ESTIMATE OF THE 2 SMALLEST SINGULAR-VALUES OF THE JACOBIAN MATRIX - APPLICATION TO DAMPED LEAST-SQUARES INVERSE KINEMATICS

The smallest singular value of the Jacobian matrix is known to be the sole reliable measure of closeness to singular configurations; its com putation via a full singular value decomposition, however, is demandin g in view of real-time applications. In this article, a standard singu lar value estimation procedure is extended to improve the handling of singular values crossing along a given trajectory. The proposed algori thm maintains the computational efficiency of the original technique a nd shows higher estimation accuracy with respect to the original proce dure. Case studies are developed to verify the effectiveness of the es timation procedure applied to an industrial manipulator. (C) 1993 John Wiley & Sons, Inc.