Writing a blog about my work and interests has always been something I wanted to do. Clearly my last attempt at that was poor, however, now that I have a new job at Optalysys, I am reinvigorated and ready to update more regularly.
Its a short entry, mostly to act a new beginning point for the future.
Taros 2017 was a Huge successes and I have enjoyed every minute. The even kicked of strong with Professor Yang Gao FIET FRAeS opening the conference with a great insight into Surreys involvement in robotic space projects and the U.K. as a whole. As it turns out we have plenty to be proud of from Surrey, with several successfully launched payloads and mission involvements. It really does look as though robotics will pay a large part in future space exploration and management. In particular ,the debris removal project looked interesting, with net based debris remove and a harpoon like device for retrieval. I felt this really did show off what modern robotics can achieve.
The keynote for the first day given by Nicolas Heess, Senior Research Scientist, Google
Deep Mind. He shows us several interesting project from google that involved reinforcement learning and deep learning for robotic control. Most relevant to my interests, was a robotic arms picking up and placing Lego bricks and stacked then neatly. The arm would learning through policy based learning, to pick up and stack Lego bricks. I still have some convincing to do when it come to reinforcement learning for control, I still have worries about the limits at play and the somewhat brute force method to learning complex tasks, however, this talk goes a long way in showing off what deep learning can achieve in robotic control.
For me the highlight was the keynote on day 3 given by Rob Buckingham, Director of Remote Applications in Challenging Environments (RACE), UK Atomic Energy Authority. His talk centred around JET, a power station that will be completely run by robots, much of which will be autonomous. The advantages that are gain here are numerous, including operating in harsher room conditions, (such as temperature and radiation) and more accurate precise controlling for repair and operation. In this highly regulated and safety conscious environment computational models and virtual testing will now count as validation of your work, hearing Dr Buckingham talk about the challenges of REAL world robotics gave me pause for though about how I conduct my own research and testing.
Finally I would like to thank all of the people that gave oral presentations, these were insightful and well delivered. I would also like to thanks every one that was involved in the organisation of this event and ensure them that the time was valuable and rewarding. I will be looking forward to my next conference in robotics, Hopefully IROS in the future.
Any question about the conference or my paper please do get in touch, I love to hear form people I can learn from or collaborate with. As always, thanks for your time reading.