Thoughts on an Autonomous Lock Picking ROBOT.
Introduction.
There are many types of locks, but to open them we have to follow a sequence of operations. Many of these operations can be simulated mechanically and driven electrically. Ideally the process needs to be controlled by a computer or a micro-processor, the size of a small finger nail. The reason why I would choose a micro-processor controlled machine is that it is very easy to program it. Mechanical computers tend to be very fixed in what they can do.
Specifically I would choose an 18F452 PIC microprocessor controlling one or more stepper motors which manipulate the lock.
Lock Types
Dial combination locks. These are probably the easiest to interface with as the lock only has a single easily accessible knob, which requires manipulating. These are the kind of combination locks one see's on safes, so it is very common to assume that they are 'secure', but it can be shown that this 'security' can be an illusion. For a more complete explanation enter the search term "combination locks" into Wikipedia. As the article mentions, there are mathematical relationships within the lock, that can be further exploited, to greatly reduce the number of combinations that have to be 'tried' to open the lock. A micro-processor can handle this easily and at a much faster speed than a human. Also a stepper motor can accurately manipulate the lock dial much faster than a human. Computer controlled lock picking ROBOT's are usually only seen at National security levels, but there is one single example on the Internet at ...
http://www.youtube.com/watch?v=6BrzOKK2exg
The lock in question could be opened manually in seconds, but that is not the point !. It is constructed from Fischertechnik components which is the German equivalent of Lego.
Cylinder locks. With a cylinder lock we have to control two things, first the torsion on the plug and second, find the binding order or the pins. Once those variables are known a ROBOT can be programmed to replicate the manual movements required to open the lock. The BIG problem is that a lot of lock picking is a sub-conscious exercise and the human often finds it difficult to put into words what they actually think is happening. Unless the manual process is fully understood it is impossible to program the micro-processor to duplicate that process automatically.
I had a chat last week with a professional time served locksmith and we were discussing the possibility, of sensing the torque wrench feedback, to detect both the binding order and individual pins clearing the shear way. As you know the amount of feedback can be very small (except with false sets) and it was interesting because we had independently come up with the idea of using a LASER diode on the torsion tool and project the beam onto a remote surface, which would greatly magnify any feedback movements on the torsion tool. Apart from that, if you can get a pick into the lock to pick it, then it would also be possible to attach that pick to two stepper motors, one for pin selection and the other for pick movement on each pin. As mentioned earlier, false sets tend to give much different torsion tool feedback that pins clearing the shear line and also lock opening is easier still to detect with an opto switch, detecting the free movement of the torsion tool at lock opening..
Mortise locks. The required torsion, on the few cheap door type mortise locks I have had a look at, has varied from 'heavy' to 'very heavy' and that is probably why they have a 'T' bar on the torsion tool !. Hopefully it is less on the better quality locks. As for picking the levers, the main problem I found was accurately placing the picking 'finger' on the individual levers, as the lever and spacer pack in cheap locks felt very 'floppy'. As mentioned elsewhere, the respective thicknesses of the lever and spacers meant that the width of the picking wire had to be less than 1.3mm. Now a stepper motor is easily capable of this kind of accuracy and could do the picking very quickly, but the problem is the floppy lever stack. I have not looked too deeply yet but there may be simple solutions for this such as using the edges of the outer lever/spacers as reference points, for the stepper motor.
Some people may think that it is a lot of work when the locks can be picked manually anyway ... but again I say that is not the point !.
Pad locks. All of the afore mentioned points also apply to padlocks, but in addition they have another problem and that is the bolt spring can be so powerful as to mask the binding order, so one may have to program the micro-processor to apply tension to the hasp until the pins start to bind and then adjust to the changing conditions inside the lock as pins clear the shear line.
Conclusions.
You like picking locks .... I like picking locks AND playing with ROBOT's. I am not suggesting that a lock picking ROBOT would be practical, or have any place in the commercial environment, it is simply an interesting challenge.