130416 – RooBot – London > words
In September 1994 I began a lecture at UCL with two images shown side by side. One image was of a male body builder, his huge arms crossed in front of acres of chest. The other image was of a female ballet dancer standing on one pointe with her other leg reaching into the air way above her head. It was an introduction to a talk on ‘Extreme Climates and Responsive Systems’ and the analogy represented by the images was immediate. The body builder represented the machines of the Industrial Revolution and the ballet dancer the machines that we needed today. The ballet dancer was fuel efficient, had a high power to weight ratio, had uniform strength throughout a full range of movement, used biomechanics over brute force, had coordination and could interact with other complex movement patterns, could be choreographed to work as part of a complex integrated team, could adapt quickly to a range of landscapes and environments. More importantly the ballet dancer was female as Gaia is female and very much part of a circular system. Biomechanics and circular systems would be the inspiration for the next decades work on ever more delicate machines that responded to the harsher climates of our planet.
We jump forward twenty years and robots are at last coming to a department store near you. Robots have been a recurring Sci-Fi topic for hundreds of years from earliest automata’s of the Renaissance to early twentieth century films such as Metropolis. The 1950’s saw a new push for automated friends and assistants as American directors popularised science fiction with films such as Forbidden Planet and The Day The Earth Stood Still. Many visions of the future from these films had man and robot working together but the realities of reaching this goal have proved a little more difficult to resolve. The computing age has made vast strides forward in the development of robotics. The seamless link between hardware and software has until now been allusive. Many early robots were simply machines doing automated repetitive tasks. The next generation will be robots that can multi-task and make elementary decision upon those tasks i.e. how to pick things up, how much pressure to apply, how to balance, how and where to put things down, how quickly to move, how to avoid obstacles etc. The majority of funded research explores aspects of one or more of these tasks i.e. Boston Dynamics location based environmental sensors, balance and balance correction. Others such as Pepper and Erica explore the interface of man and machine with their emotive recognition and responsive robots.
A lot of these projects are still based on pattern recognition and the ability to churn through huge quantities of data instantaneously. We have come a long way since Alan Turing’s work but the majority of modern robots are still Turing machines albeit very sophisticated ones. AI will enable robots to learn and this will improve their ability to make decisions. In many ways this will still be data crunching and statistical analysis but all done at such speed that it would seem like a conscience decision expressed by a piece of hardware, the robot. The Tech firms are throwing huge budgets at this. Facebook recently set up Building 8 to work on hardware that can utilise its software and data harvested from its billions of users. As early simple products come to market the revenues generated will head straight back to the R&D departments for use on the next big project. The process will snowball and progress will be quick and its progression logarithmic. This is all happening now and the journey is well underway. Facebook’s ten-year roadmap may well be reached in six.
A kangaroo can travel huge distances with great economical efficiency at cruise speeds that average 20 kph (12 mph). Kangaroos can, when required reach speeds of up to 65 kph (40 mph). When travelling at speed the Kangaroos huge tail works as a whipping counterbalance to the head and body. This forms a pendulum motion over the hips where head and tail move up and down in unison helping to create lift. The Kangaroos rear legs have elasticated tendons that stretch from the back of the knee to the underside of the toe. The tendon is stretched over an exaggerated and lengthened heal. The tendon is an energy storage mechanism. It recycles the energy used upon landing and stores it in the elasticated tendon. The heal adds leverage to the storage system. Approximately 50% of kangaroos jump makes use of this recycled energy with the rear legs powerful muscles providing the remainder. A kangaroo increases its momentum not by increasing the speed at which its hops but instead by increasing the length of each stride. In this way, combined with its recycled energy, the kangaroo uses almost the same amount of energy at whatever speed it travels. This soft biomechanical motion is a fluid integration of total body movement and will be very difficult to replicate with a machine.
Festo is a multi national that makes hydraulic and pneumatic components for industry. Festo has an R&D department that uses its technologies to experiment with biomorphic machines. This R&D is mainly used as both training for its young engineers and marketing for its products and as such presents a conflict of interest between marketing and research. Machines here are made to resemble animals and they may have some attributes of that animal but resemblance dominates. For example the Bionic Kangaroo (their name, I’ll call it a RooBot) has the plastic body of a kangaroo with a plastic tail but it is a hopping machine. It could be a frog or it could be a fluffy white cloud. The RooBot like the kangaroo has an elastic tendon but this does not recycle energy as does the kangaroos, its energy is pre loaded by a standard pneumatic cylinder DNSU fed by a high pressure storage unit and a valve. The RooBot’s tail and head are plastic add-ons that do nothing other than suggest form. The RooBot can hop, it can turn and it can balance all very commendable achievements but this is not M.I.T. Media Labs. The RooBot is a sophisticated marketing toy posing as research and here is the dilemma. Although one is aware that marketing of existing products dominates the research into biomorphics Festo’s R&D has still produced some incredibly poetic pieces. The Air Mantra is an exquisite art piece and shoals of them ‘swim’ the galleries of the world. The Festo engineers have produced other mechanical fish that have the same mesmerising beauty and the RooBot suggests very similar potentials.
So what is the point of this meandering text? The film Blade Runner (adapted from Philip K Dick’s Do Androids Dream Of Electronic Sheep) was set in Los Angeles 2019, three years from now. So somehow within the next three years we have to move on from the RooBot to the Nexus 6 perfect Replicant, unlikely indeed. However, what all these pieces do is reaffirm just how incredible and unique the biological perfection that exists on our planet is and how much we have to learn from all nature that surround us.
If I had wish for one superpower it would be for immortality just to be able to watch this fantastic journey unfold.
The Surrogate Twin