The M3 Robotics Lab at Washington State University is focused on high degree-of-freedom, compliant systems where through the use of smart materials and construction geometry, the compliance is both tunable and modular. The motivation for this research comes from the medical world in which many medical devices rely on their high intrinsic compliance to achieve bending and turning necessary to reach a target location. However, this same compliance can prohibit the device from exerting large forces at the target location. Thus, the M3 Robotics Lab seeks to develop the basic science behind tunably compliant devices through feedback control and fundamental changes in device mechanics.
Steerable needles could dramatically improve the medical field in areas such as minimally invasive surgery and tissue sampling. The goal of this research is to develop an entire needle system including both an insertion device and control device using a dexterous, flexible continuum robot consisting of concentric, elastic tube and needle.
The insertion component of this device consists of a pre-shaped nitinol needle and tube. Both have one translational degree of freedom controlled by stepper motors, but the nitinol wire has an additional rotational degree of freedom. By manipulating this rotation in sync with translational movements, the insertion direction can be adjusted. Changing the form of pre-shaped needle tip, step length of following motion, wire radius, and thickness of nitinol tube gives the user the ability to change the needle curvature inside tissue. The device is currently controlled by manual adjustments of code, but the long run goal is to integrate a joystick to manipulate the needle to avoid obstacles and reach a position inside tissue based on three dimensional vision.
The goal of this research is to make compliance a property that can be varied by precise tuning at any moment. Ideally, a device with this ability could be compliant enough to move in a complex manner at one point, and subsequently become rigid (or less compliant) so as to support some large amount of force at the target location. Currently, this concept is being explored in two ways in the M3 Robotics Lab: tendon driven actuators with variable stiffness joints and modular spring mechanisms.
Tendon Driven Actuators
The goal of this vein of research is to bridge the gap between current approaches in soft robotics and the direct actuation of rigid links in traditional robotics by investigating the geometry of compliance. Currently, this is begin accomplished by identifying methods of combining the compact nature of smart materials with existing robotic actuation techniques.
The most recent model fabricated for this purpose is a finger-like form that includes two solid bars of
a low melting point metal called “Field’s Metal” embedded in an elastomer. Power resistors are also embedded in the middle of the silicon shape and connected by wires to a power supply. When power is applied, the resistors dissipate heat into the surrounding silicone which locally raises the temperature of the metal just above its melting point. Once a particular section of the metal bars reaches its molten state, the entire elastomer finger can be bent. The actuating force is applied by tendons lying just within the top layer of the elastomer. By varying which resistor is activated, the location of the joint within the finger can be varied, thus changing the geometry of its compliance.
In the picture progression here, the heating resistors at two points are activated, causing bending much like a human finger:
Modular Spring Mechanisms
This concept is based on the various mechanical properties of springs (e.g. stiffness, buckling, potential energy) and seeks to utilize them to create a linkage mechanism with two stable states and two distinct stiffnesses. Starting with four rigid links in the form of a parallelogram, two springs are mounted between opposite corners so that deforming the shape in either direction results in compression or tension within the springs. Using MATLAB to simulate the forces and energy within the springs, it is possible to predict how the mechanism would respond to an external input. By allowing one spring to buckle (causing its force to drop) while the other remains loaded, the mechanism could theoretically have two stable states, and by varying the individual properties of the springs used, these states could determine the stiffness of the mechanism in each of the stable configurations. For a flexible, compliant application, the mechanism would have a low stiffness corresponding to one spring while for a large force application, the mechanism would have a high stiffness corresponding to the other spring (i.e. the compliance is “tuned” to the right amount for either situation).
Eventually, the goal is to move away from metal springs toward elastic materials and composites. The modularity of these small, tunably-compliant systems will make it possible to integrate them together for bigger and tougher applications in the future.