A Modular Robotic System Using Magnetic Force EffectorsBrian T. Kirby, Burak Aksak, Jason D. Campbell, James F. Hoburg,Todd C. Mowry, Padmanabhan Pillai, Seth Copen GoldsteinAbstract— One of the primary impediments to buildingensembles of modular robots is the complexity and numberof mechanical mechanisms used to construct the individualmodules. As part of the Claytronics project—which aims tobuild very large ensembles of modular robots—we investigatehow to simplify each module by eliminating moving partsand reducing the number of mechanical mechanisms on eachrobot by using force-at-a-distance actuators. Additionally, weare also investigating the feasibility of using these unaryactuators to improve docking performance, implement inter-module adhesion, power transfer, communication, and sensing.In this paper we describe our most recent results in themagnetic domain, including our first design sufficiently robustto operate reliably in groups greater than two modules. Ourwork should be seen as an extension of systems such asFracta [9], and a contrasting line of inquiry to several otherresearchers’ prior efforts that have used magnetic latching toattach modules to one another but relied upon a poweredhinge [10] or telescoping mechanism [12] within each moduleto facilitate self-reconfiguration.I. INTRO DUCTIONAdvances in manufacturing and electronics open up newpossibilities for designing modular robotic systems. As therobots become smaller, it becomes possible to use force-at-a-distance actuators—e.g., actuators which cause one moduleto move relative to another via magnetic or electric fieldsexternal to the modules themselves. Furthermore, as the costand power consumption of electronics continue to decrease,it becomes increasingly attractive to use complex electronicsrather than complex mechanical systems. In this paper, weexplore how a single device that exploits magnetic forces canbe harnessed to unify actuation, adhesion, power transfer,communication, and sensing. By combining a single coilwith the appropriate electronics we can simplify the robot—reducing both its weight and size—while increasing itscapabilities.The robots described in this paper are the result of ourexplorations into the underlying ideas of the Claytronicsproject [4], which is investigating how to design, build,program, and use ensembles comprised of massive numbersof robotic modules. Thus, one of the main driving designThis work was supported in part by DARPA/SPAWAR N66001-04-1-89XX, NSF CNS-0428738, and Intel Corporation. We want to thank themembers of the claytronics group for their many valuable insights.Kirby, Goldstein in SCS at CMU, 5000 Forbes Ave, 15213{bkirby, seth} @cs.cmu.eduAksak in ME at CMU [email protected] in ECE at CMU [email protected], Mowry, Pillai at Intel Pittsburgh Research{jason.d.campbell, todd.mowry,padmanabhan.s.pillai} @intel.comFig. 1. Three magnetic 45mm planar catoms. Videosdemonstrating their movement capabilities are available athttp://www.cs.cmu.edu/˜claytronics/iros07/planarcatom/.criteria for any individual mechanism is: will it supportscaling the ensemble to larger numbers of units?. A directoutgrowth of this design criteria is that each unit in theensemble must be inexpensive, robust, and easy to manu-facture. Hence mechanisms used for locomotion, adhesion,communication, etc., must be as simple as possible. One wayto achieve this is to use inexpensive and robust resources—e.g., computation—to reduce mechanical complexity. Fur-thermore, since we are interested in the ensemble as a whole,we do not require that individual units be self-sufficient.As long as individual units can contribute to the overallmotion of the ensemble, they do not need the ability tomove independently within the greater environment. We callthis design principle the ensemble axiom: each unit containsonly the minimum abilities necessary to contribute to theaggregate functionality of the ensemble.Choosing the right mechanism for locomotion is a keydesign decision. In addition to scalability, the size of theunit must also be taken into account. At the macroscale,complex mechanisms such as motors are effective. However,as units scale down in size other approaches become viable,taking advantage of increasing surface-to-volume ratio anddecreasing of inertial moments. Our current robots, whichwe call planar catoms1, are small enough that we canexplore a mechanism designed around magnetic field force-at-a-distance actuators. As the units decrease further in size,actuators based upon electric field forces become viable andare appealing because they use less current, produce less1“Catom” is short for “claytronics atom.”heat, and weigh less than magnetic actuators. Even smallerunits could harness surface forces such as surface tensionor Van der Waals’ forces. The size scale also affects powertransfer and storage: because electrical resistance increasesas contact size decreases, direct electrical connections be-tween robots become increasingly impractical. We chose thecentimeter scale for our initial prototypes to keep the small-scale prototyping costs of our onboard circuitry reasonable.In keeping with our design principle, we demonstrate45mm diameter cylindrical modular robots (see Figure 1)that can move in a plane and use a single, no-moving-parts mechanism—an electromagnetic coil—for locomotionand adhesion (Section III), power transfer (Section IV),and communication and neighbor sensing (Section V). Theability to implement a number of features using the samemechanism allows us to reduce the weight, volume, andoverall complexity of the unit.II. RELATED WORKThe effort to produce reliable and robust modular roboticsystems has led researchers to explore a large design spaceof mechanisms for locomotion, adhesion, communication,and power. Ostergaard, et al. survey different locomotion andadhesion mechanisms for self-actuating robots in [5].Of the many research efforts the most relevant to ourwork is Fracta [9]. Fracta is a two dimensional modularrobot which uses a combination of permanent magnets andelectromagnets for locomotion and adhesion. It is the onlyother internally actuated system which has no moving parts.As in our planar catoms, to move a module requires com-munication between the moving module and its neighbors.The two main differences between Fracta and planar catomsare due to changes in underlying technology and the use ofpermanent magnets. Fracta modules