In the 1960s movie "Fantastic Voyage," ship and crew of a submersible are shrunk down to the size of a red blood cell. This vision of Isaac Asimov's is not too far off. New technologies in intelligent systems and advanced manufacturing have resulted in preliminary assembly of human presence at scales and distances once dreamed of as fantastic. For those involved in microrobotic and microsystem fabrication, telemanipulation and telepresence, and 3D imaging, that future is drawing a few small steps closer.
Within the realm of miniaturized robots, or microrobotics, the issues of combining independent processes such as data processing, mechanical, electronic and optical functions, in addition to many others contingent upon the application, are being addressed all over the world. An exuberant Armin Sulzmann, Swiss Federal Institute of Technology (EPFL; Lausanne, Switzerland), enjoys talking about his domain for micromotor assembly, which utilizes mobile microrobots.
Sulzmann's goal has been to develop precise tracking systems for a variety of microrobots and microgrippers used in microassembly. In one instance, Sulzmann and his group create microrobots using telemanipulation to guide microgrippers in placing microrotors on micromotors. With his colleague, Silvia Allegro, they have managed to automate a high-speed microassembly system, and with Jean-Marc Breguet, to develop the precision of manipulation at the 5 to 10 nm resolution. This resolution defines the area of movement of the microrobot with an accuracy of 0.1 µm.
The beta microassembly system performs position control using external imaging sensors. Verification of their microassembly is made using 3D imaging, yielding the microsystems field of error at 2 to 3 µm. Allegro refers to her work as "visually guided microassembly for the hybrid integration of microsystems." In one of her demonstrations, Allegro used a 2D microscope that provides optical feedback to the computer, which then instructs the microgripper to a wafer. In this instance, the microgripping tool is a vacuum with a 300-µm dia. opening for suction. The microvacuum attaches to the wafer and then moves it to successfully place the wafer onto a small pin (see figure 1).
Figure 1. A microvacuum gripper, 300 µm in diameter, picks up wafers measuring 400 x 400 µm and stacks them onto small pins of 350 µm.
This position control has been applied to the assembly of micromotors as well.
One of the microgrippers being developed employs the field of piezoelectrics. A microgripper, measuring 16 µm thick, is positioned to grip a microrotor. The microrotor measures 250 microns in diameter and is 150 microns thick. The piezoelectrics move the gripper into position to grip the microrotor. The position control commands shift the rotor toward the micromotor and instruct the gripper to release the rotor onto the motor with a high degree of accuracy. The microrobots in this assembly line have achieved speeds of up to 4 mm/second.
Yves Belouard, also working with Sulzmann in Lausanne, employs a microgripper using shape memory alloys (SMA) to assemble submillimeter optical lenses (250-µm dia.) to the end of 10,000 optical fibers bundled together measuring 3 mm in diameter. When those optical fibers are attached to a CCD camera, then an application for a microendoscope is revealed. Heat is applied to the SMA microgripper, while a high resolution camera provides feedback with the 2D object recognition to establish the gripper's relative position in a 3D environment. Heat is applied to the microgripper, causing it to close its gripping fingers around an optical lens. The gripper is moved, by microassembly, to the point of attachment where the heat is quickly dissipated, via a copper plate attached to the back end of the gripper, and the gripper releases the lens onto the end of the optical fibers.
At present, one industrial partner is hoping to bring the microendoscope to the consumer market quite soon. Sulzmann says the area where microrobotics is really needed is the integration of all of these systems for assembly stations. Upon viewing the microassembly in action, it is easy to see that "desktop manufacturing" is not too far off.
Similar work is being performed at the University of Illinois (Chicago, IL) by Brad Nelson and colleagues Yu Zhou and Barmeshwar Vikramaditya. They, too, are working to establish high positioning accuracy in microassembly for hybrid microelec-tromechanical systems (MEMS). Again, real-time sensor feedback guides the assembly strategies for a particular task. Nelson and his group have achieved speeds of up to 2 mm/second with a repeatability of 0.17 µm.
If you ever wondered how transportation fits into the world of microsystems, then start to think about particle mass transportation. Dr. Felix Moesner, along with other colleagues at the Swiss Federal Institute of Technology (Zurich, Switzerland) have grappled with the problem of finding a suitable propulsion motor for a small "vessel." Since propulsion motors are yet to be developed at the micron and submicron level, like Isaac Asimov's submersible, Moesner and his group have instead created an electrostatic "motor" that propels particles in a dielectric liquid. In other words, rather than attaching a motor to move the particle, move the media in which the particle(s) sits in the direction you want it to go. In this case, the motor is fabricated by winding a number of parallel copper wires, with a diameter from 56 µm to about 236 µm, onto a 6-mm glass tube. The tube is placed in a dielectric liquidin this case, corn oiland when time-varying voltages are applied to the wires it causes an electric field to travel the length of the tube. The electric field has an effect on the liquid that pulls the particles in the direction the charge is traveling. The resulting movement forces the liquid to channel through the interior of the glass tube. The movement of the particles within the liquid is in direct relationship to the traveling electric field of the motor. Moesner has achieved a particle speed of 1.2 mm/second and he believes that a higher propulsion rate can be easily achieved.
Back to American soil, we find interesting inroads working to "improve manual precision in microsurgery without resorting to an expensive teleoperative setup," according to Dr. Cameron Riviere, Carnegie Mellon University (Pittsburgh, PA). Riviere is developing an active handheld instrument that would compensate for hand tremors and other manual positioning errors that limit accuracy and precision in microsurgery. Their current prototype addresses ophthalmological microsurgery and has an intra-ocular shaft which can sense and determine the instrument tip's position. Using neural network methods and adaptive algorithms, a miniature manipulator, controlled by three piezoelectric elements, connects the intra-ocular shaft to a handle in a handheld instrument. As information is collected, the system compensates for the errors, relocating the tip of the intra-ocular shaft. Riviere explains, "One of the primary challenges has been to find an actuator that will offer the necessary range of motion (about 500 µm) and force capability while also meeting the size constraint imposed by the handheld instrument." However, Riviere believes they have found a new material and hope to proceed successfully in the near future.
Telepresence has been determined to be a vital part of microrobotics and microsystem fabrication, but it is also a transferable skill for the macro world allowing humans to interact with an environment at great distances. As all eyes turn to the results of the Mars Pathfinder, scientists in the field of telemanipulation and telepresence gear up for higher demand in this area of research. Dr. Matthew R. Stein, Wilkes University (Wilkes-Barre, PA) comments, "NASA controllers are directly presented by the time delay and limited sensing capability problems that occur in remote control of robotic devices." Along these lines, Joe C. Parrish, NASA Headquarters (Washington, D.C.) and University of Maryland, has been working to create a dexterous telerobot for the Ranger Telerobotic Flight Experiment (TFX). The TFX has shifted its application from a free-flying satellite servicing telerobot to a telerobotic servicing unit for the International Space Station. What remains of the original design, after jettison of the spacecraft bus, are the manipulators and other necessary robotic components. While communicating with the shuttle during experimentation, multisecond time delays awkwardly affect the execution of tasks. However, with dexterous telerobotic performance at hand it is the remote control capacity that will determine whether commands will be given from a ground-based control center or not.
In the United Kingdom, Dr. John Pretlove and Ray W. Yu at the University of Surrey find themselves concentrating on the dexterous and accurate manipulation of robot arms. Pretlove and Yu have developed a man-machine interface for commercially available robots using an outfitted glove that serves as a command and input device. Utilizing touch and force feedback, the glove moves a PUMA 500 robot arm that can duplicate the motion of the human arm. Pretlove and Yu are planning to integrate two of these robot arms for a lightweight mobile robot with a visual active telepresence system that they have already developed.
Unlike the passive observation allowed by microscope and telescope, all these emerging technologies allow humans to extend their presence, and manipulate the environment, at scales beyond our normal grasp. From medicine, space exploration, and manufacturing, telemanipulation, telepresense and microrobotics will change the way we view the world. All of these technologies will be presented in further detail at SPIE's upcoming Symposium on Intelligent Systems & Advanced Manufacturing (ISAM '97) in Pittsburgh, 1417 October. For a copy of the program, contact SPIE.
Leticia Cowan is a free-lance writer based in Monterey, California.