Abstract:
In this thesis, the coordinated navigation of multiple disk-shaped robots is extended to three-dimensional workspace. The robots are assumed all to be spherical and residing within the same spherical workspace. In addition, each robot has a sensory range which determines its knowledge of the workspace. The robots task is to navigate from an arbitrary initial location to a final goal position. Two different scenarios are considered: independent navigation and order constrained navigation. In the independent case, all the robots are physically independent from each other. In order constrained navigation, we consider the case where the robots are ordered and the distance between adjacent robots is required to remain fixed throughout the motion by the help of imaginary links. It is shown that both types of motion can be achieved by using feedback-based strategies using artificial potential functions. The artificial potential functions is based on the extension and adaptation of a construction formerly proposed and proven for the two-dimensional case. Extensive simulations serve to demonstrate the performance of the approach under varying difficulty of tasks and noises. In addition, the theory developed in order constrained navigation is utilized to investigate protein folding kinetics. Results obtained from simulations are validated by the known experimental results.
