10106914Reach Industries – LumiMan – Revolutionising process automation throughout medicine manufacturing, powered by computer vision, machine learning and data analytics.ClosedCollaborative R&DInnovate UKLaboratory work is notoriously difficult to replicate. This happens for many different reasons. Small changes in conditions can affect reaction progress; individuals all carry out processes in very slightly different ways and many observations and measurements can be subjective. Reach help control the variability commonly seen in laboratory work through our novel laboratory monitoring and information management system, which improves reproducibility and thus increases efficiency. Our system captures reaction information via specialist computer-run cameras. This data (such as reaction colour, liquid level) is combined with other information we monitor (ambient temperature, humidity, pressure) and processed by our novel machine learning software to provide comprehensive feedback about how laboratory work is progressing. In this project we will develop our systems for use in scale-up labs, to enable the accurate tracking of a critical step in medicines manufacturing, where subtle changes in processes could result in the loss of £000s worth of materials and time. By developing the capability to scan and build 3D models of work environments and extending object recognition to work from multiple angles we will be able to build an end-to-end tracking system for objects in typical workspaces. Seemingly trivial differences in processes, such as the time a mixture is stirred, or the rate of the stirring, can be identified, to ensure that consistency is maintained. Should issues occur during scaling up activities, e.g. repetitions giving inconsistent results, the causes can be identified. Monitoring data can also be integrated with quality control systems to allow contract manufacturers to provide audit-quality data to stakeholders. Working with our project partners CatSci, we will implement the system to monitor oligonucleotide production processes and demonstrate technical feasibility for use of Reach technology in medicines manufacturing scale-up