Autonomous Virus-Killing Robots Fight COVID-19

Robot prototype uses Intel® Movidius™ Vision Processing Unit to navigate while sanitizing with ultraviolet light.

Stevie typically spends his days entertaining in a retirement community. On any given day, he’ll sing, lead a dance group, ask residents about their day, rattle off a painfully corny joke—whatever it takes to lift spirits.

Similar to other healthcare workers facing a global pandemic, Stevie had to learn new skills to lend a hand. But unlike his colleagues, he was able to shed his 4-foot 7-inch humanoid appearance and reboot as “Violet,” a prototype of a new robot that combines two superpowers: Using AI to navigate spaces and ultraviolet light to destroy bacteria and complex viral strains.

Germ-killing Violet and social robot Stevie are both creations of Akara, an Irish startup that specializes in designing artificially intelligent (AI) helpers for the healthcare industry. Akara engineers moved quickly to join in the fight against the coronavirus, teaming up with Intel to spin off the core technology of Stevie with the goal of assisting hospitals with faster, more effective sterilization.

AI on a New Wavelength
A certain range of ultraviolet light known as UV-C has been clinically proven to kill complex viruses and bacteria and has been used in hospitals for disinfection processes for the last decade.

Though there hasn’t been any conclusive research on the effects of UV-C on the SARS-CoV-2 virus that can cause the COVID-19 illness yet, studies have shown that it can be used against other coronaviruses such as SARS-CoV-1.1 The UV-C radiation warps the structure of a virus’s or bacteria’s genetic material and prevents the viral particles from reproducing. Early findings from studying UV-C’s effectiveness against SARS-CoV-2 at Columbia University’s Center for Disease and Immunity have been promising.2

Violet’s compact size and ability to operate safely amongst people makes it suitable for tasks outside the hospital as well.

UV-C robots have been used to sterilize hospitals, but most aren’t designed to work with humans in the room. Their bulky size makes it difficult to navigate crowded spaces and reach shadowy areas. They blast light that can be dangerous to humans in large doses, requiring rooms to be vacated for the safety of healthcare workers or patients since exposure to the light can be harmful, causing sunburn, eye damage, and even skin cancer.

In order to detect and avoid humans in the room, the Violet prototype is designed to use motion sensors, an Intel® Movidius™ Vision Processing Unit and a Luxonis DepthAI platform for AI, depth, and feature tracking. This allows the slim and nimble Violet to work around the people in the room, navigate to shadowy corners, clean surfaces and automatically switch off before anyone enters her limited field of ultraviolet rays.

Stevie needed computer vision to navigate a room and face a person talking to him, critical skills in order to be socially interactive and appear engaged. The Akara team joined an Intel-funded incubator in January 2020, where Movidius and Luxonis worked with them to build out advanced perceptual faculties for Stevie.

When the COVID-19 pandemic hit, the whole team sought a way to help. Following discussions at the incubator, Akara’s engineers sprinted to assemble a Violet prototype in only 24 hours. They adapted Stevie’s computer vision capabilities to ensure Violet could steer clear of human interaction.

“In some ways, Violet’s ‘person-detector’ has been easier to activate,” said Jonathan Byrne, a senior software developer at Movidius, which has been working closely with Akara on the robot. “Stevie needed to do 10-15 things, while Violet needs to do much less: clean and avoid people.”

Cutting Cleaning Time Without Cutting Corners
Violet doesn’t represent the first attempt to replace traditional, manual disinfection techniques.

Some hospitals today use other “no touch” approaches with chemical-based disinfectants, which are effective, but require evacuating rooms for much of the day.3 For example, hydrogen peroxide vapor can take more than five hours to fully disinfect a room before people can safely return to work in the area. And it simply can’t be applied in some hospital rooms.

Health systems around the world are already seeing long waiting lists for medical imagery and the slowdown only exacerbates cost pressure on hospitals. For example, radiology presents unique challenges: in many cases, high-tech equipment like CT scanners can’t be treated with deep chemicals and must be cleaned by hand, a process often done by the radiographers that takes up to 1.5 hours for every 15 minute procedure.

Expecting to be ready for hospitals by mid-2020, Violet was recently tested disinfecting CT scanners and isolation rooms in Ireland in order to reduce long downtimes due to the disinfection process. Akara found that the robot could speed things along and spare healthcare workers a potentially dangerous task of manually disinfecting rooms.

Violet’s compact size and ability to operate safely amongst people makes it suitable for tasks outside the hospital as well. The prototype is currently being tested for Dublin’s public transit system, and it’s also shown potential for disinfecting other high foot traffic, hard-to-clean spaces such as bathrooms, airports, hotels, cruise ships and waiting rooms.

“For the COVID-19 outbreak, we’re still in firefighting mode. What we’re focusing on is hospitals and nursing homes, where people are coming in and out every day,” said Niamh Donnelly, Akara co-founder and director of AI and machine learning.

But she acknowledges Violet may soon be needed elsewhere.

“As people are going back to work, holidays, and trying to resume their lives, the areas where Violet will be needed will change. We get more than 10-20 emails a day from different industries asking about a robot like Violet.”

Engineers are already exploring other applications of computer vision to control the spread of COVID-19. Using the Intel® Distribution of OpenVINO™ toolkit, engineers have found success monitoring the usage of masks (good and bad) and adherence to social distancing guidelines.4

Rays of Hope
In a world that is more vulnerable to pandemics, academics are looking to robots like Violet as playing an essential role in ensuring population safety.5 Future use cases are already clear: Robots like Stevie and Violet could be further adapted to use computer vision to monitor social distancing, ensure compliance with personal protective equipment (PPE) rules, and even perform contact tracing. These capabilities are currently being explored with open source artificial intelligence on the Intel® Distribution of OpenVINO™ toolkit.6

In the early stages of crisis response, much focus has been on the development of rapid response-tests and a vaccine. This work is profoundly important, but other innovations are required in the meantime. This experience has taught us the urgency of making every hospital pandemic-ready, for the sake of both patients and frontline workers. We need all the help we can get, human or not.

Explore Related Products and Solutions

Intel® Xeon® Scalable Processors

Drive actionable insight, count on hardware-based security, and deploy dynamic service delivery with Intel® Xeon® Scalable processors.

Learn more

OpenVINO™ Toolkit

Build end-to-end computer vision solutions quickly and consistently on Intel® architecture and our deep learning framework.

Learn more

Avisos y exenciones de responsabilidad

Las características y ventajas de las tecnologías Intel® dependen de la configuración del sistema y es posible que necesiten hardware y software habilitados, o la activación del servicio. El desempeño varía según la configuración del sistema. Ningún sistema informático puede proporcionar una seguridad absoluta. Consulte con el vendedor o el fabricante del sistema, o visite para obtener más información. // Es posible que el software y las cargas de trabajo utilizados en las pruebas de desempeño solo se hayan optimizado para desempeño en los microprocesadores Intel. Las pruebas de desempeño, como SYSmark y MobileMark, se miden utilizando sistemas de computación, componentes, software, operaciones y funciones específicos. Cualquier cambio en alguno de esos factores podría generar un cambio en los resultados. Debe consultar otra información y pruebas de desempeño que lo ayuden a evaluar plenamente las compras consideradas, incluido el desempeño de ese producto al combinarse con otros. Para obtener más información, visite Los resultados de desempeño se basan en pruebas realizadas en la fecha indicada en la configuración y es posible que no reflejen todas las actualizaciones de seguridad que están a disposición del público. Consulte la divulgación de configuración para obtener más información. Ningún producto o componente puede proporcionar una seguridad absoluta. Los escenarios de reducción de costo publicados se utilizan como ejemplo de cómo un producto equipado con Intel® dado, en circunstancias y configuraciones específicas, puede afectar los costos futuros y proporcionar recortes en cuanto a costo. Las circunstancias pueden variar. Intel no garantiza costos o reducciones de costos, // Intel no ejerce control ni inspección algunos sobre los datos de análisis de desempeño o los sitios web de terceros a los que se hace referencia en este documento. Debe visitar el sitio web referido y confirmar si los datos a los que se hacen referencia son precisos. // En algunos casos de prueba, los resultados estimados o simulados con modelado o simulación de arquitectura o análisis internos de Intel con fines informativos. Cualquier diferencia en el hardware, software o configuración del sistema puede afectar el desempeño real.

Información sobre productos y desempeño

4Intel is committed to respecting human rights and avoiding complicity in human rights abuses. See Intel's Global Human Rights Principles. Intel's products and software are intended only to be used in applications that do not cause or contribute to a violation of an internationally recognized human right.