The first task in Thrust 1 is Task 1.1: Surgical Systems and Applications, and focuses on systems integration and testbeds. Our initial work was devoted to providing improvements to physical steadiness and accuracy, leading to the development of the Steady-Hand robot and the Micron family of hand-held active tremor canceling tools. Reductions in tremor together with studies of associated effects on manipulation speed and accuracy were documented, force-scaling was implemented and tested, and, in Year 3, feasibility of retinal vein cannulation was demonstrated. Recent work in Task 1.1 has focused on the development of simpler, less expensive steady-hand systems. Also, we have aggressively pursued new collaborations in sinus surgery and throat surgery. The latter has led to the development of a novel family of flexible devices for remote dexterity in constrained spaces; informally labeled snake robots. Strategically, we have formed an alliance with Intuitive Surgical to provide a direct industrial partner for this and future work. This alliance has resulted in the awarding of an STTR grant, an NIH R21, an NIH R01, a new telemanipulation testbed, as well as several other grant submissions and collaborative projects spanning many tasks. We are planning on submitting an NIH grant (Hager, PI) to take the retinal project to the next level of development and testing on animal models. Likewise, NIH support for our prior work on inner-ear procedures is being pursued by Whitcomb.

Testbed descriptions: Currently, Thrust 1 has four established testbeds: the Steady Hand Robot (SHR), the Micron work at CMU, a Virtual Environment with Haptics (primarily a simulation and biomechanics testbed focusing on vascular surgery), and a Telemanipulation testbed (for minimally invasive surgery applications in the throat, sinus, and abdomen). Each of the testbeds has been chosen to emphasize particular integrative goals for the thrust.

The SHR testbed consists of the JHU Steady Hand robot, control software for cooperative guidance, image processing and visualization tools for modeling and displaying information about the surgical field, and a high-level execution architecture for multi-step cooperative tasks. This testbed is being applied to various micro-surgical applications, where high levels of precision and repeatability are essential. It drives research in many areas, including mechanism design, control system design and analysis, image processing and registration, visual tracking, human-machine systems, and human-machine system validation. Applications are being prepared for NIH funding of pre-clinical, animal studies for eye and ear procedures. In addition, we are partners on grants with Invenios, Infinite Biomedical Technologies, and with Intuitive Surgical, Inc. on several successful grants (from NIH and NIST).

The Micron system at CMU is a complementary testbed for microsurgical interventions. Rather than using a grounded robotic system, Micron seeks to perform active tremor cancellation in a mobile system. The sensing portions of Micron have been approved for clinical use under an IRB approval and have received NIH funding for pre-clinical use. This will provide a source of in-vivo data that non-clinical systems can be tested against. Many of the elements of the Steady-Hand system (multi-step task models, image-based guidance, and advanced visualization) can be mapped onto the Micron system, providing an alternative clinical and industrial path for these technologies.

The Telemanipulation testbed was initially designed to test some of the basic ideas of human-machine cooperative control in a minimally invasive surgical setting. This testbed is now comprises a complete bi-manual manipulation system. It provides a strong link to an established industrial partner (Intuitive Surgical, Inc.), and a means to explore interventions that are not consistent with the Steady-Hand paradigm. For example, our recently developed snake robot might possibly be operated in a telemanipulated fashion. This architecture provides another point of comparison against which human-machine systems using the Steady Hand or Micron paradigm can be evaluated.

The work in Task 1.1 has led to the development of a rich theory of virtual fixtures at a basic control level, the development of a high-level task description and execution architecture at the systems level, and comparative studies of how (and how effectively) these ideas can be expressed and applied in many applications contexts. To quantify the effectiveness of architectures, we are now developing expertise in evaluating human-machine systems. The maturation of our testbeds has led to a focused set of questions related to more effective modeling of tasks (for example, though intra-operatively acquired data), the extension of the notion of guidance to more complex manipulation (e.g. bi-manual) tasks, and ultimately to the question of developing context-sensitive task descriptions that are aware of the means and ends of the surgical procedures, as well as how they relate to the underlying anatomy.

These integrated testbeds enable us to work effectively on component technologies to attack systems level problems. For example, in order to develop context-aware surgery, it is necessary to dynamically and geometrically model the surgical field. However, this can be done effectively only if there is, knowledge about what should be modeled (requiring a task model), and lower-level information about the dynamics of the interventional system (e.g., the motion of the viewing system, the motion of the tools, etc.). We can now investigate ways to tune both the high-level task description and the low-level control to make the modeling process more efficient and robust.