CONTROL ALGORITHM FOR REDUNDANT DEGREE-OF-FREEDOM MANIPULATORS

CURRENT CONTROL ALGORITHMS FOR REDUNDANT DEGREE-OF-FREEDOM (DOF) MANIPULATORS ARE OFTEN LIMITED BY THEIR ABLITY TO RESOLVE SINGULARITIES (WHEN TWO OR MORE JOINT ROTATION AXES BECOME ALIGNED) AS WELL AS THE INVERSE KINEMATICS ASSOCIATEDWITH SIX DOF AND HIGHER MANIPULATORS (THE RELATIONSHIP BETWEEN THE SIX JOINT ANGLES CORRESPONDING TO THE SIX DEGREES OF FREEDOM). WE WILL DEVELOP A CONTROL ALGORITHM FOR A REDUNDANT DOF MANIPULATOR TO SOLVE MANIPULATOR INVERSEKINEMATICS WHILE AVOIDING SINGULARITIES. THE ALGORITHM WILL APPLY TO A SEVEN DOF MANIPULATOR, WHERE, BECAUSE OF THE INFINITE NUMBER OF SOLUTIONS TO THE INVERSE KINEMATICS PROBLEM, THE ALGORITHM MUST SELECT A SINGLE SOLUTION BY APPLYING A CONSTRAINT. A SINGULARITY AVOIDANCE CONSTRAINT WILL BE APPLIED AT THE JOINT MOTION LEVEL AND WILL DERIVE FROM THE GEOMETRY OF THE MANIPULATOR. THE RESULT IS AN ALGORITHM WHICH WILL ENABLE A MACHINE TO ACCOMPLISH TASKS FAR MORE SOPHISTICATED THAN THAT OFFERED BYCURRENT NON-REDUNDANT MANIPULATORS, AND WHICH WILL BRING HIGH PERFORMANCE ROBOTIC MANIPULATORS CLOSER TO COMMERCIAL VIABILITY. WITH A FOLLOW-ON PHASE II EFFORT, WE WILL ACTUALLY CONTROL AN ARM USING THE ALGORITHM.