Telerobotics---force feedback master-slave manipultor

Sytem analysis - to work out what has to be done BEFORE even thinking of which hardware, language, OS or other implementation factors.

Do some top down design of system requirements and expand each section until you know what you require. Start with the mechabical drive systems how fast they will be, how fast they respond to change and all those limitations before going any further.

Without knowing lots of details about mechanical limitations, speed of response, how this is to be controlled (human operator, automatic sequence, etc..), anything from an 8 bit micro to a Cray could do the job. Hopw well it will work is not knowable.

As Paul carpenter has already indicated, you need to do some analysis work to improve the quality and depth of your system requirements. Start with the end-effector outputs and work towards the inputs. You have a human operator in the loop so don't forget to include his/her actions as part of the focus of system analysis. This link might help your thoughts on that a little.

Additionally, you will need to look at the force sensing mechanisms that will be useful to you (motor current, strain gauges, etc). Which technology you choose will depend on your mechanical constraints.

Lots of different ones, depending on what you're trying to accomplish, and what your constraints are.

To start, the ability to identify what you're trying to accomplish, and what your constraints are.

I picked up system design skills by osmosis over the years. There is a formal association of systems design engineers who describe the job the same way I do -- you may want to dig up their web site and see what they recommend.

That's nice -- what's a bilateral position and force loop? One of the first thing you have to learn as a systems engineer is that the language is different from one specialty to another, and it's your job to understand who speaks what, and try to express things so that all the disciplines involved will understand what you mean without ambiguity.

The most important steps in identifying a microprocessor are going to be to identify the various measurements and actuators you need to take to make this work, the degree of precision that you need in your various measurements and actuators, the bandwidths that you need to sustain, and the complexity of the controller (paying particular attention to whether the math involved needs to include potentially "expensive" operations like divide or trig functions or log or exponentiation, etc).

From that information you can derive the necessary data path widths, sampling rates, and execution complexity. From _that_ you can get an idea of what the microprocessor core needs to be capable of, and then you can go shopping for a micro that has the necessary horsepower and a nice set of peripherals.

My advise, if this is going to be a one-off and you're not limited in size, weight, or power consumption -- and particularly if this is a student project -- is to get the BIGGEST MICROPROCESSOR YOU CAN AFFORD. Using 5% of the processor capability to get the job done is merely wasteful. Needing 400% more capability than your processor has means that your project has failed.

For old fashioned manipulators there are a few:

McCloy, Harris "Robotics An Introduction" Open University Press 1986 ( Has a very readable chapter on manipulators )

Vertut, Coiffet "Teleoperation and Telerobotics" "Evolution and development" Volume 3A Kogan Press 1985 "Applications and Technology" Volume 3B Kogan Press 1985 ( author: French nuclear industry ; comprehensive but a bit hard to read )

Köhler "Typenbuch der Manipulatoren / Manipulator Type Book" Thiemig 1981 ( author: german nuclear industry ; a collection of specs and photos of commercially available systems of its day ; but only few detailed descriptions )

Thring "Robots and Telechirs" ellis horwood 1983 ( Thring is a somewhat idiosyncratic but likable fellow ; the book is readable and a usefull introduction )

Vintage: NASA SP-5047 "Teleoperators and Human augmentation" 1967 NASA SP-5070 "Teleoperator Controls" 1968 NASA SP-5081 "Advancements in Teleoperator Systems" 1970 Later some of that was published as: Johnson, Corliss "Human Factors Applications in Teleoperator Design and Operation" Wiley 1971

But i guess for students a Matlab simulation will do ...

MfG JRD

or you can just use matlab/simulnk or scilab/scicos with for example

Bye Jack