Design and implementation of a millirobot for swarm studies –mROBerTO

Abstract

SUMMARYThe use of millirobots, particularly in swarm studies, would enable researchers to verify their proposed autonomous cooperative behavior algorithms under realistic conditions with a large number of agents. While multiple designs for such robots have been proposed, they, typically, require custom-made components, which make replication and manufacturing difficult, and, mostly, employ non-modular integral designs. Furthermore, these robots' proposed small sizes tend to limit sensory perception capabilities and operational time. Some have resolved few of the above issues through the use of extensions that, unfortunately, add to their size.In contribution to the pertinent field, thus, a novel millirobot with an open-source design, addressing the above concerns, is presented in this paper. Our proposed millirobot has a modular design and uses easy to source, off-the-shelf components. Themilli-robot-Toronto (mROBerTO) also includes a variety of sensors and has a 16 × 16 mm2footprint.mROBerTO's wireless communication capabilities include ANT™, Bluetooth Smart, or both simultaneously. Data-processing is handled by an ARM processor with 256 KB of flash memory. Additionally, the sensing modules allow for extending or changing the robot's perception capabilities without adding to the robot's size. For example, the swarm-sensing module, designed to facilitate swarm studies, allows for measuring proximity and bearing to neighboring robots and performing local communications.Extensive experiments, some of which are presented herein, have illustrated the capability ofmROBerTOunits for use in implementing a variety of commonly proposed swarm algorithms.

A high-performance millirobot for swarm-behaviour studies: Swarm-topology estimation

“…In our previous work, reported in the work of Kim et al 27,28 and Drisdelle et al 29 , mROBerTO was directly compared to the two best robots in the field, the Kilobot and TinyTeRP , and was determined to be clearly superior in locomotion, sensing and processing performance. Thus, the experiments presented in this section focus only on validating the global swarm-topology estimation method proposed in this article.…”

“…Validation of the proposed method’s ability to estimate the topology of a larger scale swarm is shown in this subsection for a group of 30 simulated robots. The acquired localization data for the non-physical robots included noise modelled to emulate the characteristics of real-world sensing conditions 28 for the mROBerTO s. The true and the estimated topologies for the swarm considered are shown in Figure 15.…”

“…It comprises four modules: the locomotion module, the processing and communication module and two sensing modules – proximity sensing and swarm sensing(see Figure 1). 27,28 mROBerTO ’s locomotion module contains all the components necessary to enable robot motion: two DC motors, an H-bridge for driving the motors and voltage regulators to provide power. Motion control is achieved using a kinematic model of the robot and a calibrated translator between desired motor angular velocities and pulse width modulation (PWM) levels.…”

“…Each of the above-mentioned modules, as well as their integration, is described in detail in the work of Kim et al 27,28 and Drisdelle et al 29 – as such, they are omitted herein. The specific description of the swarm-sensing module below, thus, is only due to its immediate relevance to the swarm-topology estimation algorithm proposed in this article.…”

Slip and Magnetic Attraction Effects in a Microrobot with Magnetic-Wheels and Skid-Steering

“…Caprari et al [9] developed a microrobot called Alice with dimensions of 21 mm × 21 mm × 18 mm as a research and development platform. Kim et al [10] developed a millirobot (mROBerTO) for swarm studies.…”