Task 1.3: Surgical Enhancement began with initial investigations into hidden Markov models as a means for recognizing surgical actions. Core funding, plus an NSF RHA grant, supported work on vision-guided virtual fixtures and related studies of their efficacy for improving the speed and accuracy of fine-scale motion tasks. An NSF ITR grant has provided support to develop an architecture that integrates task modeling, human-motion modeling using HMMs, and online human-machine task synchronization with lower-level motion augmentation primitives. This project is targeting retinal vein cannulation, retinal membrane peeling and micro-suturing. In recent work, we have developed a platform-independent architecture for specifying and executing surgical assistance tasks. A recent NSF SBIR grant with Invenios, Inc. is evaluating the commercial potential of human-machine collaborative systems for both manufacturing and surgical applications. In addition, we have now applied virtual fixtures to telemanipulation systems with various underlying control laws and demonstrated that such systems can be made stable (and therefore safe for application in the OR).