[Kevin Sevcik]

A preliminary Inventor model suggests that a 150 lb robot with a CoM 18 inches off the center of the pipe will load a single pipe to 50% of yield.

This is using the cheesy "Stress Analysis" Inventor package we get with the kit, so it's somewhat less trustworthy for the coarse mesh it's limited to. But it's also assuming the pipe is a cantilevered column with support only at the bottom, which is clearly not the case. The rigid connections to the other side of the tower will mean some load will be carried through the top of the pipe as well. So I think one robot per vertical pipe is a reasonable load for the tower.

However.... I was assuming a 12 inch separation between the "hands". The hands generate two equal and opposite forces to create a moment to balance the moment generated by the robot's CoM. The closer the hands, the greater the point force required to generate this moment. At around 6" of separation, this point load starts to become significant. At this point you have to worry about crushing the pipe between your hands in addition to bending the pipe at the base. This crushing type failure will NOT be improved by rigid connections at the top and middle of the platform. The only way to improve it is to increase the separation of your hands.

Something for to think about for teams considering this approach: This normal loading on your hands increases linearly with the weight of your robot(s). And you can customize magnitude of this normal force by choosing the separation between your hands. I wonder how this could be useful.....