Mobile Microrobotics
Introduction
Magnetic principles have proved successful for the untethered submillimeter microrobotics, although challenges still exist in areas of propulsion and control. Our work in this regard is on the design, analysis, and performance results for a bimorph thin film magnetic microrobot utilizing the magnetostrictive principle as a secondary oscillating operation mode. The microrobot is no larger than 580um in its planar dimension and its total thickness is less than 5um. As a robot with magnetic material, it can be operated in a pushing/pulling mode in orthogonal directions for movement in a plane, while it’s powered with an external magnetic field as low as 1mT. For the secondary oscillating mode utilizing the magnetostrictive principle, the induced in-plane strain, resulting in bending and blocking forces were theoretically calculated to prove enough drive force can be generated in this mode. The design is further abstracted and translated into a piezoelectric cantilever FEM model to confirm the theoretical results. Microrobot fabrication and test-bed development based on this analysis was performed which enabled us to participate in the final competition in the 2010 NIST Mobile Microrobot Challenge, with good performance in the Two Millimeter Dash and Freestyle events.
Micro-scale Magnetic Asymmetric thin film Bimorph (µMAB) microrobot
We call our robot a micro-scale Magnetic Asymmetric thin film Bimorph (µMAB) microrobot. Magnetostrictive bending, made from a magnetic film bonded to a nonmagnetic substrate, occurs when the film is magnetized by anapplied field. A magnetostrictive stress is produced in the film. Bending occurs if one end of the two layer structures is clamped. Further, if the deflected end is in contact with some ground or face, a blocking force is produced which is able to provide mechanical work through the friction force it causes. Legs with different geometry dimensions and different contact line/areas are able to lead to different blocking forces and then friction forces. Making use of the friction difference along the contact face between the robot and the supporting substrate can push or pull the robot mass. If the robot is on the substrate and the external magnetic field is static and constant, the robot legs will remain bent; if the magnetic field is a pulsing signal on and off, the legs will perform bending and straightening, which means a walking/crawling motion will result. Further, when the surrounding magnetic field is a high frequency pulsing signal, for example, with the natural frequency of the robot, the robot would be expected to perform very fast walking and/or running motions. The direction control can be readily realized by changing the direction of magnetic field to align the magnetic body.
µMAB Actuation Modes
Micro-scale Magnetic Asymmetric thin film Bimorph (µMAB) microrobot Concept (click image to enlarge)
Modeling & Verification
Determining whether the microrobot has enough power/force to conquer the resistance to move is a critical step that must be validated before manufacturing the µMAB microrobot. We need to calculate how much driving force we can expect based on the commonly available surroundings, such as off-the-shelf electronic components. To calculate the friction force caused by the blocking force, we need to evaluate the deflection of the magnetostrictive bimorph layers. Few theories and software tools are able to simulate and predict planar magnetostrictive bimorph’s behavior. What we do here is translate the magnetostrictive problem to a piezoelectric one, because the later situation has more available research and analysis software tools. The relationship between the two domains is shown in the table below (click table to enlarge).
Microfabrication
µMAB Microfabrication Process Flow
Fabricated Prototypes
Mobile Microrobotics Test-bed
The mobile microrobotics test-bed includes two orthogonal coil pairs to provide a magnetic field. The coil pairs surround a testing platform where the robot moves, and are supported by a machined aluminum column. Computer-controlled drive electronics modulate the coils’ operating current and voltage signals.
In the initial platform design, the robots are actuated and controlled with the magnetic field generated by the two coil pairs. The small and large coil pairs are made from copper magnetic wire with 110 and 180 turns, with diameters of 2.2” and 3.1”, respectively. Solid iron cores can be inserted into each coil to provide increased magnetic field strength in the workspace of the robot. The coil pairs are supported by laser-cut acrylic braces along with the testing platform, creating a coil assembly unit.
Modifications to the initial design have been made to include a bottom coil, embedded under the working platform, made of 50 turns of copper tape. With constant current flow through the bottom coil, it can pull down the robot piece to the working plane and we can more easily observe the oscillating working mode due to in-plane field and strain. Because of the narrow working space, a simple flanged chamber also made from acrylic has also been added that can easily slide in and out. We can fill the chamber with water or any type of substrate as the working plane at the bottom. The robot’s performance is observed through an overhead firewire CCD camera (Flea2, Point Grey Research, Inc., www.ptgrey.com), a 0.7X to 3X focus lens (VZM 300i, Edmund Optics, www.edmundoptics.com), fiber optic light source and light ring (MI-150, Edmund Optics). The camera and lens are mounted on a rack and pinion focusing mount (NT54-793, Edmund Optics) that is attached to a vertical shaft on a circular baseplate. The camera is connected to a control computer to capture real-time images of the robot.
The drive electronics box is hooked up to two 35V/10A variable power supplies, each with two channels. Since the magnetic field is determined by the amount of current flowing through the coils, the drive electronics were designed to adjust the level of current through each coil independently. A series of solid state relays are used to control which of the four coils are activated. Power resistors and supply voltages are set to determine appropriate coil current levels. Two sets of power resistors (5 and 10 Ohm resistances) are used, allowing for high and low power mode operations for each of the coils. The modes are selected with a switch on the drive electronics control box. Also through a switch on the box, the two orthogonal coil pairs can be set to operate in either Helmholz or Maxwell coil pair configurations, allowing for the same or opposite current directions through the coils in the pair, respectively. Four relays are used to control one small coil and one big coil (i.e. one of the two coils in the coil pair). The relays (coils) can be opened and closed individually or simultaneously. The electronics have been designed to provide either constant or pulsed signal waveforms to the coils yielding constant or oscillating external fields to the robot workspace. These control signals, along with the control signals to operate the solid state relays, are sent via computer control of a data acquisition board (LabJack U3-HV and CB15 Board, www.labjack.com). The system is able to pulse signals with a frequencies as high as 100kHz, which will cover the range of the first few natural frequencies of the µMABs. The electronics can provide up to 6A of current per coil. With an input current of 2A in one coil pair, a magnetic field of about 0.5mT is able to be produced at the center area of the coil pairs, which is enough to drive the robot on a dry, flat substrate. A Matlab-based graphical user interface was created to operate the system.
Results
The vibration actuation mode, due to induced vibration of the robot from an oscillating (pulsing) magnetic field based on the magnetostrictive mechanism, has been observed. A pulsed frequency of approximately 6 kHz was applied to the small coil pair, in the Helmholz configuration and the high power mode (~ 3A/coil), in 20 pulse increments. The pulse signal caused the main body of the robot to vibrate/deflect and the robot to translate across the substrate. At the conclusion of the pulse train, the robot motion ceased. Upon the application of another magnetic field pulse train, robot movement resumed in a similar manner. A video of one of these such tests is shown to the right.
In this working mode, the robot moves relatively steady and controllable because the end point of each “walking” step is predictable and in limited distance. The bottom coil is on with constant current at the same time, pulling the robot body
down for steady gait. Although the translation direction is not accordant with the robot’s front and rear leg as expected, it shows different light reflections between stable and moving phase, which indicates the film vibrating. It also shows that the deflection power (force) is much bigger than the frictional resistance generated by gravity, because even a friction difference, not due to the asymmetric structure, is able to translate the robot.
The second operating mode exhibited by the µMAB's is pushing/pulling the magnetic body with a magnetic field. In this mode, the magnetized nature of the robot is responsible for the motion. Different currents are input into each coil in a coil pair (small or large) to create a gradient magnetic field to direct the magnetic force and results in translation of the µMAB on a dry surface with very fast speeds.
This mode was used very successfully in the Two Millimeter Dash event at the challenge. The µMAB achieved one of the fastest individual runs of the competition for the dash at only 27 ms, as shown in the below video.
Moreover, the translation direction for the µMAB is along the magnetic field line, which means the translation direction is predictable and controllable.By creating gradient fields with adjacent coils in the orthogonal sets in different pairs (i.e. one small coil and one large coil) diagonal translation is possible.
We took advantage of this controllability in the Freestyle event in the competition, realizing automated horizontal, vertical and diagonal translation through a simple series of commands. This video is shown here.
In subsequent work, we have tested the robot’s performance with new testbed components and in a fluid environment for more uniform, larger damping and better control. A bottom coil was also utilized to increase the stability of the µMAB’s movement. Much smoother movement trajectories were realized in the chamber filled with water. A curvilinear C-shaped path on the substrate was successfully traversed. Similar continuous translation tests were conducted, which were composed to trace out our school’s initials, S-I-T, with the robot. The trajectories from these experiments were extracted off-line using the image processing toolbox in Matlab. Plots of these trajectories along with videos are shown below.
Related Publications
Jing, W., Chen, X., Lyttle, S., Fu, Z., Shi, Y., and Cappelleri, D., “Design of a magnetostrictive thin film microrobot”, Proceedings of the ASME International Mechanical Engineering Congress & Exposition, November, 2010.
Jing, W., Chen, X., Lyttle, S., Fu, Z., Shi, Y., and Cappelleri, D., “A Magnetic Thin Film Microrobot with Two Operating Modes”, Proceedings of the IEEE International Conference on Robotics and Automation (ICRA), Shanghai, China, May, 2011.
Copyright Notice
Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the publishers.