Get the Flash Player to see this rotator.
 
 


search


Robotics     UMD

Search robotics news archives

Bookmark and Share

logo


Ant-Like Micro-Robots: Fast, Small, and Under Control

Main Participants

P. Abshire, S. Bergbreiter, N. Martins, E. Smela

Sponsor

This project is sponsored by NSF.

Keywords

Bioinspired, adaptive integrated circuits, micro-robotics, antbots, distributed algorithms, rapid prototyping, MEMS, polymer actuators

Motivation

Ants are amazing. They are tiny machines with tiny brains that nevertheless work together to achieve very complex goals. They employ a variety of specialized forms for different tasks and exploit favorable scaling laws to move around in, and manipulate objects in, their environment.

We are performing research to provide the core fundamental knowledge needed to realize artificial ants: how to embed communication, computation, and actuation in a very small form factor to create adaptable, coordinated behavior in robots on a size scale that has not previously been realized. The ultra-small size of these systems absolutely requires co-development of the computational and physical components, which will be severely size and power constrained.

The realization of micro-robots could lead to technological breakthroughs and could impact many different application areas, including search and rescue, infrastructure and equipment monitoring, and possibly even the medical field. Very small robots have a number of advantages: they are small enough not to be seen, and small enough to crawl through small crevices.

Objectives

We are performing fundamental research on realizing cooperative, sub-cm3, ant-sized micro-robots that self-organize into an optimal formation, and are developing the methods and technologies in the context of extreme power and size constraints. There are five key objectives.

1. Develop distributed algorithms for formation control of the micro-robot fleet.

2. Design minimal electronics hardware for robot control using event-based communication and computation, ultra-low-power radio, and adaptive analog-digital integrated circuits.

3. Implement the algorithms in the event-based hardware so that they perform robustly even under extreme power and size constraints.

4. Develop efficient actuators using both rapid-prototyping and MEMS technologies that are robust enough to operate under real-world conditions, and develop methods of locomotion using these actuators.

5. Integrate the algorithms, electronics, and actuators into a fleet of ant-size micro-robots and demonstrate formation control.

Overview of approach

Cooperative Motion. Coordination between micro-robots requires communication between them. However, the antbots will operate under the constraints that resources for communication are severely limited, that communication links will be stochastic, and that both coordination and communication success are strongly linked to the locomotion methods. We are designing distributed control and coordination algorithms for wireless networks of micro-robots so that they self-organize into a desired formation from an arbitrary initial position, taking into account the constraints of the physical system, such as the facts that actuation is stochastic and under-actuated, that measurement of positions depends on the antenna configuration, and that power is precious. The algorithms will be informed by the development of new, more realistic locomotion models, taking into account fabrication and integration challenges. The algorithms will be implemented in a miniature robot test-bed based on ZigBots (for ZigBee robots) which will be cm-scale in size and constructed using commercial off the shelf (COTS) components.

Robot Architecture. Most robots today use COTS electronic components for communication, computation, and control. Relatively little development has occurred for extremely resource-constrained robots, and little has been done to understand the minimal infrastructure requirements for communication, computation, and control in such systems. To get to the sub-cm scale requires a significant restructuring of the electronics hardware. Somehow ants do it without microcontrollers. They use a relatively simple event-driven infrastructure based on neural wetware. New electronics architectures are therefore needed to realize these tiny robots. This project is adapting new technologies from the field of wireless sensor networks, and at the same time is exploring new ideas for hardware implementations of robotic control systems, such as an event-based infrastructure to connect the communication, computation, and control components that make up the micro-robot.

Computation. The electronic infrastructure for the micro-robots is clearly critical; just as important is the issue of how to efficiently use it. This research must proceed in tight coordination with the algorithm and hardware development since the specific algorithms and implementation drastically affect how efficiently they may be realized in specific hardware. To satisfy power and space constraints, the computational infrastructure will be implemented using hybrid computational approaches including analog, digital, and mixed-signal approaches, adaptive integrated circuits, time-based computation, and distributed algorithms that partition computations between local and global hardware.

Locomotion & Actuation. The antbots must be able to move around in a real world environment, and to do so without being tethered to a power supply. This requires that the actuators be both highly robust and highly efficient. This project will tackle challenges such as incorporating compliant materials and mechanisms into micro-fabricated systems in order to address the first issue, and will explore highly efficient actuators such as dielectric elastomer actuators (DEAs) to address the second. While DEAs have been realized on the meso-scale, microfabricating them is a challenge because none of the materials and none of fabrication methods used on the meso-scale can be used on the micro-scale. As part of this research, we are developing the necessary approaches to achieve the required miniaturization.

It is also important to realize various micro-robot designs in order to better understand their relative advantages, constraints, and challenges for locomotion, so this project is using a rapid micro-robot prototyping (RAMP) process in order to understand device designs and to optimize micro-robot mobility.

Multiphysics Integration. Many technical challenges must be overcome in order to realize the ant-like robots, including how to provide the high voltages required for actuation, how to physically interconnect and configure the hardware components, and how to electrically interconnect the components. Our work is considering these integration issues right from the start, approaching the system design in a holistic way.

Contact

Dr. Pamela Abshire
Department of Electrical and Computer Engineering and Institute for Systems Research
2211 A. V. Williams Bldg.
University of Maryland
College Park, MD 20742
Phone: 301-405-6629
Email: pabshire@umd.edu

Dr. Sarah Bergbreiter
Department of Mechanical Engineering and Institute for Systems Research
2170 Martin Hall
University of Maryland
College Park, MD-20742
Phone: 301-405-6506
Email: sarahb@umd.edu

Dr. Nuno C. Martins
Department of Electrical and Computer Engineering and Institute for Systems Research
2259 A. V. Williams Bldg.
University of Maryland
College Park, MD 20742
Phone: 301-405-9198
Email: nmartins@umd.edu

Dr. Elisabeth Smela
Department of Mechanical Engineering
2176 Martin Hall
University of Maryland
College Park, MD 20742
Phone: 301-405-5265
Email: smela@umd.edu


 

   
Back to top      
Clark School Home UMD Home ISR Home