A METHOD FOR DYNAMIC BALANCE
Dynamic balance is a very important function of the human biological control system. Humans have learned to walk on uneven terrain, keeping the body posture in the vertical position by sensing the direction of gravity. The adaptation of this skill to operations of a truly unstable autonomous mobile robot would be useful in dealing with an unstructured environment. This thesis describes the dynamic balance of a single link inverted pendulum in flat and changing terrains. A gravity sensing technique has been developed and tested in simulation to measure the imbalance of the inverted pendulum from the vertical position regardless of the terrain conditions. A task oriented motion planning algorithm has been developed and tested. The planned motion allows the inverted pendulum system to perform the desired motion with an unbalance-balance cyclic motion in flat and changing terrains, similar to the way humans plan their motions. The simulation results on the developed techniques, gravity sensing and motion planning, have indicated that the proposed methods can be used to balance an unstable mobile system in motion and at rest in changing terrains. A self guided unstable autonomous mobile robot could be built by incorporating the techniques with a modest control system.
Master of Science (M.Sc.)
Electrical and Computer Engineering