Kinectbot

The folks at the MIT personal robotics group didn't waste any time getting the Kinect depth camera working with a mobile robot, with this only approximately a week after the sensor was hacked.



Achieving good spatial awareness has been a fundamental obstacle for robotics, and in the foreseeable future this could be a solved problem. Whilst experimenting with humanoids, such as the Rodney robot, it became obvious to me that unless the robot has a good level of awareness of the shape of it's surroundings then its intelligence and range of skills remains highly constrained. Once 3D information is available then detecting large objects such as people, walls, doors, chairs and desks becomes a far easier proposition than is the case with 2D vision.

This is also a demonstration of the power of building upon earlier work, using existing open source algorithms wherever possible. A big problem in the past was that robotics software would usually be re-invented from scratch, with little or no transferability of software from one system to another either due to hardware incompatibilities or because the software was regarded as being secret or only released under restrictive licences. Being able to get a rapid result such as the above means using software components, such as the SLAM algorithms, which themselves will have taken many years to research and develop.

Don't believe the cyberwar hype

One problem caused by the rise of the world wide web and 24 hour news is that news items are now often generated with very little fact checking or journalistic analysis.

A fairly good example of this is a story about an alleged Chinese web traffic hijack.  There's a lot of suppositions in this story, and when faced with such notions it's a good idea to apply occam's razor.  Is it more likely that:


(a) Chinese technicians made a configuration error, then corrected it 18 minutes later.

or

(b) The Chinese government conspired to hijack a vast amount of internet data, in order to perform unspecified data mining or tampering on it.

In the absence of any other information, I know which hypothesis makes the fewest assumptions.

There's also some fearmongering in the article about the Stuxnet virus.  The appearance of viruses which are specifically targeted at industrial infrastructure, as opposed to the usual constituency of hapless internet users, is cause for concern.  However, it's important to note that even though considerable expertise, time and probably financial resources may have been required in order to create Stuxnet, and that there does seem to be credible information suggesting that it was targeted at a specific location (it only affects inverter drives whose manufacturers are based in Finland and Iran), as far as we can tell it was completely and utterly unsuccessful at causing any non-trivial amount of industrial disruption.  This should indicate something to us about the difficulty and likelihood of any nation being able to launch a cyberwar capable of causing significant inconveniences to an opponent in a targeted way which doesn't simply degrade the whole internet globally.  The introduction if IPv6 should make malware writers (whether they're state sponsored or not) jobs even harder, since the address range becomes so incredibly gigantic that brute force attacks really have enormous odds stacked against their success.

Measuring virtual objects

A video showing that depth cameras can also be used to take measurements from objects fairly accurately.



Being able to make such measurements easily with a low cost sensor should mean that 3D models of objects or people can be made very easily by non-experts, then used in games or online environments such as Second Life, without needing to have any expertise in using complicated CAD tools.

Another related application of interest would be the ability to generate models from objects for use with rapid prototyping (RepRap) type devices. This could already be done using shape from silhouette techniques or with lasers, but a depth camera probably makes it even easier.

It seems likely that before long these sensors, or variations on the theme, will be available on all games consoles and probably also available separately for PCs. One possible consequence might be the emergence of the "3D web", which was something which was much anticipated in the mid 1990s, but never happened. Back at a time around 1996/7 I had a VRML web page, where the clickable items were 3D objects such as a spinning globe. One of the reasons for the failure of the 3D web was the poor rendering performance of the hardware which was available at the time, and this is now no longer an issue, but the main reason was that it's difficult for the average user - or even a quite knowledgeable user - to generate 3D content of their own in a way which doesn't require a considerable investment of time and effort.

A possible scenario might be that you take your object and present it to the camera, rotate it a few times, and then you have a 3D model automatically extracted and with any missing details filled in which can be posted onto ebay, all in the space of one or two minutes.

Does this mean that the game is up for stereo vision ?

For indoor scenarios in which mobile robots might typically operate perhaps it does.  But there are still some big unknowns, and they are how well these systems perform in natural lighting, and also how they perform when two or more such systems are operating within the same space.  For the latter problem possibly slightly different IR wavelengths could be used to prevent interference.

For outdoor situations, which would include automotive or agricultural systems, I think it's possible that this kind of sensing might not work very well unless you're pumping out large amounts of energy, in which case bleedover and interference becomes more of an issue.  So within those realms stereo vision or multi view stereo might still remain the best option in terms of cost and performance.

Tobin tax

A news item about the possible introduction of a Tobin tax (aka "Robin Hood" tax). It's mildly frustrating in these kind of discussions that they never consider the financial system from a cybernetics perspective.

The financial system as we know it now is a bit like steam engines prior to use of mechanical governors. There's nothing to damp out fluctuations or prevent the whole system from blowing up. Transaction taxes levied on each trade could provide the principal mechanism which really does make boom and bust economics a thing of the past - or at least turn tidal waves into minor ripples.

It's probably true that there will always be speculators making speculative trades in this or that, and that sometimes these trading decisions will be bad, based upon superstition, faulty information, the vagaries of fashion or bad maths.  Occasionally such bad decisions will compound according to positive feedback to generate a bubble. But the financial system is not a force of nature outside of our influence. It's an artificial system devised by humans, and as such can be subject to the moderating influences of laws, regulations, taxes and so on.  We don't have to become victims of our own limited outlook or tendency to follow the latest fad, and I expect that future generations will look back on the economic booms and collapses of the past with a combination of amusement and horror.

Philosophy and Science

The Kinect hack extravaganza

Not long after it's release the depth camera on Microsoft's new game console has been hacked, with an open source driver being available for Linux.



It's a similar situation to the early days of webcams. When the Connectix Quickcam was first released around 1996 there was no open source driver or API available, and people also wanted to try using it for robotics.

It would be good to have someone do an all round test of one of these cameras under different lighting conditions and in different environments, to see what kind of limitations exist. How well it copes with natural lighting is one of the main concerns.

But even if you don't care about robotics there are also other more traditional applications for depth cameras. In a video conferencing scenario - the typical use case for a webcam - you could have not just a flat image but a 3D model of the users being displayed, and if you tie this in with face detection to estimate head pose then you could have a situation where it really appears that the screen is a window through which you're looking at the other person in quite a natural way.

Mob mentality

Watching some of the news coverage of rioting students this seems to confirm the apprehension of myself and others about what might be on the horizon over the next few years.  If the economy doesn't pick up and unemployment remains stubbornly at acutely high levels there could be a protracted period of ongoing social unrest, similar to the various strikes and riots which took place in the 1980s, except with civil servants substituting for miners.

Students, their fees, and the decommissioning or loss of funding of some academic departments are probably only a small component of what looks like quite an extensive laundry list of potential problems.

• Inflation is higher than the official target of 2%.

• Soon the VAT increase to 20% will automatically inflate most prices by an additional 2.5%. 

• Fuel prices remain at levels similar to those during the 2008 oil price spike.  There may be longer term benefits if fuel prices remain high, but in the short term when considered in conjunction with other factors it's likely to lead to further discontent.

• If upper limits are not centrally imposed then local taxes seem likely to rise, as local councils attempt to recoup some of the budget shortfalls coming from central government. 

•  Cuts in police budgets may mean that they're less able to respond to any unrest when it occurs.

• Even for those workers who havn't taken voluntary pay cuts, disposable income and the purchasing power of their pounds will relatively decrease as other costs rise, so this is bound to add inflationary pressures into the system as people seek pay rises to try to keep their living standards on an even keel.

There is probably more malaise than I've listed here, and I havn't touched on the problems of the now retiring baby boomers and their pensions, but that's enough collapsitarianism to be going on with for the moment.

So I expect more than just students to protest, and people should protest about matters of importance, but if at some time in the next few years you find yourself (in the style of the Talking Heads song Once in a Lifetime) in a mob which is becoming angry and violent and taking an unscheduled detour off the planned protest route, as the students did today, then my advice is to make a swift exit from the scene.  Avoid populist rhetoric, elaborate conspiracy theories and extremists claiming to have easy answers to complicated problems.

Also in addition to the budget deficit there's a more general democratic deficit.  This doesn't look like a situation which you can easily vote your way out of, and there's a lack of political biodiversity.  The problems we have now originate from the previous Blair/Brown administration, and looking at the photo from today of an anarchist kicking in a window - which will presumably become an iconic image from 2010 - you can trace that smashed window back to Gordon Brown's mistaken belief that the era of "boom and bust economics" was over, his failure to recognise that house prices were getting seriously out of line with average earnings and his belief in "light touch regulation" of the financial industry and the mystical "financial innovation" which turned out to be not much more than fraud, bad maths, or some arcane mixture thereof.

Neither the Conservatives nor the Liberals seem to have much by way of solutions to offer, and talk of a "Big Society" only seems like so much airy rhetoric - extremely hard to pin down to anything practical.  So there is discontent and distrust of all the mainstream political parties, compounded by the expenses scandals of 2009 and ongoing investigations into the alleged criminality of former politicians accused of stealing public money.  At this point it seems that it would be very easy for extremist elements to step in and appear to offer a way out of the impasse.

Marvin Minsky's robot arm

Considering the relatively primitive state of technology in 1968 compared to what's available today this is a fairly impressive attempt at building an articulated robot appendage. This would have been at a time when there was still a great deal of optimism about AI and its rate of progress, where nothing like this had been done previously, and prior to the production of the Lighthill report and Perceptrons.



It looks like the arm is hydraulically actuated, since pneumatics probably wouldn't be strong enough to hold the weight of a child. The complicated kinematics, which would normally require quite a bit of processing power, are bypassed by having a small model which contains a few potentiometers, with Minsky manually solving the inverse problem.

Ewok rampage

A system called PTAMM (Parallel Tracking and Multiple Mapping) looks interesting, and this type of structure from motion augmented reality is closely related to robotic visual perception systems.  Here is an example:



Unfortunately the license associated with this system makes no sense.

"No part of the Software may be reproduced, modified, transmitted or transferred in any form or by any means, electronic or mechanical, without the express permission of Isis"

Ok, so this isn't open source. No wait...

"Isis’s permission is not required if the said reproduction, modification, transmission or transference is done without financial return, the conditions of this Licence are imposed upon the receiver of the product, and all original and amended source code is included in any transmitted product."

So it is kinda open source.  ish.

"You may be held legally responsible for any copyright infringement that is caused or encouraged by your failure to abide by these terms and conditions."

So it's a pseudo-open "throw code over the wall" sort of endeavour, but they can sue you if you make any derivative work which according to them "encourages" financial return.  That's pretty vague.  Any redistributed modification could constitute encouragement to do something or other, dependent upon the prejudices or theories of the observer.  Probably best not to touch this code with a barge pole.

This highlights the perils of "roll your own" software licenses.  In such cases there's a good chance that the language used doesn't make sense, is self-contradictory, or may not actually be legally valid if it came to the crunch.  Probably they would be better off either keeping PTAMM as a closed proprietary system, or else using a standard open source software license which preferably is OSI approved.

Within the academic sphere, of which this could be a part, there might be a gap in the pantheon of licenses for a validation only license.  That is, where the code is open for inspection and independent use, but only specifically for the purposes of checking that particular algorithms described in an academic publication do what it's claimed they should.

Robot carers

A BBC article about robots which might be used for care of the elderly.  Rather than having sceptical comments from caring organisations a better strategy would be to collaborate with them in the research.

If I were someone doing this research probably the way I'd approach it would be to consult social services organisations who work in care homes on exactly what it is that they do on a daily basis, and I'd also do some time and motion studies for different scenarios.  Only then can you identify what the relevant needs are and which of those might be amenable to automation.  My guess is that having an anthropomorphic talking head, although it's very mediagenic, is not a high priority in this line of work.

Mobile robots probably do have their place, and will be useful for communication, picking things up and moving things around.  Even more lucrative though will be wearable technology.  A device which you can wear which continuously monitors your heart rate and blood pressure, and which can alert a nurse or doctor automatically in an emergency would be useful and probably also commercially viable.  Wearable devices which assist mobility are also the lowest hanging fruit in terms of possible commercial robotics opportunities for the elderly.  An exoskeleton-like device which could be strapped to the legs which assists in walking or standing up, with at least equivalent support functionality to a walking stick would have no problems in selling if it could be manufactured in the right price range - even if it were a relatively primitive electromechanical type of system.

The reputation economy

Receiving "points" for carrying out certain tasks has been quite well established in the domain of retail brand loyalty schemes.  People will sometimes go to extraordinary lengths to receive "points" of incredibly low economic value, often amounting to tiny fractions of a pence each.  There's now a proposal to extend a similar points based system to casual labour.

If the "points" have some non-trivial economic value then this may be a good idea, basically clearing out a lot of the bureaucracy surrounding the process of hiring casual labour.  This isn't really the same as the "big society" idea, which is based upon the notion of volunteering without economic reward.  I think the big society will fail to boost volunteerism by a significant percentage (most people are not highly altruistic), but that altering the economy to be more focussed around alternative currencies - such as a person's reputation - may be more successful.

It might be a better idea to have a single, officially approved, reputation system which is exchangeable for goods and services, rather like money.  Performing certain tasks either boosts or decreases your reputation, with the votes of other people being used to decide who has a higher or lower reputation.  In this way even if structural unemployment rises due to automation people who are unemployed in the conventional sense will still be encouraged to maintain their reputation and engage in activities which enhance their social standing in the opinion of others.

An advantage of a reputation based points system is that you can't borrow your own reputation from others and get into reputational debt.  You have to build up your reputation from your own actions, and antisocial activities will damage your reputation and reduce its economic value.  Also, reputation systems should be non-zero sum, such that if I increase my reputation this doesn't mean that someone else is excluded from having the opportunity to also do the same, although relative competition would still exist.

It's worth remembering that there is nothing inalienable or inherently valuable about money as we now know it.  It's just an imaginary man-made system, and could be replaced by other kinds of system which in some way indicate value and facilitate the fair distribution of goods and services.  New technology and ubiquitous internet access may mean that value representations more complex than physical cash (metal or paper) or credit cards can be implemented and made practical.