Presentations

• “Dynamic Trajectory Planning for Robotic Manipulation in Unstructured Environments” (poster) – AgriFoRwArdS CDT Annual Conference 2024: Robots in Action [July 2024] – Norwich, UK.
• “Exploration of LLM-Enhanced State-Machine function-calls for Planning Robot Actions” (oral) – AgriFoRwArdS CDT Summer School: Robotic Phenotyping [July 2024] – Wageningen, The Netherlands.
• “Adaptive Hybrid Control Architectures for LLM-Controlled Robot Arm” (poster) – AgriFoRwArdS CDT Summer School: Robotic Phenotyping [July 2024] – Wageningen, The Netherlands.

Activities and Outputs

• Member of the AgriFoRwArdS CDT Summer School 2024 Organisation Committee (Feb to Jul 2024)

MSc Project

Adaptive Hybrid Control Architectures for LLM-Controlled Robot Arm

This project seeks to develop an adaptive control architecture that leverages the capabilities of Large Language Models (LLMs) for enhanced decision-making. By dynamically selecting and switching between various learning or planning methods—such as reinforcement learning and imitation learning—based on real-time environmental data, the architecture will enable robots to perform complex tasks more efficiently in unpredictable settings. The initiative promises to improve robotic performance, thereby enabling robots to understand and execute task instructions with unprecedented flexibility and accuracy.

PhD Project

Developing an Intelligent Soft Robotic Manipulator for High Efficient and Automated Cross – Breeding and Phenotypic Analysis in Agriculture

The primary aim of this PhD project is to develop an advanced robotic manipulator system specifically designed for the delicate handling of soft and fragile agricultural products, such as berries and various food items. This project seeks to improve current agricultural and food processing practices by enhancing robotic perception and learning capabilities, enabling precise and gentle manipulation that minimizes damage and improves operational efficiency. The key objectives of the project are: (1) to design and develop soft, adaptive end-effectors capable of handling delicate objects; (2) to integrate tactile and visual sensory systems for providing real-time feedback on the texture and firmness of objects; and (3) to apply deep learning techniques to refine the robotic perception and decision-making processes.

The training environment for this PhD project will be rich and interdisciplinary, providing comprehensive exposure to advanced robotics, machine learning, and agrifood technology. The candidate will work within BIRL, a robotics laboratory equipped with the latest tools and technologies necessary for designing, building, and testing robotic systems. This includes access to high-precision robotic arms, advanced tactile and visual sensors, rapid prototyping tools such as 3D printers, and high-performance computing systems for training and deploying machine learning models. The PhD candidate will also benefit from collaboration with a diverse team of experts in robotics, and machine learning . Regular interaction with these experts will enhance the candidate’s knowledge and skills, fostering a deep understanding of both theoretical and practical aspects of the project. Furthermore, the candidate will have the opportunity to attend specialized training sessions, workshops, and conferences to stay abreast of the latest developments in the field and to network with other researchers and professionals.

In addition to technical training, the project will also emphasize the development of essential soft skills, including project management, interdisciplinary collaboration, and effective communication. The candidate will engage in regular project meetings, presenting progress reports and receiving feedback to refine their work. This holistic training approach will ensure that the candidate is well-prepared for a successful career in academia or industry.

Amanda’s PhD project is being carried out under the primary supervision of Prof Fumiya Iida.