Belated prophecy

Citation

Loughlin, C. (2002), "Belated prophecy", Industrial Robot, Vol. 29 No. 5. https://doi.org/10.1108/ir.2002.04929eaa.001

Belated prophecy

Belated prophecy

About eight years ago (IR 21,2) I mentioned (in reference to GPS systems) that mobile robots did not need to know their position in the world but only their position in their world. And I called at the time for the equivalent of an indoor local GPS system.

I would like to be able to claim that Jan Fäger of MEEQ AB had read my Viewpoint and was instantly inspired to produce the PosEye system described in our feature "Does your robot really know where it is?" (Monica Schofield, pp. 407-11). Unfortunately I cannot make this claim because he patented the idea about 12 years earlier.

His method of localisation uses vision and heavy number crunching, rather than precise timing signals, and his stated accuracy of 0.1mm is a lot better than the 5m claimed for the satellite systems (although interestingly the resolution is comparable). Of course other localisation systems exist, and we have covered quite a few in this journal over the years; but the real beauty of the PosEye system is that it has no moving parts and that the really clever bits are purely computational. And as we all know, computation is becoming cheaper and faster all the time.

Rough ride

If you want to see a picture of something that I would never do, just take a look at Plate 1 in "Robots give rides at Hanover Fair"(Anna Kochan, pp. 399-403). The prospect of being thrown around on the end of a robot arm (and I mean any robot, not just the manufacturer concerned) fills me with dread. I did not see the system myself, but if it is theoretically possible for human limbs to be smashed into any of the robot's surroundings then I would not like to be providing the insurance cover for the system. This could be an age thing. My first memories of robot behaviour include a Puma 500 going wild on passing through a discontinuity. Divide by zero on your pocket calculator and a harmless "E" for error symbol appears. Divide by zero as part of a Denavit-Hartenberg robot positioning transformation matrix, and weird and scary things can happen.

I am certain that safety will have been uppermost in the minds of the engineers who designed the system, and I am sure that the system is as safe as it can possibly be made to be, and the actual risk is probably very small indeed. Robots are after all, considerably more reliable now than they were at first, with MTBF measured in tens of thousands of hours. However, to me the whole thing looks like a bad idea if for no other reason that if an accident were to happen, and even if the robot was not at fault at all, the effect on the robot industry would be disastrous.

Am I being over cautious? Probably. I am also inconsistent because I look forward to a time when people are in regular and intimate contact with service robots without any safety barriers to get in the way. Perhaps it just comes down to the magnitude of the forces available in the Robocoaster ride.

I also welcome the introduction of medical robots where "intimate contact" reaches a new set of heights. But with medical robotics the risks are balanced by the greater risk of the solely human alternatives; and the potential gains are considerably more worthwhile than a five minute adrenalin rush.

One major reason for the widespread introduction of robots in Japan is the cartoon character "Atom" (see "The humanoid robot scene in Japan" – Yoshihiro Kusuda, pp. 412-19) who was very friendly towards people and went around rescuing them. This has engendered a strong feeling of trust and "comfort" emotions between the Japanese and their robots that has only been heightened by the robotic toys which are now flooding the market.

Here in the West we have Robocop, Terminator and Robot Wars in which extreme violence is the main attraction. So I suppose that Robocoaster is just another perfectly logical step down the same path. The question is, whether or not this is a path we should be going down at all?

Clive Loughlin