Task #1: Programming a Wire Cut
Our initial challenge for ChatGPT involved programming the Yaskawa robot to perform a wire cut. This is a very simple task. However, ChatGPT isn't intrinsically familiar with the INFORM programming language, which is integral to Yaskawa robots. As such, our first step was to delineate the fundamental commands of this language.
Furthermore, ChatGPT had no understanding of the physical robot, its movements, or the typical process of wire-cutting. To address this, we established several coordinates using the robot's teach pendant and outlined the basic principles of operation.
With these prerequisites met, we put forward our request for ChatGPT to create the required program. The AI successfully rose to the challenge, generating a program that we then transferred to the robot for a test run. The outcome was encouraging, with the robot effectively performing the wire-cutting task as directed.
However, it's important to note that creating the prompts for ChatGPT required a considerable amount of time and a deep understanding of robot programming. Without this specific knowledge, generating such a request would likely prove unfeasible.
Task #2: Programming a 'Square' Robot Path
Next, we challenged ChatGPT to write a program for the robot to draw a square. This task is a little more difficult than the previous one. This involved feeding the AI the necessary coordinates. The first run of the program, however, did not yield the expected result. Instead of a square, the robot drew a rectangle. We gave feedback to ChatGPT, emphasizing the mistake and asking for a revised program.
The AI responded with a new attempt. Nevertheless, the second try again resulted in a rectangle, not a square. It appeared that despite the feedback, ChatGPT was unable to accurately comprehend and execute the concept of a square.
We made adjustments to the set of coordinates and the robot's orientation and send the request once more. This time, the robot managed to draw a square, demonstrating that while the Large Language Models (LLMs) could improve its output, it required significant human intervention and guidance.
ChatGPT and Industrial Robotics: Still a Long Way to Go
The experiment provided some insights into the potential and limitations of ChatGPT in the realm of industrial robotics. For basic, clearly defined tasks, ChatGPT demonstrated some degree of competence. However, for more intricate assignments LLM was lacking.
Furthermore, the experiment underscored that ChatGPT, in its current form, is far from being a viable replacement for advanced software like ABAGY. The latter possesses the capability to intelligently utilize machine vision and adapt to unexpected changes in the geometry of parts, a feature that is currently absent in ChatGPT.
As LLMs continue to evolve, it will be interesting to see how they can complement and enhance industrial robotics solutions.