Safe Sphere

Let's Discover

Innovation Fest 2025

About Safe Sphere

Our project is an incident response system that uses a camera with a custom object detection model to detect falls in the home. The safe sphere incident response system is a travel friendly, portable device that can be mounted to the wall of the user’s home or even healthcare environments. The incident response system uses a camera to monitor the vulnerable individual 24/7 to detect if the user has fell over. If the user has fell , an alarm will sound, the emergency services and loved ones will be contacted immediately and notified of the incident by a message that appears on the app. Once the emergency services are contacted, the microphone will be activated therefore the user can engage in a live call with the emergency services

Our project aims to protect vulnerable groups such as the elderly and people with chronic illnesses. Families are concerned that their vulnerable loved ones are unsafe and at a high risk of harmful injuries whilst at home alone .

UN SUSTAINABILITY DEVELOPMENT GOALS

Empowered and Driven:
Delivering Solutions for Global Issues

The UN Sustainability on Goal of Good Health and Wellbeing focuses on promoting healthy lifestyles and improving societies health and wellbeing. It aims to reduce maternal mortality , prevent deaths of newborns and children under the age of 5, fight and cure transmittable and infectious diseases , raise awareness of mental health and address drug and alcohol abuse .It also seeks to reduce harm from hazardous chemicals and pollution , ensure access to affordable medicines and vaccines especially in developing countries and it supports an increase in health financing that is required to enhance healthcare services.

One of the goals of the UN Sustainability Goal on Good Health and Wellbeing highlights the need for an increase in health financing. However, the UK is suffering from inflation therefore the government does not have the extra funds to allocate towards the NHS and to increase healthcare workers’ pay. The lack of funding is impacting healthcare services resulting in appointment delays, shortage of hospital beds and patient backlogs. As a consequence of this, healthcare workers are leaving the profession causing a shortage of staff. With the incident response system in place within healthcare settings care workers will only be required to care for individuals who are incapacitated, require hands on care and specialised medical attention.

In addition , safe sphere believes the elderly should not have to suffer with isolation, patronising treatment and restricted freedom in a care home. Instead the incident response system provides elderly people with independence , safety and freedom while enjoying the comfort of their own home.

THE STORY BEGAN

DESIGN STORY

Challenges

The prototype is a simple proof of concept hosted locally. It displays the webcam (or images in a specified folder) and uses the trained model to detect falls. If a fall is detected a bounding box is drawn, and information regarding the fall is saved to a log file. This is then viewable on the website. There’s also an option to send notification emails when valid SMTP settings are configured in the code. However, this is awkward to set up, so a version of this that is fully controlled by the group is preferrable. For example, using an app to send push notifications to a mobile device. These things are secondary, however, as the main feature is simply the object detection model. It’s completely plausible that notification related features can be made, they just need more time for this to be implemented. In addition , obtaining a pi zero and camera to use as an external camera and powering it in a portable way (with a battery pack) is a complex process to complete in such a short amount of time near the deadline.

Solution

Falls are the second leading cause of unintentional injury-related deaths globally, responsible for approximately 684,000 fatalities each year. When vulnerable individuals fall while alone, they often face delayed medical attention, leading to severe complications, prolonged pain, or even death. The Safe Sphere device offers a life-saving solution by using smart technology to detect falls and immediately alert emergency services. This rapid response reduces the likelihood of long-term injuries, hospitalisation, or further deterioration. With the NHS under immense pressure due to funding shortages, staff deficits, and increasing patient backlogs, Safe Sphere eases the burden by ensuring that only individuals needing urgent, hands-on care are prioritised. It also empowers elderly individuals to maintain independence, safety, and dignity within their own homes, avoiding the isolation and restricted lifestyles of care homes. Families gain peace of mind knowing their loved ones are protected 24/7, while healthcare systems benefit from more efficient resource allocation and reduced care demands.

Use Case

Our project is an advanced incident response system that uses a camera equipped with a custom object detection model to accurately detect falls within the home. It is designed to protect vulnerable groups such as the elderly and individuals with chronic illnesses, ensuring they receive immediate assistance when accidents occur. Families often worry about the safety of loved ones at home alone and the high costs of care homes. The Safe Sphere system monitors movements in real time, alerting emergency services if a fall is detected, providing peace of mind and enabling vulnerable individuals to live safely and independently. Many elderly individuals feel like a burden or guilty for care-related expenses; this system helps maintain their dignity and autonomy. Additionally, it reduces the pressure on healthcare systems like the NHS, allowing staff to focus on patients who require intensive care. By supporting at-home safety, the system offers a scalable solution to the ongoing healthcare worker shortage.

TECHNOLOGIES

List of Essential Tools and Technologies

Raspberry Pi Camera

Raspberry Pi

Adafruit I2S 3W

Python

Software Design Sketchfab

Mixamo

THE PROCESS

PROTOTYPE DESIGN

The prototype is a simple proof of concept hosted locally. It displays the webcam (or images in a specified folder) and uses the trained model to detect falls. If a fall is detected a bounding box is drawn, and information regarding the fall is saved to a log file. This is then viewable on the website.

There’s also an option to send notification emails when valid SMTP settings are configured in the code. However, this is awkward to set up, so a version of this that is fully controlled by the group is preferrable. For example, using an app to send push notifications to a mobile device.

These things are secondary, however, as the main feature is simply the object detection model. It’s completely plausible that notification related features can be made, they just need more time for this to be implemented.

MicrosoftTeams-image (5)_Nageenah Sagheer

PLEDGE

Impact on the market

Fall detection systems are increasingly important due to the rise in the ageing population. Traditional systems rely on wearable devices or smartphone apps, which can be inconvenient for elderly users, especially those with memory issues. These systems often require users to press a button or wear a device, which may not be practical during an emergency.

SafeSphere addresses these limitations by offering automatic, 24/7 monitoring without requiring user intervention or wearable technology. Using advanced machine learning algorithms and a camera, SafeSphere can accurately detect falls in real time, distinguish between different types of movements, and immediately notify emergency services and designated contacts.

Housed in a waterproof, portable case, SafeSphere is travel-friendly and easy to set up in different locations. The device can be mounted on a wall, enabling the camera to track body movements accurately. Unlike systems that may trigger false alarms when wearables are dropped, SafeSphere remains fixed in place, reducing errors.

Its simple setup and affordable pricing make it especially suitable for working-class families who cannot afford expensive care homes or high-tech alternatives. SafeSphere ensures elderly users remain safe and independent in their homes, providing peace of mind for both users and their loved ones.

RESULTS

FINAL PRODUCT

Setup for the system should be as simple as possible, with a minimal amount of user configuration. Where configuration is needed, it should be done using a GUI, as opposed to entering commands in the terminal, or manually altering configuration fiels (json, toml, yaml etc.).

External cameras should use mDNS to locate the central hub. This allows the cameras to set themselves up, making them more user friendly. A fallback option should be present, however, as not all networks will support this. The website (when accessed externally) should require authentication as privacy is very important for the project. Connecting the app to a specific system should also require authentication, initially through the same network. This should then make the app and system aware of each other, such that some form of whitelisting can take place.

The case for the hub should be sturdy, and possibly wall-mountable for the maximum amount of flexibility in positioning (although this is more of a priority for external cameras).