Simulation and offline programming

Industrial Robot

ISSN: 0143-991x

Article publication date: 28 August 2007

644

Citation

Young, K. (2007), "Simulation and offline programming", Industrial Robot, Vol. 34 No. 5. https://doi.org/10.1108/ir.2007.04934eaa.002

Publisher

:

Emerald Group Publishing Limited

Copyright © 2007, Emerald Group Publishing Limited


Simulation and offline programming

Robot simulation has been around for many years, but like many of the technologies that we take for granted today it has not always been quite so useful or accessible. My own experiences of it started in 1986 when I started work for BYG systems developing GRASP. At the time the software ran on either a mini computer the size of a wardrobe or one of the then new Unix-based workstations. Neither the software nor the hardware was cheap and given that the population of industrial robots in the UK was then 3,500 there were not many potential customers for it.

Even though this was early in the adoption of this technology, there were already issues with people who had been sold earlier solutions which did not deliver against expectations. These early solutions had text-based interfaces, used wire frame representations, could not do hidden line removal in real time and could not interface well with CAD systems. While it was possible to use them for off-line programming their main use was as a planning tool and even then they required an experienced user.

The technology has moved a long way since then. Systems run on PCs, display full solid models in real time, have graphical user interfaces, interface seamlessly to CAD databases, have application specific menus that allow complex programs to be created with minimum effort, allow the user to work in the robots native language so that none of the program is lost in translation, give exceedingly accurate predictions of cycle time and are significantly cheaper than they were. There are, however, still issues with using this technology.

Within the UK the population of robots has grown to an estimated 15,000 but the average price of these robots has fallen to between £10 and 30K. This makes the cost of a simulation system comparable to an extra robot. For most applications (the most common one in 2006 was handling) the programming is sufficiently simple as to not require it. At the other extreme robots are vision controlled and here the ability to program and simulate the interaction with the vision system would be required. To date offerings in this area are not useful as different vision systems have different programming languages and would be complex to program off-line. The same is true for other sensor driven applications, be they laser seam tracking or through arc tracking. This is a big limitation for arc welding where many systems require this form of sensory feedback in order to guarantee a quality product with variable components. In order to be useful in these situations the system would need to be able to simulate components at all points in the tolerance range and not just CAD nominal as is done at present. They would also need to be able to simulate the arc welding process. This would be incredibly complicated and beyond any system on the market today.

The biggest users of simulation are still in the automotive industry. Here the ability to check out large production lines for spot welding and adhesive laying offer massive advantages in terms of elimination of design errors. The ability to reuse existing designs is also a massive help in developing these systems. Other users exist, working in small niche areas where the ability to program complex tasks using CAD data can give more accuracy and much faster programming times. These users often work with the software provider to develop application solutions and what they end up with is a very simple user interface dedicated to their own process.

Interestingly most of the suppliers of this software that were in the market in 1990 are still around today. Some are now owned by main stream CAD vendors, showing that this technology is now seen as core. Many of the robot companies also offer their own packages though it is not clear what level of market penetration these have gained. Clearly in the future the need for simulation and off-line programming will only increase. These systems will require a fast development curve, however, to keep up to speed with the development of robotic technology and remain useful in all application areas. The ability to model product variation and sensor-based programming will be crucial. Advanced robotic functions such as servo compliance and co-operative manipulators may be simple to simulate but not easy to program off-line with confidence.

Ken Young Warwick University Manufactoring Group, International Manufacturing Centre, University of Warwick, Coventry, UK

Related articles