About politics of networks..

It is a great initiative by  UK Engineering and Physical Sciences Research Council ( EPSRC ) to fund the UK Robotics and Autonomous System Network. UK robotics community being young, has the advantage of learning from each other as well as from expensive mistakes done by groups in other countries with a longer history for robotics.

The common mistake of judgment – play to media or slow down to address hard problems? By trying to be “attractive” to media, some groups in other countries more known for robotics made robots that looked like humans that could walk with bent knees on flat floors, wave at audiences, shake hands with politicians, etc., but failed outside calibrated environments. Some robots could even play violins but went nuts if a string broke! They of course brought more funding in the short run, but lost public trust in the long run. Recent DARPA grand challenge is an eye-opener to all of us that the robotics community has a long way to go to understand the basic secrets of the nature of computation of perception and action by biped walkers. In fact, these problems are not limited to biped walkers, but common in all robots including medical robots that are expected to be efficient and stable during interactions in uncertain environments.

The hard truth: The answers to these pressing problems have to be found by doing simple and basic experiments that people like Newton did few centuries ago to understand basic laws of dynamics, and what people like Marc Raibert did few decades ago at MIT by rolling a simple rimless wheel down a ramp to see what causes variability in passive dynamic walking. We have to re-visit the fundamental nature of embodied computation and recognize that it transcends programs running in microprocessors to something that also encompasses the whole body and the environment.

We at the Centre for Robotics Research (CoRe) in the Department of Informatics, King’s College London (KCL) have worked with the belief that breakthroughs in basic research in robotic sensing, morphology, perception, internal representation, and actuation will help to advance the field of Robotics by learning from other countries with longer histories for robotics. Since 2003, members of CoRe have consistently ranked the highest in the UK to publish in the most prestigious international robotics conferences with 45 publications in the IEEE International Conference on Robotics and Automation (ICRA) and 26 publications in the IEEE International Conference on Intelligent Robots and Systems (IROS). Those who are in the field of Robotics know how hard it is to publish in the above two flagship conferences of the IEEE Robotics and Automation Society. In addition, we have published 50 top quality journal publications in IEEE transactions related to robotics (ToR). With other papers, members of CoRe have published 177 peer reviewed top quality papers since 2003.

We have also been part of a large number of EU funded robotics and autonomous systems projects like HANDLE, CONPHIRMER, TOMSY, DARWIN, COSMOS, STIFF-FLOP, SQUIRREL, and FourbyThere.  When it comes to impact, the 2015 Research Excellence Framework (REF) would have immediately shown our place as a university at 7th in terms of quality and 6th in terms of research power, and as a department according to the power score, we achieved position 8 out of 89 institutions. In terms of specific projects, STIFF-FLOP is testing the surgical robot in human cadavers, and details of other projects are online.

We have 8-academic staff members, each supervising on average 5-PhD students in robotics. The Centre also enjoys a team of on average 15-postdoctoral research associates. This makes our Centre one of the largest Robotics groups in UK. Our members hold Associate Editor positions in IEEE Transactions on Robotics, IEEE transactions on Fuzzy Systems, ASME Journal of Mechanisms and Robotics, and leading IEEE robotics conferences like ICRA and IROS.

Why do founding visions of Robotics Networks matter? To be successful, ideally the network should encompass the industry and social interest groups as proper members from the beginning. Then, the network should be allowed to grow on its own with energetic members taking the lead without imposing hard barriers like “capital grants”. When we launched the London Robotics Network, the Centre for Robotics Research, KCL, partnered withImperial College London, Shadow Robotics Company, and the UK Knowledge Transfer Network to be founding members. This way, we opened the doors to all those who are interested in robotics to come on-board. The result was the recent expansion of the network to the South Eastern Robotics Network.

However, establishing a formal National level Robotics and Automation network sponsored by the country’s premier fundamental research funding agency – EPSRC – has a different gravity, and should have clarity of founding principles. The criteria to be founding members should clearly reflect academic excellence extent of international engagement. Preferably, the criteria should respect international norms used to invite founding members, so that the international image of established robotics groups are not damaged.

Disappointment: It is disappointing to notice that the EPSRC funded UK Robotics and Autonomous Systems Network does not include established large robotics groups like the Centre for Robotics Research at King’s College London in the founding list of institutes. This damages the reputation of our centre in the absence of clarity in the press releases. We were told by the UK RAS network that the criterion to be a founding member was to be Capital Expenditure and Doctoral Training Centre grant recipients from EPSRC. We disagree with this criterion to be one that reflects academic excellence and international engagement in Robotics. Moreover, this criterion can lead to the potentially dangerous precedence where future funding decisions in Robotics will be based on a flawed image claimed by the founding institutes. We are a bit frustrated as to why members in the list of the founding members did not notice the fundamental flaw in this criterion.

The damage can be reduced if EPSRC corrects the press release with a clear reference to the above criterion used. Then we do not have to explain to international robotics colleagues as to why our centre was not among the founding list of members. They will understand that UK has a right to do things their own way even by defying international norms used to invite research centres to be founding members of a National Robotics and Automation network.