Jacob Hicks
A great deal of work has been done to solve the problem of robotic motion planning. There are a large variety of packages and techniques that have been developed to deal with different types of robots operating in a variety of different situations. Very little work however, has been done on the theory and algorithms required to plan robot motion through fluid of arbitrary viscosity.
The goal of the project is to create a robot motion planner that is able to interface with a fluid simulator. The motion planner will be able to access the global state of the fluid environment to help the robot direct the forces that it will have to apply in order to navigate from a given start position to a goal position.
In this project several things must be achieved. An interface that provides information between a fluid simulator and a motion planner must be created that will allow the motion planner to be used on multiple different kinds of fluid simulators. The motion planner must be equipped with heuristics and algorithms that will allow it to make intelligent decisions for the directions in which forces must be applied in order to obtain proper movement.
The current state of the art does not deal with the problem explicitly at an algorithmic level. Most of the papers that have been published rely on sensor data from a real physical craft to plan their local motion. These robot construction papers gloss over how their planning algorithm works normally only referring to available packages that were used. They seem to be more interested in the performance of the craft than the actual robot underneath. Some work in the areas of nanomedicine takes a slightly similar approach to the simulator that I desire to create, but most of those make huge simplifying assumptions that make them less robust than they could be.
I hope to be able to make several novel contributions in this project. The current state of the art in underwater robotics is confined to a single known viscosity, that of water. Based on the design of my program it would inherently be able to handle liquids of different viscosities. This would have the advantage of being able to work with different characteristic lengths in the simulation, so the same program that would be able to work with a microrobot in the bloodstream, a submarine in the ocean, or a robot trying navigate tree sap.