This project tackled the automation of apple packaging, employing a robotic arm (iFactory uARM5) and a depth camera (ZED2) to identify, orient for optimal presentation (vibrant colours facing up), and carefully place apples into diverse packaging. This aims to increase efficiency, reduce manual labour, and improve the visual appeal of packaged fruit in agricultural processing.

A custom GPS-based security system was developed for a Jackal UGV, leveraging the ROS2 Nav2 framework for autonomous patrol missions. This initiative seeks to create a reliable and cost-effective solution for outdoor security, surveillance, and potentially initial incident response by enabling robots to navigate and monitor designated areas.

An innovative "observer rover" was designed for the challenging task of inspecting underground piping tunnels. Equipped with over 12 sensors, including LiDAR and a depth-camera, and managed by ROS 2, the rover generates detailed 3D maps and surveys, enabling remote assessment of infrastructure integrity and identification of potential maintenance needs, improving safety and efficiency.

This research focused on enhancing the mobility of Unitree A1/Go1 quadruped robots through advanced gait planning and control. By developing ROS/2 wrappers and utilizing the TOWR optimization framework, the project aims to enable these robots to navigate complex and uneven terrains with greater stability and efficiency, expanding their potential applications in diverse environments.

This project explored the use of Husky A200 equipped with various sensors to capture data for generating precise 3D models of vineyards. The goal is to provide valuable insights for precision agriculture, including detailed mapping, plant health monitoring, yield prediction, and optimization of resource management for improved vineyard productivity and sustainability.

The development of a retail service bot (ai image due to NDA) aimed to automate various tasks within retail spaces to enhance customer experience and operational efficiency. Potential functionalities include guiding customers, managing inventory, providing product information, and potentially even assisting with order fulfilment, showcasing the growing role of robotics in the retail sector.

This project served as a demonstration of how humanoid robots can be effectively utilized in logistics and warehousing. By showcasing tasks such as object manipulation, transportation, and interaction within human-designed environments, the demo aimed to highlight the potential for humanoid robots to augment or automate labour-intensive processes in supply chain operations.

The creation of a robotic bar focused on automating the process of preparing and serving drinks. This involves robotic arms for manipulation, precise dispensing mechanisms, and a control system to manage orders and create consistent beverages, potentially offering benefits in terms of speed, accuracy, and operational hours in the hospitality industry.

This project focuses on utilizing the ALOHA robot, a low-cost open-source hardware system designed for bimanual teleoperation. The system typically involves a human operator controlling a master set of robotic arms, which in turn precisely mirror their movements on a follower set of arms (the ALOHA robot itself). Often equipped with multiple cameras for enhanced perception, ALOHA enables the collection of high-quality real-world demonstrations for imitation learning in robotics.

The Robot Web Control Interface is a web-based platform designed to facilitate the control and monitoring of robotic systems. It typically includes features such as real-time video streaming, sensor data visualization, and control interfaces for various robotic components, enabling users to interact with and manage robots remotely through a user-friendly web interface.

R&D that automates robotic standards validation through simulated testing, verifying compliance, identifying missing functionalities and software bugs in heterogeneous robots.

Functional near-infrared spectroscopy (fNIR) detected finger taps from brain with a 92% test accuracy using a Deep Neural Network.

Design and fabrication of an automated wireless coil-winding machine for stator coils of 70cc motorcycles.

Research and development of automated property-based testing for the Toyota HSR robot’s perception, manipulation, and navigation stacks.

Implementation of a fault-tolerant manager that detects errors within the \href{https://robotnik.eu/products/mobile-robots/summit-xl-en-2/}{Summit XL robot} and executes the recovery process during a mission around an airport.

Development of entity detectors, image classifiers, text-extractors, text-summarizers, chatbots, form-fillers, text-preprocessors and more.