Science, technology & design

Self-sufficient, plant-eating, ‘Steam Punk’ robots

This is fabulous. Scientists working for the US army have created a wandering veggie-eating robot that can supply power both to it and to soldiers in remote regions. The robot – known as EATR (Energetically Autonomous Tactical Robot) is able to collect raw biomass such as leaves, wood, grass and convert these into fuel for its steam-powered engine. The robot uses smart software to tell what’s edible and what’s not and uses a laser-guided robotic arm to grab the biomass and put it in a hopper that connects with an internal combustion engine, which in turn powers an onboard battery. If plant material isn’t available the robot can be operated on petrol, diesel, kerosene or cooking oil. Why bother? Because soldiers need power. Within a few years every US soldier will get through the equivalent of 120 AA batteries each day, powering communications equipment and support devices and this power has to come from somewhere. There are other uses for such robots too. They could wander fields in search of landmines in war zones or they could be programmed to look for invasive weeds or lost walkers in national parks. According to EATR’s creator, Robert Finkelstein, Director of the Intelligent Systems Laboratory at the Clark School of Engineering at Maryland University, a fully functional foraging robot will be operational sometime in 2012. If you think this is all a bit too Sci-Fi it’s not. It’s merely a logical development of the domestic cleaning robots that already wander around private homes and offices cleaning up and then plugging themselves into a socket when they need more power.

Ubiquitous sensing

Back in the 1990s, David Gelernter, a professor of computer science at Yale University, wrote a book called Mirror Worlds. In it he described a world that had a digital reflection. Everything that existed in reality had a twin in a virtual reality. Now, thanks to the falling cost and increasing power of mobile computing, the idea is becoming a reality and the real world is converging with the virtual. One of the reasons for this is so-called smart infrastructure projects. For example, Hewlett Packard is developing what it calls a Nervous System for the Earth (CeNSE).

The idea is to scatter a trillion tiny sensors across the globe to create a ubiquitous sensing network. Such devices would be placed on office buildings, bridges, railways, and water pipes to create mountains of data that could help buildings monitor their own energy use or help bridges to find problems and issue warnings to engineers – possibly even to dispatch repair robots to fix their own faults. Intel is designing similar sensor ‘motes’ or ‘smart dust’, as is IBM. The primary reason for doing this is to make infrastructure more efficient and to save money. But what works on pipelines works for people too. For example, patients could be monitored at home and doctors called only when they are needed - or small children could be monitored 24/7 so that they didn’t get lost or get into trouble. However, there are problems. First, these sensors have to be cheap – almost free – if they are to be scattered by the trillion. Second, the sensors need to be rugged enough to cope with whatever Mother Nature or mankind throws at them. Third, there needs to be significant computing power available to crunch the vast amounts of data that these devices would produce. Fourth, sensors need power to function or at least to transmit data.

There’s also the issue of privacy. If you are walking around with a smartphone you are generating vast amounts of data, much of which is sent back to, and held by, the network. This can be useful. Many people already trade locational privacy for information. For example, ‘Layar’ is an augmented reality application. Switch on your phone’s video camera, point it at a street and you will receive information about the street, giving you, for example, user reviews of a restaurant in front of you or information that a house behind you is currently for sale. Google’s ‘Goggles’ can similarly identify books, paintings or landmarks from uploaded images. Perhaps that’s it. Rather than spending billions developing sensors why not simply tap into the 4 billion plus mobile phones currently roaming the earth and use people themselves (or at least the cameras, microphones and GPS they are walking around with) as the sensors? To some extent this is happening already but we are not really aware of it or the potential consequences. What if, for instance, governments start to use the data generated by such devices as an instrument of control? What if the data generated by individuals is not made available to individuals? What if everything becomes so connected that volatility and cascading failure become commonplace? What if people become so dependent on intelligent systems and devices that they stop thinking for themselves and defer important decisions to machines?

Word changing scientific ideas

Technology is one of the most powerful forces for change and is always expanding the limits of both our imagination and of reality. Having said this, few technologies end up reshaping the world - but a few do. Recently the World Wide Web and before that the transistor, the automobile and the steam engine all divided history into ‘before’ and ‘after’. So what new ideas are on the horizon that have the power to create large discontinuities? A report from Scientific American magazine suggests a few contenders. Here’s an extract. Game-ification is the idea that in the future almost every part of your life will include game-like experiences in order to make you healthier and more productive. In other words, we will use games to motivate and engage people and the playing of such games will provide data to make such experiences more and more compelling. Critical to this are technologies such as GPS and cheap sensors that provide data on personal behaviour.

Personally I find the whole idea rather creepy. I, for one, do not want my bathroom mirror to give me a scorecard on the health of my teeth every single morning or make me play a game to get me to brush my teeth a little longer. Furthermore, I do not want my shower reminding me that if I shut the water of in 30-seconds I will qualify for another discount coupon at my local supermarket. Much more up my street is the idea of cheap water. Given that one in six people still doesn’t have access to clean water, a filter that costs half a cent and can be used to purify a bottle of river or pond water seems like a good use of human brainpower and resources. There are two possible technologies that could make this idea a reality. The first is essentially a small bag filled with carbon granules coated with a microbicide. The clever bit is that the chemicals are encapsulated within nanofibres to increase the surface area, so a small filter can fit inside an ordinary bottleneck. The second option is to use woven cotton coated with electrically conducting carbon nanotubes and nanowires. The silver ions act as a bactericide and bacteria are also killed by an electric charge supplied by a couple of cheap batteries connected to the material. Next up is Gas from trash – modified microbes that are created to eat naturally occurring and man-made waste and then secrete (or excrete) fuel in the form of hydrocarbons. For example, we could develop a gene-altered bacterium that uses sunlight plus carbon dioxide to create the components of diesel fuel. This probably sounds a bit fanciful, but naturally occurring algae already turns carbon dioxide into fatty acids that can be refined into fuel. The only problem is that getting the algae out of the water and the fatty acids out of the algae takes up too much energy to make the process a commercial proposition in most cases. Other ideas on the list include a cheap diagnostic test to alert parents to more than 100 rare genetic diseases and wandering, plant-eating robots (see story above).

Bionic eyes and ears

When cochlear implants were first available they did little more than help people that were deaf vaguely understand what other people were saying by creating a series of basic sounds that aided lip reading. At the time this was a big leap forward, but two and a half decades on the technology seems almost quaint. Now cochlear implants allow deaf people to engage in telephone conversations or hear what’s being said in noisy environments. So could retinal implants do for blind people what cochlear implants did for deaf people? The answer is yes. A company called Second Sight, based in California, has created a commercial implant called Argus II for people suffering from retina-wasting diseases.

At the moment the technology only provides basic assistance, but given a decade or two it’s likely that these devices will have progressed at a similar rate to hearing devices. Around 100 retinal devices will be made available in the first year with trials starting in hospitals in London, Geneva and Paris and then the US, once FDA approval has been given. The cost of each device is around $100,000, which is obviously expensive, but early cochlear implants cost about the same. As for the future, it should eventually be possible to offer devices to anyone that wants one (or can afford one) and also for implants to include video cameras within each eye, at which point the whole of an individuals life presumably becomes recordable. The mind, as they say, boggles.

Smart roads

In the future most freeways won’t be free. Thanks to GPS, CCTV and other devices it’s easily possible to work out, in real time, how busy roads are and where particular cars are going. This means that vehicles can be charged for entering specific areas or fined for going where they shouldn’t. Congestion zones are already popular in some cities, but we should expect the idea to develop much further, with local councils and national governments putting a value on road use and charging road users in real time depending on time of day or volume of traffic. This is a good idea on some levels.

At the moment most individuals pay for roads through fuel taxation or vehicle registration. However, such schemes aren’t very smart and someone that drives infrequently can end up paying almost as much as a heavy user. In theory, charging drivers per-mile or per-minute is a better idea, although don’t expect the other forms of charging to disappear. In the Netherlands the government recently devised a plan to charge all road users on a per-kilometre basis by 2012. In a road test, 70% of drivers changed their behaviour due to charging. The recent collapse of the Dutch government has scrapped this idea for the time being, but it is likely to return sooner or later. Similarly, Norwich Union (a UK insurance company now known as Aviva) ran a test whereby vehicles were fitted with technology so that insurance could be charged according to road use and even driving style. In the end the scheme proved too expensive to administer but accident claims did fall by around 30% during the trial. So what’s next? Expect to be charged to use all motorways, bridges and tunnels in the not too distant future with fast lanes incurring additional costs. Secondary routes and minor roads will be free for the foreseeable future but even these may eventually be fee paying.