HOME    |    RESEARCH    |    TEACHING    |    GROUP    |    NEWS    |    PUBLICATIONS    |    CONTACT

Optimal Communication for Fast Sensor Network Coordination  (Support: NSF NeTS)

The goal of this project is to develop a new framework to control the structure of wireless robot networks so that the performance of networked dynamical processes carried out by the robots, such as decentralized estimation, information spreading, or synchronization, is improved. As the behavior of such processes directly depends on the network eigenvalue spectra, new metrics need to be designed that relate the network spectra to the network structure and capture the complexity of the wireless channel and the richness of the communication space, e.g., the end-to-end rates and routing decisions. This project develops efficient approximations of these metrics that can be computed in a decentralized way, as well as ways to control them in the joint space of robot positions and network configurations. It points to a new direction in the design of networked dynamical processes where the network structure is not only controlled for connectivity, but also for optimization of the dynamical process itself.

Selected Publications:
[1] Distributed Network Design for Laplacian Eigenvalue Placement. V. M. Preciado and M. M. Zavlanos. IEEE Transactions on Control of Network Systems, Vol. 4, No. 3, pp. 598-609, September 2017.
[2] Spectral Control of Mobile Robot Networks. M. M. Zavlanos, V. M. Preciado, and A. Jadbabaie. Proc. 2011 American Control Conference, San Francisco, CA, June 2011, pp. 3245-3250.


Controlling Teams of Autonomous Mobile Beamformers  (Support: NSF NeTS)

The goal of this project is to develop a new framework to control teams of mobile robots, cooperating in a beamforming fashion, to transmit information between multiple source-destination pairs, while meeting quality-of-service constraints and consuming minimum power. This optimization depends on the wireless channel, but also on the robot positions, which offer an additional degree of freedom due to robot mobility. It gives rise to a novel system of mobile beamformers that allows for significant performance gains compared to static systems that do not consider mobility. Additionally, unlike other communication protocols, such as multi-hop, it does not require high node density which results in packet collisions and delays, unreliable links, and difficulties in routing. The approach proposed in this project ensures robust communications and longevity in challenging environments, arising during the transmission of high-rate data, such as video or images, or in environments where there is no line-of-sight.

Selected Publications:
[1] Distributed Cooperative Beamforming in Multi-Source Multi-Destination Clustered Systems. N. Chatzipanagiotis, Y. Liu, A. P. Petropulu, and M. M. Zavlanos. IEEE Transactions on Signal Processing, Vol. 62, No. 23, pp. 6105-6117, December 2014.
[2] Controlling Groups of Mobile Beamformers. N. Chatzipanagiotis, Y. Liu, A. P. Petropulu, and M. M. Zavlanos. Proc. 51st IEEE Conference on Decision and Control, Maui, Hawaii, December 2012, pp. 1984-1989.


Controlling Teams of Mobile Microrobots using External Electromagnetic Fields  (Support: NSF RI)

The goal of this project is to develop a new framework to individually control teams of mobile mictorobots using external electromagnetic fields. External electromagnetically powered microrobots require the same control signal to be sent to all the microrobots in the workspace. Control strategies for this situation exist, but the tasks that can be achieved by multiple agents are limited by constraints such as a low number of robots, non-smooth trajectories, the robots meeting in the same location, congregating close together, and collision avoidance. In some cases, the control strategies assume microrobot capabilities that are not yet realizable in practice. The key idea that enables this work is to decompose the workspace into cells, each one defined by the convex hull of a set of active microcoils, and develop hierarchical control strategies to drive teams of microrobots, jointly, through sequences of cells until they reach their final destinations. This research will allow for truly independent and coordinated control of mobile magnetic microrobots.

Selected Publications:
[1] Control of Magnetic Microrobot Teams for Temporal Micromanipulation Tasks. Y. Kantaros, B. Johnson, S. Chowdhury, D. J. Cappelleri, and M. M. Zavlanos. IEEE Transactions on Robotics, under review.
[2] Towards Mobile Microrobot Swarms for Additive Micromanufacturing. D. Cappelleri, D. Efthymiou, A. Goswami, N. Vitoroulis, and M. M. Zavlanos. International Journal of Advanced Robotic Systems, Vol. 11, No. 150, pp. 1-14, September 2014.


Graph Theoretic Connectivity Control of Mobile Robot Networks

Communication and network connectivity has emerged as one of the most important and critical requirements in numerous cooperative tasks, such as formation stabilization and consensus seeking problems. While the agents’ primary task is typically detection of certain physical changes within their proximity, their communication capabilities enable them to share the individually collected data with their peers, in order to achieve a global coordinated objective. Consequently, network connectivity has become a critical design specification of a network and, due to its global nature, distributed solutions are rather hard to obtain. The focus of this research is on distributed and provable algorithms that maintain connectivity in networks with dynamically changing communication topologies.

[1] Graph Theoretic Connectivity Control of Mobile Robot Networks. M. M. Zavlanos, M. B. Egerstedt, and G. J. Pappas. Proceedings of the IEEE: Special Issue on Swarming in Natural and Engineered Systems, Vol. 99, No. 9, pp. 1525-1540, September 2011.
[2] Distributed Connectivity Control of Mobile Networks. M. M. Zavlanos and G. J. Pappas. IEEE Transactions on Robotics, Vol. 24, No. 6, pp. 1416-1428, December 2008.


Distributed Multi-Robot Assignment and Placement

A fundamental yet poorly understood problem in multi-agent coordination is the assignment of multiple tasks to multiple agents. Most existing approaches decouple the assignment process from the execution process that follows any choice of task assignment. Despite the enormous combinatorial complexity of static, discrete assignment problems, in highly unpredictable environments, where the number of targets and agents dynamically change, dynamic assignment methods are clearly much more preferable. In addition to role assignment being dynamic, ideally one would also desire distributed coordination protocols for task assignment, as the best feasible way for addressing the computational complexity of the problem. The focus of this research is on distributed coordination algorithms for task assignment.

Selected Publications:
[1] A Distributed Auction Algorithm for the Assignment Problem. M. M. Zavlanos, L. Spesivtsev, and G. J. Pappas. Proc. 47th IEEE Conference on Decision and Control, Cancun, Mexico, December 2008, pp. 1212-1217.
[2] Dynamic Assignment in Distributed Motion Planning with Local Coordination. M. M. Zavlanos and G. J. Pappas. IEEE Transactions on Robotics, Vol. 24, No. 1, pp. 232-242, February 2008.


Identification of Gene Regulatory Networks in Molecular Biology

Gene regulatory networks capture the interactions between genes and other cellular components, resulting in the fundamental biological process of transcription and translation. In some applications, the topology of the regulatory network is not known, and has to be inferred from experimental data. The experimental data consist of expression levels of the genes, which are typically measured as mRNA concentrations in micro-array experiments. In a so called genetic perturbation experiment, small perturbations are applied to equilibrium states and the resulting changes in expression activity are measured. The focus of this research is on novel optimization techniques to identify genetic networks that explain data obtained from noisy genetic perturbation experiments or time course measurements. The impact of this research is not only in promoting biological knowledge, but also in drug discovery, where a systems-wide understanding of regulatory networks is crucial for identifying the targeted pathways.

Selected Publications:
[1] Inferring Stable Genetic Networks from Steady-State Data. M. M. Zavlanos, A. A. Julius, S. P. Boyd, and G. J. Pappas. Automatica Special Issue on Systems Biology, Vol. 47, No. 6, pp. 1113-1122, June 2011.
[2] Genetic Network Identification using Convex Programming. A. A. Julius, M. M. Zavlanos, S. P. Boyd, and G. J. Pappas. IET Systems Biology, Vol. 3, No. 3, pp. 155-166, May 2009.

Michael M. Zavlanos
Department of Mechanical Engneering & Materials Science

Michael M. Zavlanos    |    Last Update 02.22.2018