Periodic Reporting for period 3 - AVRICS (Akara Violet Robotic Infection Control System)
Période du rapport: 2023-08-01 au 2025-04-30
Hospitals face growing pressure from rising demand, aging populations, chronic disease, and surgical backlogs worsened by COVID-19. At the same time, they struggle with workforce shortages, outdated infrastructure, and slow-moving procurement systems. Infection prevention remains labour-intensive and inconsistent, and most existing UV or HPV disinfection systems are impractical; they require empty rooms and offer no visibility into performance.
Digital infrastructure is also limited. Key processes like disinfection, scheduling, and room tracking are often handled manually, leaving hospitals without reliable data to identify inefficiencies or evaluate improvements. This is especially problematic in operating rooms, where up to two hours per day are commonly lost to avoidable delays.
AVRICS was launched to address these issues by developing and commercializing a CE-marked UV disinfection robot, Violet, designed for safe use in occupied rooms. It delivers traceable, measurable disinfection through AI-powered dose control and optimized navigation. The project also developed foundational infrastructure involving hardware, software, user interfaces, and integration pathway to enable scalable adoption of autonomous systems in hospitals.
Five core objectives guided the project:
(1) Product Development: Deliver a hospital-ready, compliant robot.
(2) Evidence Generation: Validate safety, traceability, and efficacy.
(3) Market Discovery: Identify early adopters and refine GTM strategy.
(4) Adoption Pathway: Build a repeatable pilot-to-procurement model.
(5) Strategic Positioning: Prepare for global scale through team, partners, and investment.
During US market engagement, the team identified a broader need: hospitals lacked objective visibility into workflows. In response, AVRICS developed a privacy-preserving thermal AI sensor that runs all inference at the edge, providing real-time room activity insights without capturing or transmitting identifiable data. Initially built to support the robot, the sensor evolved into a standalone product for measuring baseline inefficiencies and informing operational decisions.
Digital infrastructure is also limited. Key processes like disinfection, scheduling, and room tracking are often handled manually, leaving hospitals without reliable data to identify inefficiencies or evaluate improvements. This is especially problematic in operating rooms, where up to two hours per day are commonly lost to avoidable delays.
AVRICS was launched to address these issues by developing and commercializing a CE-marked UV disinfection robot, Violet, designed for safe use in occupied rooms. It delivers traceable, measurable disinfection through AI-powered dose control and optimized navigation. The project also developed foundational infrastructure involving hardware, software, user interfaces, and integration pathway to enable scalable adoption of autonomous systems in hospitals.
Five core objectives guided the project:
(1) Product Development: Deliver a hospital-ready, compliant robot.
(2) Evidence Generation: Validate safety, traceability, and efficacy.
(3) Market Discovery: Identify early adopters and refine GTM strategy.
(4) Adoption Pathway: Build a repeatable pilot-to-procurement model.
(5) Strategic Positioning: Prepare for global scale through team, partners, and investment.
During US market engagement, the team identified a broader need: hospitals lacked objective visibility into workflows. In response, AVRICS developed a privacy-preserving thermal AI sensor that runs all inference at the edge, providing real-time room activity insights without capturing or transmitting identifiable data. Initially built to support the robot, the sensor evolved into a standalone product for measuring baseline inefficiencies and informing operational decisions.
The project raised the technology from TRL 5–6 to TRL 8 and produced a robust, validated, and CE-marked solution. Work was structured across six dimensions:
(1) Hardware and Systems – Two commercial-grade robot versions were developed and manufactured. To support real-world integration, the team also created supporting tech: the AI thermal sensor and a wearable UV sensor for dose monitoring during co-presence.
(2) Autonomy and Optimization – Navigation software was validated to ISO standards. A simulation platform was used to model UV coverage, and AI trajectory planning was added to improve efficiency and coverage.
(3) Safety and Remote Ops – Real-time exposure monitoring, secure remote diagnostics, and over-the-air updates were deployed. Cybersecurity testing ensured compliance with hospital IT standards.
(4) User Interfaces – Interfaces included touchscreens, a mobile app, a web dashboard, and a prototype voice assistant, each co-developed with hospital staff to ensure usability.
(5) Scientific Validation – Field trials demonstrated consistent bioburden reduction. A computer vision tool automated colony counting, replacing manual methods with reproducible imaging.
(6) Go-to-Market – Pilots across Europe, the US, and Japan shaped a scalable commercialization framework. These efforts highlighted the value of the AI sensor and led to its refinement as a complementary product.
(1) Hardware and Systems – Two commercial-grade robot versions were developed and manufactured. To support real-world integration, the team also created supporting tech: the AI thermal sensor and a wearable UV sensor for dose monitoring during co-presence.
(2) Autonomy and Optimization – Navigation software was validated to ISO standards. A simulation platform was used to model UV coverage, and AI trajectory planning was added to improve efficiency and coverage.
(3) Safety and Remote Ops – Real-time exposure monitoring, secure remote diagnostics, and over-the-air updates were deployed. Cybersecurity testing ensured compliance with hospital IT standards.
(4) User Interfaces – Interfaces included touchscreens, a mobile app, a web dashboard, and a prototype voice assistant, each co-developed with hospital staff to ensure usability.
(5) Scientific Validation – Field trials demonstrated consistent bioburden reduction. A computer vision tool automated colony counting, replacing manual methods with reproducible imaging.
(6) Go-to-Market – Pilots across Europe, the US, and Japan shaped a scalable commercialization framework. These efforts highlighted the value of the AI sensor and led to its refinement as a complementary product.
AVRICS led to the development of two complementary technologies: a UV disinfecting robot and an AI-powered room monitoring sensors which together form a synergistic platform for both action (disinfection) and insight (workflow optimization).
The robot delivers:
(1) Safe co-presence: Operates in occupied rooms using selective irradiance and real-time UV exposure monitoring.
(2) Active dose control: A closed-loop system delivers precise UV doses using physics-based algorithms validated in peer-reviewed research.
(3) Optimized navigation: AI planners and digital twins select paths that reduce shadows and ensure effective coverage.
The sensor delivers:
(1) Thermal-based sensing: Provides strong privacy without video, suitable for sensitive areas like ORs.
(2) Edge AI: All processing is local, no data is stored or sent, making it easy to deploy securely.
The robot delivers:
(1) Safe co-presence: Operates in occupied rooms using selective irradiance and real-time UV exposure monitoring.
(2) Active dose control: A closed-loop system delivers precise UV doses using physics-based algorithms validated in peer-reviewed research.
(3) Optimized navigation: AI planners and digital twins select paths that reduce shadows and ensure effective coverage.
The sensor delivers:
(1) Thermal-based sensing: Provides strong privacy without video, suitable for sensitive areas like ORs.
(2) Edge AI: All processing is local, no data is stored or sent, making it easy to deploy securely.