Manipulation approaches for hyper-redundant modular robots

Lee D Kress, Purdue University

Abstract

This work examines how to utilize a robot that consists of many modular units with revolute joints to accomplish a given task. The task may be to move the modules of the robot themselves to another position or to move an object from a specific position and orientation to another position and orientation. For example, a 100 modular unit robot has 100 different motors but there are only six degrees of freedom in three dimensional space, and an approach to find the best solution to a given task becomes difficult. The problem addressed in this thesis is to manipulate an object in two dimensions using a hyper-redundant modular chain-type robot where each module can potentially contact the object or dynamically guide the object. Using modular robots to complete such tasks is beneficial as modular devices have the potential to be very robust and complete a variety of different tasks. The approach is to specify the position and orientation of a few specific modules throughout the movement of the robot. Solving for the motor torques of the modules required to satisfy the position and orientation constraints results in many unknowns and relatively few equations. A pseudo-inverse is then used to determine the set of solutions which best minimizes joint motion of the revolute joints of the robot. This thesis specifically examines the problem of configuring the modules into a specific shape and manipulating an external object. Various results of configuring the modules into a specific design and manipulating an external object to a desired position and orientation are presented. The computer simulation is written in Matlab. The results of this research show that it is possible to have controlled manipulation of an external object by only using module contact of a modular robot. This thesis presents the theory, the program created to implement the theory, and the results of the research.

Degree

M.S.M.E.

Advisors

Cipra, Purdue University.

Subject Area

Mechanical engineering

COinS