure that the information leads to timely, appropriate assistance – without compromising the independence and dignity the technology is intended to preserve.

Editorial sources: Nottingham Trent University, “Elderly people could live independently for longer thanks to new home sensor system which detects emergencies,” 8 October 2026; Mohammed et al., “Privacy-Preserving Ambient Sensing for Activities of Daily Living,” Sensors, August 2026, DOI: 10.3390/s26165066; pressetext, “KI passt künftig daheim auf ältere Menschen auf,” 12 October 2026.