CEE Seminar - Integration and Control of Pattern Recognition, Computer Vision and Robotics in Immersive Platforms

This seminar summarizes human-computer interfaces for quality inspection aided by robotics applications, enhancing human decision-making with Artificial-Intelligence (AI) and machine-enabled visual analysis. Accordingly, interfaces of image-based automatic defect-detection and visual inspectors are explored focusing on Augmented Reality (AR) systems. A standalone human interface for automatic defect detection and localization is developed by integrating an image-based pattern recognition algorithm in AR headset's platform. The interface also quantifies anomalies in the pixel units using a horizontal measurement algorithm improved by the edge gradient analysis and employs a nonstationary pixel unit conversion algorithm to translate the size of anomalies to the engineering units. Additionally, human interfaces for intuitive robot programming are presented. Compared to other approaches, AR interfaces enhance precision, reduce setup time, and allow for more intuitive interaction. This seminar also summarizes the research developing a new interactive interface for intuitive robot programming that allows humans to adjust the kinematic controller's parameters or use a different outer-loop controller that is an alternative to the time-consuming adjustments of the holograms. Automatic robot control in human proximity is examined where human intention prediction is essential for safe robot planning. This automated control with the AR interface open a bi-lateral communication line between humans and robots. The practical application and validation of the model is examined through experiments related to inspections and manufacturing involving human upper limb movement in interaction with robots.