A four-legged robot?
Too many things to consider.
For stable locomotion each leg would need great responsiveness…
Deciding how each of the four legs should move in every possible situation felt like far too much work.
And there seemed to be far too much to account for across different terrain.
Then again, if attitude control worked well it would practically fly…
Setting all that aside, the motors just get way too expensive
So then?
I figured a two-legged robot would be much the same. If anything, there was far more to think about regarding balance than with four legs:
- When a leg slips
- When it jumps and assumes a landing posture, but some variable makes a stable landing impossible
- When it’s just standing there…
- etc.
Nothing but hard problems.
One leg..?
Oh…? The number of things to think about dropped. Dramatically.
With only one leg there aren’t many ways to move at all. The first thing that came to mind was hopping. Huh? That’s perfect.
One leg it is!
While designing a one-legged robot, the parts that got dramatically easier were the leg and the body. If it moves by hopping, the leg needs no complicated hydraulics or joints — a spring and a linear bearing are enough to solve the robot’s propulsion outright. The same goes for attitude control. Instead of choreographing legs in perfect sync, you just spin the body directly, right?
Attitude control
Attitude is controlled in two main ways:
- Extending the leg to reduce angular velocity
- Rotating the body with a reaction wheel
If you set the body to the angle you want and then jump, you can travel in the intended direction; and on landing, the spring recharges while absorbing the impact, making a stable landing possible.
