This paper presents an algorithm for collision free path planning of autonomous manipulators among obstacles. The algorithm uses powerful features of Non-Uniform Rational B-splines (NURBS) to determine the path of the manipulator in the complimentary space of the obstacles. Initially the algorithm artificially transforms the obstacles to point size. It then uses these point obstacles and start & target position of the manipulator as control points to determine a NURBS path which is the initial collision-free path. Obstacles are then iteratively increased in size by small increments. After each increment the new collision control points with obstacles are determined and a new collision-free NURBS path is generated to guide the manipulator through crowded obstacles. This procedure is continued till all the obstacles attain their full size and resulting path is maintained free of collision. The developed algorithm is most suited for manipulators facilitated with range sensors for remote detection of obstacles. The paper details the technique for obtaining a smooth collision-free path of a manipulator moving around a crowded environment with static obstacles.