%page_title%

Industrial robots, most often the heavy 6-axis robots, have been operating in production plants around the world for decades. They are responsible for many production processes, handling, palletizing, spot welding, arc welding, gluing, etc. The preparation of a robotic production process involves several stages. However, when it comes to programming the trajectory of robots, there are two, offline and online.

Offline programming is done with the help of dedicated CAD / CAM software. These are complex tools provided by leading manufacturers of software and instrumentation for the industry. While this software allows you to simulate the entire production process, despite the accuracy of the models and a fairly good representation of the real process, in most cases we are still doomed to programming robots on the spot, on the finished line. This is for a number of reasons.

Time and accuracy are at the premium
The main factor responsible for the inaccurate representation of the finished line in relation to the simulated model is time. Unrealistic deadlines for particular stages of designing (or bad planning) and execution of the production line cause their undesirable overlapping. This affects the accuracy of modeling individual elements that do not have a 1: 1 mapping in the real world. It is easy to imagine that the tightly programmed trajectory of the robot’s movement in such a situation may in real conditions collide with an element of the device that has not been modeled.

Still inaccurate modeling methods
Despite the continuous development of the above-mentioned software and modeling algorithms, we are constantly dealing with places where we are helpless. An example is how the robot’s cable package behaves, or other moving parts equipped with loose cable bundles or wires. Due to the randomness of their behavior, in order to avoid their damage at the stage of line start-up, it is still necessary to manually check the robot’s runs.

Online programming is not only about moving the robot
It is a mistake to think that the people who start the line on site are only responsible for uploading to the robot the programs prepared at the simulation stage. The online programmer’s tasks also include commissioning the entire robot, first aligning the axes, configuring and balancing the tool, starting the application, etc. It is also the adjustment of programs to the applicable standard, when it comes to automotive concerns or other industries where such a standard applies. An important element that seldom matches the assumptions and results of the simulation is the cycle time. Its optimization is also the responsibility of the online programmer and is performed on site during the observation of a working line.

Are we doomed to online programming and corrections?
The answer to this question is not simple. The quotation of the above examples shows how many components of the preparation of the entire robotic production process are. Nevertheless, the direction of CAD / CAM software development or better and better measurement methods show that within a few years it may be possible to move away from online programming.

More and more companies responsible for the design and execution of production lines try to do it modularly, i.e. individual fragments are designed and made in their own halls, and then everything is disassembled and put back at the customer’s place. This reduces the online stage to a minimum as the entire design office and programmers are on site and corrections can be made offline. While this approach allows for the delivery of a finished product and the elimination of all problems before its launch on site, it does not guarantee that the line built at the customer’s site will match perfectly with what the company had at its place. It is mainly about assembly methods that do not allow us to rebuild everything with the desired accuracy.

However, we can also deal with this today. Advanced scanning and measurement methods allow us to perform a secondary measurement, i.e. after building the line, it is precisely measured again and the result is compared with the component placement in the simulation. Then the simulation is adjusted to the realities of the finished line, and the robot programs are corrected and transferred to the robot without having to check them manually. Thanks to this approach, it was possible to eliminate online trajectory programming from spray applications where accuracy is particularly important. Online programming is also absent in many industries where robots perform very precise work, such as the assembly of electronic components. It should be remembered, however, that we are only talking about trajectory programming and about robots with a different degree of accuracy. These are also applications where precision is more important than the execution time and the whole process can be carefully prepared. The problem is extensive production lines with a large number of 6-axis robots with high capacity.

The melody of the distant future, when it comes to the most popular and mass applications of robots, are vision systems combined with AI algorithms, which will allow the robot to learn its tasks on an ongoing basis or correct the trajectory to achieve the best cycle time.

We already have the technology and tools to move away from online programming. However, the correlation of all components of this process is still a problem. Once these difficulties are overcome, online robot programming will be a thing of the past.