This dissertation proposes a new type of modular, long-reach, truss-type manipulator. Variable Geometry Trusses (VGT’s) are used to construct a reconfigurable manipulator system in which all primary members are loaded in pure tension or compression. Each module of the manipulator system is either a static truss link or one of several possible VGT actuators. This results in an extremely stiff and strong manipulator system with minimal overall weight. While many potential applications exist for this technology, the present work was largely motivated by the need for a robotic waste remediation system for underground radioactive waste storage tanks. This new manipulator system provides several advantages when used for this application. The reconfigurable nature of the proposed system allows the manipulator to be adapted on site to unforeseen conditions. Additionally, the kinematic redundancy of the manipulator ensures that solutions can be accomplished even in a highly obstructed workspace. The parallel structure of the truss modules enables the manipulator to be withdrawn in the event of a structural failure. Finally, of particular importance to this task, the open framework of the modules provide a passageway for waste conveyance or additionally, could act as a shielded conduit for control and power cabling. Kinematic analysis algorithms tailored to address the peculiarities of this new manipulator system have also been developed. In this work, the kinematic redundancy of the system is exploited to provide alternative solutions, to avoid numerical difficulties at singularities, or to avoid workspace obstacles. These issues are addressed through a combination of null space optimization procedures and order reduction methods. The null space optimization procedures are accomplished by extracting information from a full singular value decomposition of the Jacobian matrix. This method is shown to converge quickly, even for systems with thirty or more degrees of freedom. This represents a significant increase over most of the current literature which typically addresses systems of eight or fewer degrees of freedom. This dissertation presents the first application of null space optimization techniques for the positional control of a high degree-of-freedom parallel manipulators. This work also formalizes the concept of a canonical input specification set. The application of this concept results in greatly simplified analyses of many parallel manipulators. Although the manipulator system discussed was specifically developed for robotic handling of radioactive waste, the final resulting methodology is suited to a much broader class of problems, namely, under-constrained, redundant manipulator systems in general. / Ph. D.