So I thought I would start to begin to explore my robotics framework. One of the prevalent ways of thinking on robotics is that actions are in fact made up of a set of behaviours that all work together collectively. This is most easily seen with the subsumption architecture where you have low-level behaviours such as "don't walk into walls" and "wander around aimlessly", combined with a high-level behaviour such as "stop when you get to point X". Together these behaviours are sufficient to achieve the objective of navigating to a particular point, without having to worry about all the details of path planning. Can we apply this same principle to Tai Chi? I thought I'd give it a go.
So here's a starter for 10 on some really simple behaviour ideas that should hopefully combine together to give you a tai chi capable robot.
1) In order to achieve a movement, move your joints/limbs sequentially, starting with the ones closest to your centre of mass. In tai chi speak: "Everything comes from the Dan-tien". My supposition for this is that the dan-tien and is actually the centre of mass of the human body.
2) Given a particular configuration of your points of attachment (i.e. feet), arrange your joints/limbs, so that your centre of mass, is located at the centre of all possible stable centre of mass positions. In other words this means that your centre of mass (dan-tien), has the maximum possible movement potential and is therefore is in the most flexible and stable position.
3) No joints move in isolation. All joints move simultaneously in every movement. In tai chi speak this is whole-body coordination. Practically this means recalculating all joint positions dynamicaly for every movement, but if me do this in a distributed, rather than centralised manner it should be posible to achieve.
Anyway that's just some thoughts that I've jotted down, more to follow...
These are great!!
ReplyDelete