Automation surprise (AS) has often been associated with aviation safety incidents. Although numerous laboratory studies have been conducted, few data are available from routine flight operations. A survey among a representative sample of 200 Dutch airline pilots was used to determine the prevalence of AS and the severity of its consequences, and to test some of the factors leading to AS. Results show that AS is a relatively widespread phenomenon that occurs three times per year per pilot on average but rarely has serious consequences. In less than 10% of the AS cases that were reviewed, an undesired aircraft state was induced. Reportable occurrences are estimated to occur only once every 1–3 years per pilot. Factors leading to a higher prevalence of AS include less flying experience, increasing complexity of the flight control mode, and flight duty periods of over 8 hr. It is concluded that AS is a manifestation of system and interface complexity rather than cognitive errors.
As a logical consequence of the advancements in automation of production of composite aircraft structures, more attention is paid to the automation of maintenance. Current repair procedures involve manual labour and exposure to harmful particles (such as dust, vapours) while final quality and evidencing depends largely on the skills of repair technicians. The current study aims to automate composite repair procedures for the aviation sector with the objective to counter these disadvantages. Main research question: ‘What is required for a robot system to assist in composite repairs’This research is part of a larger, SIA-RAAK funded project FIXAR, running in three Universities of Applied Sciences in the Netherlands and a cluster of knowledge institutions and industry partners.In the repair process of aircraft structures, repair by means of scarf or lap joints is common practice. First paint layers must be removed to inspect the area and prepare for further repair. Then damaged material is removed. Material is replaced and the repair is finished and painted. Tasks within the repair process that are considered dull or harmful are sanding and material removal. Current investigation focussed on automation of these tasks.
The increasing amount of electronic waste (e-waste) urgently requires the use of innovative solutions within the circular economy models in this industry. Sorting of e-waste in a proper manner are essential for the recovery of valuable materials and minimizing environmental problems. The conventional e-waste sorting models are time-consuming processes, which involve laborious manual classification of complex and diverse electronic components. Moreover, the sector is lacking in skilled labor, thus making automation in sorting procedures is an urgent necessity. The project “AdapSort: Adaptive AI for Sorting E-Waste” aims to develop an adaptable AI-based system for optimal and efficient e-waste sorting. The project combines deep learning object detection algorithms with open-world vision-language models to enable adaptive AI models that incorporate operator feedback as part of a continuous learning process. The project initiates with problem analysis, including use case definition, requirement specification, and collection of labeled image data. AI models will be trained and deployed on edge devices for real-time sorting and scalability. Then, the feasibility of developing adaptive AI models that capture the state-of-the-art open-world vision-language models will be investigated. The human-in-the-loop learning is an important feature of this phase, wherein the user is enabled to provide ongoing feedback about how to refine the model further. An interface will be constructed to enable human intervention to facilitate real-time improvement of classification accuracy and sorting of different items. Finally, the project will deliver a proof of concept for the AI-based sorter, validated through selected use cases in collaboration with industrial partners. By integrating AI with human feedback, this project aims to facilitate e-waste management and serve as a foundation for larger projects.
De glastuinbouw in Nederland is wereldwijd toonaangevend en loopt voorop in automatisering en data-gedreven bedrijfsvoering. Voor de data-gedreven teelt wordt, naast het monitoren van de kas-parameters ook het monitoren van gewasparameters steeds meer gevraagd. De sector is daarbij vooral geïnteresseerd in niet-destructieve, contactloze en persoonsonafhankelijk monitoring van gewassen. Optische sensortechnologie, zoals spectrale afbeeldingstechnologie, kan veel waardevolle informatie opleveren over de staat van een gewas of vrucht, bijvoorbeeld over het suikergehalte, maar ook de aanwezigheid van plantziektes of insecten. Echter is dit vaak een te kostbare oplossing voor zowel de technologiebedrijven die oplossingen leveren als voor de telers zelf. In dit project onderzoeken wij de mogelijkheid om spectrale beeldvorming tegen lagere kosten te realiseren. Het beoogde resultaat is een prototype van een instrument dat tegen lage kosten met spectrale beeldvorming een of meerdere gewaseigenschappen kan kwantificeren. Realisatie van dit prototype heeft een sterke Fotonica-component (expertise Haagse Hogeschool) maakt gebruik van Machine Learning (expertise perClass) en is bedoeld voor toepassing op scout robots in de glastuinbouw (expertise Mythronics). Een betaalbare oplossing betekent in potentie voor de teler een betere controle over kwaliteit van het gewas en automatisering voor detectie van ziekte-uitbraken. Bij een succesvol prototype kan deze innovatie leiden tot betere voedselkwaliteit en minder verspilling in de glastuinbouw.
Various companies in diagnostic testing struggle with the same “valley of death” challenge. In order to further develop their sensing application, they rely on the technological readiness of easy and reproducible read-out systems. Photonic chips can be very sensitive sensors and can be made application-specific when coated with a properly chosen bio-functionalized layer. Here the challenge lies in the optical coupling of the active components (light source and detector) to the (disposable) photonic sensor chip. For the technology to be commercially viable, the price of the disposable photonic sensor chip should be as low as possible. The coupling of light from the source to the photonic sensor chip and back to the detectors requires a positioning accuracy of less than 1 micrometer, which is a tremendous challenge. In this research proposal, we want to investigate which of the six degrees of freedom (three translational and three rotational) are the most crucial when aligning photonic sensor chips with the external active components. Knowing these degrees of freedom and their respective range we can develop and test an automated alignment tool which can realize photonic sensor chip alignment reproducibly and fully autonomously. The consortium with expertise and contributions in the value chain of photonics interfacing, system and mechanical engineering will investigate a two-step solution. This solution comprises a passive pre-alignment step (a mechanical stop determines the position), followed by an active alignment step (an algorithm moves the source to the optimal position with respect to the chip). The results will be integrated into a demonstrator that performs an automated procedure that aligns a passive photonic chip with a terminal that contains the active components. The demonstrator is successful if adequate optical coupling of the passive photonic chip with the external active components is realized fully automatically, without the need of operator intervention.
