How do you teach a robot to do its job?

Updated on : January 17, 2022 by Blake Hawkins



How do you teach a robot to do its job?

How do you teach a robot to do its job?

Robot implies a machine that assumes a function previously performed by a human. That leaves a giant blurry area. Do we define a tractor as a robot? Here I include an automation kit that has certainly taken over the human work, but it doesn't look like a 2-arm bipod.

4 ways I can think of: scheduled, performed, supervised, and unsupervised.

Programmed - The machine is built and programmed to perform a specific function with just a few physical controls. The least flexible technique. Control software written to do something or perform a

Keep reading

How do you teach a robot to do its job?

Robot implies a machine that assumes a function previously performed by a human. That leaves a giant blurry area. Do we define a tractor as a robot? Here I include an automation kit that has certainly taken over the human work, but it doesn't look like a 2-arm bipod.

4 ways I can think of: scheduled, performed, supervised, and unsupervised.

Programmed - The machine is built and programmed to perform a specific function with just a few physical controls. The least flexible technique. Control software written to do something or perform a specific purpose that rarely changes.

Interpreted: This is probably the most common technique in 2020. It does not mean writing system program code. It means generating Gerber file data (G code) to run robotic / automation machines. G-code files are often ASCII (text files) that are read by an interpreter to produce specific mechanical / electronic control actions. The interpreter reads a line of G code, parses it, and then executes the code to perform the requested physical movements. The G code is produced as an output file from a computer-aided design (CAD) program, although simple files can be generated by hand. Most CNC and 3D printers work with Gcode files, for example.

Supervised learning: teaching a machine how to perform some activity by physically moving it through the task. When placed in "learn" mode, the robot reads the same closed-loop control (feedback) sensors that it uses to operate. This way you can learn how a function should be performed. This is part of the entire learning process complemented by native (programmed), human and interpreted operation.

Unsupervised learning: I have not used or programmed this learning method. My opinion is that this type of teaching will become the dominant way of doing things in the future. Here, scheduled, interpreted, and supervised learning will be part of the unsupervised process. The difference is that when the robot is in the field, it will be able to analyze and adapt to situations with little or no human interaction. Think of the robotic vehicles that are sent to our Moon and Mars, for example.

Well, that depends.

There are two ways:

A: Programming

B: machine learning

Let's talk programming first. Programming is when you write code on the robot's hardware and allow it to complete tasks. You program it to do what you want. This is the most common way to teach a robot its task. This form of teaching is extremely efficient in performing specific and non-variable tasks.

Machine learning is something very new and complicated. Machine learning allows a machine to "learn" through observation. Many chat bots use this feature. Because there are so many words and phrases, coding it is ha

Keep reading

Well, that depends.

There are two ways:

A: Programming

B: machine learning

Let's talk programming first. Programming is when you write code on the robot's hardware and allow it to complete tasks. You program it to do what you want. This is the most common way to teach a robot its task. This form of teaching is extremely efficient in performing specific and non-variable tasks.

Machine learning is something very new and complicated. Machine learning allows a machine to "learn" through observation. Many chat bots use this feature. Because there are so many words and phrases, coding it is difficult. Machine learning enables the bot to browse the web and learn words over the internet. This is very silly for doing specific tasks that have no variability, but it is really good for tasks that are massive and require expertise.

In the future, robots can be so good at learning that they can be considered "smart." However, the Internet is quite toxic and there are a lot of bad words. This problem is getting worse every day and therefore in the future chat bots might start saying inappropriate things.

Hope you enjoy my answer:

-Howard Xing

Other Guides:


GET SPECIAL OFFER FROM OUR PARTNER.