[Closed] Help me learn about manufacturing robotics.

Never teach them religion.....
You didn't mention Battlestar Galactica!

I am well aware of Moores law and Asimov's rules. There was a robot in Battlestar???? I only noticed Wilma!!! But really what I want is to know about is what kind of kit is likely to transform 1st world manufacturing in the next 5-15 years. Deployment of robots in a mixed human/robotic environment etc.

I think too much emphasis is put on 'smart' robots. You want them to carry out the mundane repetitive and boring tasks 1st and foremost. We don't mind doing the interesting stuff ourselves. Saying that AI scares the crap out of me and I don't think they have signed up to Asimov's laws.

"robotics" is very vague and far reaching. Could be anything from small mobile platforms that are at most a trip hazard, up to large Kuka arms that could throw you through the air or snap you in half!

Surely you know what you're doing/wanting to do with robots? Give us a clue...

(i work in robotics and autonomy R&D, less on the manufacturing side, but I'm less clueless than most)

I mean what could go wrong...

robotics sans politics = disaster for the working man.

IA (and others). The kind of robotics I am interested in (I think) are automated pick and place, and robots that can either transport and move stuff or be easily configurable for a range of jobs that would have been done by hand previously. So info on ease of programming/adaptability. Sensor technology to allow safe integration without excessive guarding. Drive systems (motors,servos, stepper motors?) reliability and maintenance. General robots (and I think I am talking arms) that can be turned to a wide range of adaptable uses. Also interested in things like the cool robotic conveyor belt (that uses camera systems and 360 degree steering to move and sort products).

I work in automation. I moved from a pure c++ programming role into it. Quite different but interesting if I can help I will. Pick and place is quite simple.

Pure pick and place might be quite simple, but what if you want the same robot to do 150 different products and configurations and be easy to configure/teach for the 151st. I know camera systems are coming into self driving cars, has much been done in manufacturing environment to use these type of systems instead of light beams and pressure sensors for safety?

Define a robot. I would consider a CNC mill/turn machine to be a robot as much as a traditional robotic arm. We use them when we need the flexibility to manufacture several similar components on the same production line. So you have several (say 8 or so) CNC machines in a cell. These are high speed machines, so you might have a grinder, drilling machine (laser or mechanical), milling machine and maybe some form of inspection machine. The component arrives on a conveyor belt, paced up by a robotic arm, a dot matrix code etched on the fixture or the component itself is read and the system manages several components in the cell at the same time in terms of sequencing, moving them from machine to machine. There is no AI involved, its all 'dumb' programmed. There is some clever algorithms going on in the sequencing and planning of the choreography of the movement of the components between machines when you've got more than one component across different part numbers being worked at the same time in the cell to minimise dwell or down time and maintain the production rate.

The other area where robotics are used where I work is with additive machining and our Carbon Fibre lay up machines that lay up carbon fibre strand by strand rather than laying down woven sheets in a 1970's fibre glass stylee.

However again it's all 'dumb' programming, no need for AI. Not sure why you would want or need AI in manufacturing. Manufacturing is all about doing the same thing repetitively removing all opportunities for variability.

Depends what you need to know. Though i'm a Manufacturing Engineer by trade i've not been working in Manufacturing Engineering for about 15 years now, but still spend some time on the shop floor keeping upto date with the latest bits of kit and methods. The last project I was working on was with adaptive machining on 5 axis CNC milling machine (which I consider to be a robot) measuring 3D surfaces and adapting the machining for each individual component and features on a single component. This was to optimise the performance of the component by 'designing and manufacturing' features on the fly e.g. if the throat area of a space between two features had to be a certain amount the pre-machined features would be measured and CNC machine code automatically altered to adapt the machining. But that's all old hat by now but was quite interesting at the time.

If you're doing a proper project or something for a qualification at a College or Uni or something then I can probably put you into contact with some of the proper boffins where I work, but if its just for your own amusement or work then I might be able to get you some material or pointers of where to go to get some further info.

I've got a bit of experience with a small ABB six axis arm. We thought it may be interesting to see what one could do in the digital printing field. It was good fun learning how to program it and getting it to interface with other kit. There's a lot of info on the ABB website, and a forum which may give more info.

thanks wobbliscott. Main areas of interest are in mixed use robots. I have seen specific robots (mainly arms) used with great effect in end of line (palletisation, stacking etc). I have also seen (again not cutting edge) visual inspection camera systems where products are inspected in real time to track and eliminate defects.
I suspect that with increases in CPU processing power, and better cheaper camera systems (like the thing in a PS4 motion camera) that robots will be more easily configurable to picking, packing and forming materials. I am fairly familiar with CNC machines, and definitely consider them a robot, the ability to not just repeat a set of pre given instructions, but to use sensors and situational awareness to stop if unsafe, and adapt to different roles and have positive feedback of succsess of operation n before commencing step n+1.

Thanks Richmars. The ABB website is very interesting!

I work in a pharmaceutical manufacturing environment with lots of pick and place robots. They're pretty much bread and butter now. We do have a team working on collaborative robots which are said to be the next big thing for us. No "zero access" guarding...designed to work in the same space as human co-workers. I'll try to find some vids.....

Love those baxter robots, and some great applications. I will have to Google how they keep safe in a collaborative environment.

British automaton and robotic association (Bara) for some information and safety requirements.
Just don't stand in front of a class a say "search for bra".
Programing doesn't look much harder than cnc. Just position moves linked together. Though this should be 3 rotational co-ordinates for a jointed arm, we did have a cartisen robot on an injection moulder.

Edit: Collaborative robots look interesting, adds to next lesson(s).

You've not clarified why you're interested, which makes it hard to help. You're asking very open ended questions (though maybe don't realise it).

E.g. Why the focus on what makes robots safe - really you should ask what makes software safe. I can write software to spot a human, and avoid moving an arm into them - but then the question is how safe is that software? How reliable? Will it always detect and avoid in time?

However there is some work making robots designed to be safe in that they're elastic, enough strength to do a useful task but some give to keep humans safe. HRI (human robot interaction) is also a big field of study (that I've worked in). I know there's research within the ERC (Edinburgh and Herriot watt) into collaborative robotics for manufacturing tasks.

And the software isn't just about guards, it's about understanding intention - what is that person going to do? How do you help them? Avoid them? Keep them safe?

Programing doesn't look much harder than cnc. Just position moves linked together.

Gets harder when it's not set moves, but "pick that thing" and you have unconstrained space with obstacles to move the arm through -solving fast enough to obstacle avoid is hard.

Programming the ABB arm isn't too bad, you manually jog it to each position and record that point. The problem is, as IA mentions, you can't just enter the start and end point, you have to think about what's in the way (or may be in the way), and how you want to approach the target.
There's also the programming of the thing on the end, which may be an actuator that you need to open, or a paint dispenser.
I'm no expert, but I enjoyed the small amount of programming I've done.

There an entire standard for safety of machinery. Your control software should not be dealing with this. It is a different system. Look at piltz for examples.

Re using cameras you tend to have a dedicated processee for this. The cameras have larger sensors rather than relying on hi pixel count. You would buy a camera with the right reduction etc for ytapplication