Interest in robots on legs is far and wide, with many big projects turning heads. Such as the BigDog, from Boston Dynamics, which was a very famous attempt to create a bot that could carry items across rough hills and trails for troops. Human-like, bipedal (two legs) bots such as Honda’s ASIMO project explore the miracle of how we (and many other animals) can walk. Bots are in our video games, in our homes cleaning our floors and in our phones to answer our questions.
Legged locomotion is a simple idea at first. It’s easy to see one foot in front of the other, then repeat. It is how most of us humans get around, after all. The problem is, we only see a very small part of what is actually happening when we – or any machine – moves on legs.
The human foot isn’t just a stand for our bodies. It’s a complex system of 26 bones, 33 joints and 19 muscles. When we walk, our foot makes many small movements that we aren’t even aware of that keep us standing.
The sheer complexity of the human foot is not exclusive to us. When we create walking robots we have to account for the same level of intricacy. Projects such as this failed walker show us that robots with many more legs than us still need a more complex system than “foot goes in front of foot”. In this case, a model on a computer is needed. It has to guess and correct where the legs are at a given time during the process of walking.
In the case of the famous BigDog, about 50 sensors are used to perform this. Some of them track the elevation and speed of the whole machine. Others track the servo motors – their position and how much force with which they are moving. The computer in the BigDog does the same “guess-and-correct” method I talk about above. It uses the same complex system – on a greater scale using more complex systems!
In short, moving on legs is far more complex than most of us realize. To move on legs requires systems to track and manage the tiniest movements. It also requires a vast and complex group of systems to track and make these motions accordingly.