Cohort 3 ART-AI student, Doug Tilley, has had a busy few months in Vienna and China. He writes about what he has been up to:
“Back in November, I attended the 25th Anniversary IEEE Sensors Conference 2023 in Vienna, Austria from Saturday 28th October to Wednesday 1st November. The conference runs annually and is a global conference intended to act as a forum for scientists, industry and engineers to collaborate and discuss their latest research and applications on a wide range of topics to do with sensing and sensing technologies. The different topics range from sensor design and material properties to artificial intelligence, machine learning, robotics and data processing. Near the end of the conference, a 25th anniversary dinner was held at the Hofburg. This was a black tie event, with a show of traditional ball room dancing, live music as well as a range of different local delicacies.
I was presenting my conference poster and paper titled: “Shallow Hierarchical CNN-LSTM for Activity Recognition to Integrate Postural Transition States”, the full paper has been published and can be found in IEEE Xplore in the following link: https://ieeexplore.ieee.org/abstract/document/10325044. My paper was peer-reviewed and ranked in the top 5% of the papers at the conference and was invited to be expanded and invited to the special issue which I am in the process of writing. The paper aims to account for some of the issues occurred when using unbalanced datasets for machine learning and by increasing the accuracy of human intention with the use of shallow models making them more accessible for use in prosthetics. By using shallow models, we reduce the size and complexity of models to potentially enable them to be deployed on edge devices. This can make them usable in offline methods and hence, real time applications for robotics without requiring a large amount of compute. The poster was received well, with interest from a host of different researchers, as well as producing some collaboration opportunities with others. I was attending with my supervisor Dr Uriel Martinez Hernandez who was presenting a journal article titled “Soft Tactile Sensor with Multimodal Data Processing for Texture Recognition” and presenting for Gregory West in his paper titled “Towards Low-cost Plastic Recognition using Machine Learning and Multi-spectra Near-infrared Sensor”, Dr Gorkem Anil-Al presenting his paper “Multimodal Barometric and Inertial Measurement Unit-Based Tactile Sensor for Robot Control”, and Professor Manuchehr Soleimani presenting his paper: “Shape Self-Sensing With Mutual Inductance Sensor Array”. This was a good team activity representing the Department of Electronics & Electrical Engineering, Department of Computer Science, ART-AI CDT and the University of Bath. In my paper, the corresponding acknowledgements to ART-AI and the University of Bath were also included.
Outside of our presentations, I had the honour to meet some of the leading researchers in the field on the use of multimodality and IMUs for activity recognition. The majority of lunch time presentations were rife with discussion and meeting researchers in the field.
After the conference in Vienna, I left straight away to go to China for a 3-month academic Exchange at the Academy of Medical Engineering and Translational Medicine, Tianjin University. Tianjin University is the oldest university in China and has a well established reputation for work and collaboration. For the academic exchange, under the supervision of Professor Lin Meng, I focused on leading a team to conduct three multimodal studies: Human Activity Recognition, Cognitive Walking and a Sit to Stand study. The dataset is composed of a wide range of sensors which requires a vast amount of processing. The dataset aims to use multimodality to help increase classification accuracy in prosthetics to reduce harm to patients, but also allow a comparison between different ranges of sensors that are currently used. The study has been fantastic and we have achieved our viable criteria, hoping to process and publish both the results of our processing as well as the dataset within the next few years. It has been a fantastic opportunity to work and collaborate with others, improving on my interpersonal skills as well as help support neural network development. Furthermore, this activity contributes to strengthen the academic and research links between the UK and China for future academic exchanges.
China is an incredible place, with rich culture, and I am honoured to have visited a few different areas and cities. There is a wide range of incredible food with local curiosity for new people, with an interest in the cultural differences that we all have. Over my stay, I managed to see the Great Wall of China, Beijing’s imperial palace district as well as Shanghai’s Central Business District. Three months felt like too short of a time to be able to explore properly, as leaving just weekends to explore meant I couldn’t go to too many areas off the beaten path. I fully intend to visit China again, both in an academic sense and as a tourist to cover some of the other great areas of culture. Whilst Christmas isn’t typically celebrated, the New Years celebrations were great with plenty of fireworks displays and as I was getting ready to head back, everyone was getting ready for their spring festival break (Chinese New Year), where there was some incredible decorations and sculpture work going on as well as plenty of celebration. There is such a wide range of beautiful architecture, and diverse foods. The food had incredible flavours as well as being a nice social occasion, such as with Hot Pot. Although, the Sichuan province food was very spicy! Some of the plates were basically only chillis… But it was also a lovely experience to try a range of different provincial food such as those from Shaanxi province, though I had a slight favourite, TangHulu. The university is an incredible place to work with a collaborative and positive work culture. In the future, I am hoping to carry on the collaboration with Tianjin University.
