-- MarkCutkosky - 21 Apr 2008

NSF / CCC / CRA Roadmapping for Robotics Workshop: A Research Roadmap for Medical and Healthcare Robotics

Co-Chairs: Allison Okamura, Johns Hopkins University, Maja Mataric', University of Southern California, Henrik Christensen, Georgia Tech

June 19-20, 2008, Arlington/Washington DC area

• See workshop proposal from A. Okamura - email "position paper" due April 23.

Version as submitted is attached at bottom of page.

Theme

Concurrent advances in robotics, haptics, biology and medicine are poised to enable a revolution in healthcare, in which people interact directly with robotic technology to improve their quality of life.

In the following sections we explore two exemplary manifestations of this confluence of technology with applications in human motion analysis and training and in human-interactive robotics for home healthcare. The first example involves the adaptation of fast algorithms developed for computing and controlling the dynamics of robots to the analysis and improvement of human motion. The second involves adaptation of principles from biology to the development of robots that are inherently safe for physical interaction with humans. Both projects are the subjects of new multi-disciplinary collaborations at Stanford to address growing societal needs.

Robotic Analysis Applied to Human Movement

Faculty contributors: Oussama Khatib (CS), Mark Cutkosky (ME), Scott Delp (BioE), Thor Besier (Orthopedics), Terry Sanger (Neurology)

Robots have proven to be an effective rehabilitation tool for persons with sensori-motor deficits, however the concept of adapting robotic sensing, control and display technology to wearable devices for improving or restoring human performance has received less attention. Figure 1 illustrates an approach in which motion tracking, dynamic modeling and control, and display are applied to motion training.

INSERT FIGURE 1 FROM KAUST

Advances in biomechanics, dynamic reconstruction algorithms and wearable devices will soon make it possible to provide real-time feedback of biomechanical information to patients undergoing rehabilitation. The sequence in Fig. 1 begins with dynamic reconstruction of human motion. Building upon algorithms and models developed in the context of humanoid robotics (Khatib et al., 2004) we have developed and implemented a novel muscle effort criterion for predicting physiologically accurate upper limb motion (Khatib et al., 2004, De Sapio et al., 2006). More recently, we proposed and implemented a new algorithm to reconstruct human motion from motion capture data through direct control of captured marker trajectories (Khatib et al., 2008). The approach directly projects marker points onto a simulated human model and tracks the trajectory in marker space without costly inverse kinematics computations (Demircan et al., 2008). The study has also shown the algorithm to easily accommodate anthropometric scaling of the musculoskeletal model to human subjects.

The next stage in Fig. 1 is dynamic simulation of muscle activity. Simulations of movement allow one to study neuromuscular coordination, analyze athletic performance, and estimate internal loading of the musculoskeletal system. Simulations can also be used to identify the sources of pathological movement and establish a scientific basis for treatment planning (Delp et al., 2007).

The third stage in Fig. 1 involves interactive display of results to the subject. Wearable haptic feedback has been proposed as a way to augment or even replace visual or audio feedback for motion training (Lieberman and Breazeal, 2007). Vibration is the most common feedback mechanism, but is imprecise at providing a sense of the magnitudes or velocities of motions. Nonetheless, vibration has shown promise in improving balance (e.g., Bach-y-Rita et al., 2005) and grasp force control in patients recovering from stroke and other pathologies that affect the neuromuscular system (e.g. Jiang 2008) . Wearable vibration devices are also being as a means to help patients with dystonia to isolate the effects of particular muscle activiations (Sanger NEED REFERENCE ).

For applications involving motion training and rehabilitation, skin stretch is a useful complement to the cues provided by vibration. Skin stretch is a component of the human proprioceptive apparatus (Edin, 2001,Collins et al., 2005) and artificially stimulated skin stretch near the joints can provide subjects with useful information about joint movement and velocity (Bark et al., 2008). As with vibration, the inherently low power requirements of skin stretch lend themselves to the design of compact wearable displays.

In addition to providing subjects with information about the motion of their limbs (which can also be conveyed visually) we expect that conveying real-time information about muscle force will be more intuitive using tactile displays. The ability to convey multimodal information is also important. For example, for neurological rehabilitation after surgery or stroke, we expect that providing proportional proprioceptive feedback will be most useful as subjects relearn movements. In athletic applications, where movements are often fast and highly stereotyped, we expect that providing discrete event cues may be more useful (e.g. alerting the subject that a particular muscle group has been activated too early or late in a golf swing or if the force in a muscle becomes high such that injury is likely).

Bio-Inspired Human-Safe Robotics

Faculty contributors: Oussama Khatib (CS), Mark Cutkosky (ME), Ken Salisbury (CS/Surgery). Take text from GM proposal, with home healthcare flavor.

The growing proportion of elderly individuals in industrialized societies will soon constitute a crisis unless cost-effective ways can be found to help them maintain dignity and quality of life, as they become increasingly dependent on daily assistance and healthcare services. Today, home healthcare is either very expensive or very demanding of families and friends. Robotic technology has the capacity to augment, and in some cases replace, the services that people provide.

To realize this goal, robotic platforms will be required to navigate the unpredictable and dynamic environment found in homes with greatly improved sensing, planning and reactivity than seen in today's robots. In addition, robot platforms must be developed that are intrinsically safe when working in direct physical contact with humans. Inadvertent collisions, although mitigated with sensing technology and intelligence, are inevitable.

Consequently, there is a need for inherently safe designs that have sufficient power and precision to carry heavy loads while also exhibiting low inertia and soft exterior surfaces. During the past several years research groups have investigated new actuation techniques to overcome the limitations of existing approaches. Zinn et al. (Zinn, 2004) have developed the distributed macro–mini (D2M? ) actuation approach to address the problem of a large reflected inertia by partitioning torque generation into low- and high-frequency domains, which are controlled by distributed pairs of McKibben (macro) and electromagnetic (mini) actuators. Pneumatic McKibben actuators provide high power and force density and inherently low mechanical impedance. However, the underlying nonlinear compressible gas dynamics involved make precise control difficult. By combining them with small electromagnetic actuators we were able to achieve a 10-fold reduction in effective inertia while maintaining high-frequency torque capability. The combination of two different actuation technologies comes at the expense of complexity in comparison to traditional robot design. To make this complexity manageable, we use miniaturized integrated pressure controllers and multi-material structures.

In addition to being low-inertia and low-impedance (for safety), these robots will rely upon having more tactile sensing than traditional systems. Then eed to be more responsive to a dynamic environment is driving the development of new technologies for sensor arrays with many sensors and their attendant communications and signal processing.

Further integration calls for use of Shape Depostion Manufacturing (SDM) techniques. SDM allows multiple materials, as well as sensors, actuators and other discrete parts, to be integrated in a single heterogeneous structure. The technology has been demonstrated for various bio-inspired robots in Cutkosky’s lab. The ability of SDM to provide local variations in materials properties also permits structures with high specific strength and stiffness in selected areas while providing high impact energy absorption in other areas.

Salisbury et al. has developed numerous bio-inspired robotic designs that mimic the dexterity of humans, enabling them to interact with the human- centric environment with minimal modification.

Older summary text. See if anything below needs to be merged into what is above...

Bio-inspired design + advances in robotic control and actuation are leading to robots inherently safe for direct interaction with people. This has important consequences for the goal of introducing robots into healthcare, at institutions and in the home. in addition to being low-inertia and low-impedance (for safety), these robots will rely upon having more tactile sensing than traditional robots. they will need to be more responsive. This requirement, in turn, is leading to the development of new technologies for sensor arrays with many sensors and their attendant communications and signal processing. Efforts underway at Stanford include the macro-mini actuation schemes underlying S2rho by Khatib and new dynamcially counterbalanced robots from Salisbury's lab + new tactile sensor arrays and communications mechanisms from Cutkosky's lab. New combinations of materials are also being exploited for a combination of high specific strength and energy dissipation when unexpected collisions occur. %ENDCOMMENT%

References

Bach y Rita, p., Danilov, y., Tyler, M., Grimm, R. Late human brain plasticity: vestibular substitution with a tongue brainport human-machine interface. Journal of Plasticity and Restorative Neurolgy, 4:31–34, 2005.

Bark, K., Wheeler, J., Premakumar, S., Cutkosky, M., Comparison of skin stretch and vibrotactile stimulation for feedback of proprioceptive information. IEEE/VR Symposium on Haptic Interfaces, March, 2008.

Collins, D.F., Refshauge, K.M., Todd, G., Gandevia. S.C. Cutaneous receptors contribute to kinesthesia at the index finger, elbow, and knee. Journal of Neurophysiology, 94:1699–1706, 2005.

Delp, S.L., et al. OpenSim? : Open-source software to create and analyze dynamic simulations of movement. IEEE Transactions on Biomedical Engineering, vol 55, pp 1940-1950, 2007.

Demircan, E, Sentis, L, De Sapio, V, Khatib, O, Human Motion Reconstruction by Direct Control of Marker Trajectories. Advances in Robot Kinematics, June 2008 (in press).

De Sapio, V., Warren, J., Khatib, O., Predicting Reaching Postures Using a Kinematically Constrained Shoulder Model. J. Lenarcic and B. Roth, (Eds.), Advances in Robot Kinematics, pp. 209-218. Springer, 2006.

Edin. B.B. Cutaneous afferents provide information about knee joint movements in humans. The Journal of Physiology, (531.1):289–297, 2001.

Khatib, O., Sentis, L., and Park, J., A Unified Framework for Whole-Body Humanoid Robot Control With Multiple Constraints and Contacts, Springer Tracts in Advanced Robotics - STAR, European Robotics Symposium, 2008.

Khatib, O. Sentis, L., Park, J. and Warren, J. Whole Body Dynamic Behavior and Control of Human-Like Robots, International Journal of Humanoid Robotics, 1(1): 29-44, 2004.

Khatib, O., Warren, J., De Sapio, V., Sentis, L. Human-Like Motion From Physiologically-Based Potential Energies. On Advances in Robot Kinematics, J. Lenarcic and C. Galletti (Eds.), pp. 149-163, Kluwer P.C., 2004.

Khatib, O., Brock, O., Chang, K., Conti, F., Ruspini, D., Sentis, L. Robotics and Interactive Simulation. Communications of the ACM, vol. 45, no. 3, pp. 46-51, March 2002.

Lieberman, J., Breazeal, C., TIKL: Development of a Wearable Vibrotactile Feedback Suit for Improved Human Motor Learning. IEEE Transactions on Robotics, vol 23, no. 5, pp 919-926, 2007.

Shin, D., Sardellitti, I., and Khatib, O. A Hybrid Actuation Approach for Human-Friendly Robot Design. {\em Proc. of the 2008 IEEE International Conference on Robotics and Automation,} Pasadena, CA, 2008.

Zinn, M., Khatib, O., Roth, B., and Salisbury, K. ``A New Actuation Approach for Human-friendly Robot Design,'' {\em International Journal of Robotics Research,} 23(4/5) 379-398, 2004.

• Khatib_GMv3c.doc: MS Word version of GM human-safe