The Intelligent Cognitive Assistance Systems (INCASE) is the product of a successful Enterprise Ireland Innovation Partnership between The Technological University of the Shannon, Cook Medical Ltd, and Vistamed Ltd. The aim was to identify solutions to problems those companies were facing in relation to the high costs of quality inspection, training of staff and responding to batch-of-one. We worked collaboratively for four years designing a solution driven by their needs. 


This is the feature that has had the ‘wow’ factor in demonstrations, and the one that IRDG said ‘our network would love this’. INCASE creates a digital history record (DHR) in video for each individual unit produced. This record is proof that units were produced correctly within this factory. The seamless manner in which this is introduced without any input from an operator is highly favourable and has the potential to set an industry standard. Closing the loop to a manufacturing execution system (MES) by data integration massively reduces the manual effort in sign-off and signature record keeping. Using a unique algorithm this solution stores the DHR at a fraction of the storage requirement of other competitor solutions, and provides a rich set of tagged data for later analysis. So not only can customers say they’ve built the product correctly, they can prove it. This is invaluable for demonstrating an incredibly high level of conformance to the rest of the value chain, provides proof in litigation should there be a claim against defective parts, and avoids the necessity for a very expensive full batch recall in the event that one unit in that batch was found to be faulty as that faulty unit can be uniquely identified. Alternatively if a unit should fail a quality check, the video record of the mistakes made is available to identify how that happened and intervene with training or adapted processes to ensure it doesn’t happen again.


INCASE performs real-time assembly quality checks and instructions delivery. Using a novel method captured by know-how the system is able to predict with 99.94% accuracy whether the work is being done in accordance with the process requirements. INCASE uses this method to assist the operator by providing feedback if they are not completing the process correctly. Therefore the operator has the opportunity to fix the issue immediately before the next stage in the process. Thus quality checking is performed in real time. It reduces the costly overheads of post-assembly quality and integrity checks, and is more responsive than post production sampling necessary in a fast-changing environment to deliver zero-defect approaches. Production steps that do not follow the recommended production process or where the system calculates a lower confidence in its checks, are flagged as being products that require a quality check by the quality team. Thus, quality efforts are directed at those units not randoms selection.


INCASE promotes seamless and intelligent integration between itself and the operator and provides simple intuitive access to the SOP. The operator is provided with just-in-time context-specific instructions as to the next step in production. Thus training and support is real-time. This enables flexible manufacturing. In addition, we propose that we can reduce the amount of time required for training. The solution also learns from the operator, using a unique method for the neural net to train itself. Thus tacit knowledge is captured which in turn assist understanding of new ways in which to work and improve production. 


We thank our ongoing partners for there continued support.