Problem with Humanoid Robots

According to me a Human Robot is an automated system which is supposed to automate various behaviors offered by a Human body as a system.

In my training programs, many times my participants ask me why I am discussing an unrelated topic like Human Robots in correlation with Information Systems.

It is not an unrelated concept as there is a direct correlation between a Human Robot and a Software/Information system.

Go through this section carefully as it will help you understand a very important aspect that will change the way you look at software systems.

Let me give you a task of designing a Human Robot. Since the system is a Human Robot; needless to say it is supposed to walk like a Human Being.

So how would you do something like this? How would you make a Human Robot exhibit a behavior like walking?

We walk all the time. Do you think it's easy to make a Robot exhibit the behavior of walking?

Do you think Human Robot is a simple system to develop?

According to me a Human Robot is one of the very complex systems to develop. The best engineering companies across the world have been investing huge money and resources in developing Human Robots. In spite of all the financial and technological might, the companies have been struggling to design a robot which can behave as naturally as a human being or a robot that can walk as naturally as a human being.

If it had been a simple system to design, let me tell you there would be millions of human robots moving on this earth.

Can you think what is so complex about Human robot as a system and walking as a behavior?

According to me one of the most important and complex aspect of walking as a behavior is "Balancing" or "maintaining the centre of gravity".

With so many systems like cars, aircrafts, submarines that are designed by human beings can successfully maintain their centre of gravity, what is so difficult in doing it for a Human Robot?

We have already studied a little about Human body in the previous section. Can you think of something that can help us understand this complexity?

What are the prerequisites for an automated system to behave as naturally as a real world system?

For any automated system to behave as naturally as the real world system the prerequisites are:-

  • The structural and behavioral complexity of the real world system should be completely understood.
  • You should be able to mimic or represent similar or better structural components in the automated system.
  • Once you have similar or better structural components in the automated system, you need to make these components interact in the same or better way as they interact in the real world system.


The first prerequisite of understanding the structural and behavioral complexity of a human body is satisfied as human anatomy as a subject is well studied and documented in medical science.

Since we understand human body as a system, we know that it is made up of millions of structural components nested at different levels of granularity.

As per the second prerequisite the automated system should have similar or better structural components as compared to the human body.

Is this possible?

The challenge is to arrange millions of structural components nested at different levels of granularity in a structure which is about 6 feet tall, 3 feet wide and does not weigh more than 100 kilograms.

According to me, in spite of the technical know how it is not possible to arrange millions of structural components nested at different levels of granularity in a structure that meets the above statistics.

Although it is possible to have similar structural components on the automated side, it is extremely difficult to restrict the size. It can be done but the structure would go up to 100 feet tall and weigh more than 1000 kilograms.

Let me share my personal experience with you. I have been working on designing a "Humanoid Cricket Bowling Machine" which has a natural arm that throws the ball.

At the prototype stage I had structurally tried to mimic the Human hand with most of the structural components of a real hand.

Though it was a very difficult thing to do, it was a rich learning experience as I was able to simulate certain very important behaviors of the hand. But the biggest fault with my mechanical hand was that it weighed more than 30 kilograms with double the dimensions to support different levels of flexibility and strength needed.

That day I realized that if we look at our hands as a system, which doesn't weigh more than 2 kilograms is a marvelous piece of engineering.

According to me currently we don't have the technical know how to structurally mimic a human body and that is the biggest limitations. But with the work going on in the area of Nano- technology, it might be possible for us the see a human robot with all the structural parts.

The third prerequisite says that once you have similar or better structural components on the automated side, you need to make them interact with each other in the same way.

We know that in a behavior like walking most of the structural components at nested levels of granularity interact with each other. The next big challenge is to synchronize the interaction between these millions of components to deliver the behavior.

We have now understood human robots along with the problems associated with them.


 
Hemant Jha
Founder - VPlanSolutions
Researcher, Trainer

www.VPlanSolutions.co.in