Robotic Figures: Future Directions
Much of the figure engineering would be carried over from the existing robots. The upper body would likely be the same. But the back-entry cable group and its up/down revector block is new, and the hip area, being narrow and short, presents serious spatial challenges. The main problem is that "twist cylinders" (the tightly packed parallel groupings of cables which twist) should be as long as possible to reduce cable wear, and we don't have enough verticle space for the new twist cylinders. The most obvious way to solve this hip area space shortage is (sketched at right) to divert each leg's main group up and then down over ball-bearing revector pulleys (similar to those used throughout serpentine arm). This can buy us a precious 4" of verticle which will be needed for the lower torso twist cylinder as well as the upper leg twist cylinders, and it allows the cone to come in lower in the back which bean the upper torso twist cylinder can be longer.
The legs must be lightly built since they are longer than the arms and we have limited cable force available. The foot will be a very light non-articulated casting riding on a 2 d.o.f. ankle.
Unlike the Slave Zero / One robots which have heavy epoxy-clay sculpted panels, the body panels here must all be very light, since the whole body must be moved. This is especially important for the leg panels which, unlike the arms, are required to visually "bulk up" the lightly built leg structures. Thin shells of fiber reinforced resin would be cast off of sculpted masters.
It has been an obvious goal from the beginning of the Slave Zero robot project to create a FULL BODY version; and better still, a full body which can TUMBLE continuously in space.
Such a "floating" robot performer would be unprecedented in its expressiveness, going far beyond the current fixed-base machines. In addition to moving its body, its body moves in ways only divers and astronauts can approach.
There are major technical hurdles on the way to such a machine, but the Slave robots and the Serpentine Arm project have blazed much of the trail. Such a project would be costly in money and skilled man-hours, and probably will require institutional support.
The desgn philosopy of Slave Zero would be expanded in the proposed machine: No actuators on the figure, all joints tendon driven. Not only does this keep the figure clean and "structural", it gives a subtle, almost self-motivated look to the robots.
Whereas Slave Zero has its cone of tendon cables below the upper body, this machine would have legs, leaving the back as the only option for tendon placement. This arrangement creates additional 90 degree revectors as the groups enter the body, difficult in terms of verticle space. Another difficulty caused by such a back cone is that the arms must not go there - somewhat problematic if the figure is to pivot on its verticle axis.
Lessons learned in the original Slave robots and the Serpentine Arm will be applied. There would likely PC mastered serial network, with microcontroller subsystems at each node. There would be manual, rythmic and programmed moves available.
Another control area which would be explored in this project is sound-driven automatic rhythm generation.
A critical design issue is safe and easy transportability. The complete object, or the parts it breaks down into, must fit through doorways and should be liftable without special equipment. If it is to break down into 2 or more components, the assembly and disassembly should be easy and safe.