Swarms that "hear the shape of the drum"

Funded by: NSF

The concept in this research effort is to develop robotic swarms that collectivly make up a spattially distributed associate memory, that is robust to noise and uncertainty. Basically, a distributed neural network is implemented in locally interconnected chunks over the swarm members. This robot swarm would ultimately fall into formation, scan an area of interest and collect measurements, and decide on the pattern of the data collected on their own, without human supervision.

Publications

coming soon...

Micro Autonomous System Technologies

Funded by: US Army Research Lab

Processing for Autonomous Operation

The vision of the Autonomous Operations group, consisting of several groups from academia (UPenn, Berkeley, GATech, UNM, MIT, Univ of Sydney, Vanderbilt, and Univ of Maryland) and industry (BAE Systems) is to develop Autonomous Multifunctional Mobile Microsystems (Am3), a networked group of small vehicles and sensors operating in dynamic, resource-constrained, adversarial environments. While individual units may be specialized, Am3 will be multifunctional because of its heterogeneity, the ability of individual units to automatically reconfigure and adapt to the environment and to human commands, and its distributed intelligence. Am3 will need to operate with little or no direct human supervision, because groups like this will be very difficult, if not impossible, to efficiently manage or control by programming or by tele-operation. The deployment, monitoring, and tasking of such multifunctional groups will be challenging and will require the application of new, yet-to-be-developed methods of communication, control, computation and sensing, specifically tailored to Mast applications.

In this effort, UD works closely with UNM, UPenn and Berkeley to develop algorithms that would enable robotic crawlers to navigate optimally in uncertain and complex environments. Our approach is based on a combination of artificial potential fields (navigation functions), and model predictive control design using randomized algorithms.

An AM3 prototype. Such bioinspired vehicles are envisioned to work in groups to provide situational awareness.

The graphical user interface of a matlab tool we developed for simulating receding horizon control strategies. The control design aims at communication-aware navigation in uncertain, cluttered environments, and utilizes randomized algorithms.

Publications

Herbert G. Tanner and Jorge L. Piovesan, "Randomized Receding Horizon Navigation," IEEE Transactions on Automatic Control, 2009 (submitted)

Jorge Piovesan and Herbert Tanner, "Relaxed stability conditions for switching systems," Systems and Control Letters, 2009 (submitted)

Jorge Piovesan and Herbert Tanner, “Randomized Model Predictive Control for Robot Navigation,” IEEE International Conference on Robotics and Automation, 2009, pp. 94-99

Formal Cooperative Planning
of Decentralized Robot Actions

Funded by: NSF

Overview

The goal in this project is to develop a methodology with which groups of robots of different types can decide on their own how to put together their capabilities in order to solve cooperatively problems that none alone can solve. The envisioned methodology has several stages, in which decentralized group controllers are first designed to make each homogeneous group behave as a single entity, then the controlled group behaviors are modeled as letters of a behavioral alphabet, and finally cooperative grammars are synthesized to allow the composed heterogeneous systems to reason about how to jointly work together to complete the task. Our proposed composition of behaviors demonstrates that the whole can be more than the sum of the parts

This material is based upon work supported by the National Science Foundation under Grant No. 00447898. Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.

Publications

Ye Yuan and Herbert G. Tanner, "Sensor graphs for guaranteed cooperative localization performance," Control and Intelligent Systems, 2010 (to appear)

Herbert G. Tanner, Jorge L. Piovesan and Chaouki T. Abdallah, "Finite asymptotic abstractions for hybrid systems with stable continuous dynamics," Journal of Discrete Event Dynamical Systems, 2009 (submitted)

Jorge L. Piovesan, Chaouki T. Abdallah, and Herbert G. Tanner, “Modeling Multi-Agent Systems with Hybrid Interconnected Dynamics,” American Control Conference, 2009 (to appear)

Jorge L. Pioveasn, Chaouki T. Abdallah and Herbert G. Tanner, "Preliminary Results on Interconnected Hybrid Systems." 16th Mediterranean Conference on Control and Automation, pp. 101-106.

Wenqi Zhang and Herbert G. Tanner, "Composition of Motion Description Languages," Hybrid Systems: Computation and Control, M. Egerstedt and B. Misrha (eds), Springer 2008

A. Clauset, H.G. Tanner, C.T. Abdallah, and R.H. Byrne, "Controlling across complex networks; emerging links between networks and control," IFAC Control Engineering Practice, 32:183-192, 2008.

Jorge L. Piovesan, Chaouki T. Abdallah, and Herbert G. Tanner, "A Hybrid Framework for Resource Allocation among Multiple Agents Moving on Discrete Environments," Asian Journal of Control, 10(2):171-186, 2008

A. Closet, B. Tanner, R. Byrne and C.T. Abdallah, “Controlling Across Complex Networks: Emerging Links between Networks and Control,” Proceedings of the IFAC Time-Delay Symposium, Nantes France, 2007.

D. Kumar and H.G. Tanner, “How Sensor Graph Topology Affects Localization Accuracy,” European Control Conference, 2007 pp. 868-873

Jorge Piovesan, Chaouki T. Abdallah, Magnus Egerstedt, Herbert Tanner and Yorai Wardi, “Statistical Learning for Optimal Control of Hybrid Systems,” IEEE American Control Conference, 2007, pp 2775-2780

Herbert Tanner, Jorge Piovesan, Chaouki T. Abdallah, "Discrete Asymptotic Abstractions of Hybrid Systems," 45th IEEE Conference on Decision and Control, pp 917-922

Herbert G. Tanner and Amit Kumar, "Formation Stabilization of Multiple Agents Using Decentralized Navigation Functions," Robotics: Science and Systems I, S. Thrun, G. Sukhatme, S. Schaal and O. Brock (eds), MIT Press, 2005, pp 49–56.

Distributed Cooperative Robotic Radiation Mapping

Funded by: DoE

Overview

Both in emergency management as well as nuclear non-proliferation, time and personnel may be a critical resource in limited supply. Robots can help in providing quick situational awareness and assessment by working in groups mapping radiation levels over certain areas of interest. In this project we focused on detecting and mapping the presense of weakly radioactive sources: small isolated specs of radioactive material (gamma emitters), or larger quantities of shielded special nuclear material (alpha emitters). In such cases, the search needs to take into account the statistics of nuclear measurement, and results are presented at certain levels of confidence, depending on the exposure time of the radiation detector. Our approach to cooperative search and mapping is based on prioritizing measurement collection using the mutual information associated with each local measurement. Thus, robots take measurements first in locations where data maximize the operator's knowledge about his workspace, and reduce the uncertainty of the map in a locally optimal way. The radiation map that robots construct is built incrementally, the most interesting locations are mapped first, and the map's confidence increases as more time for measurement collection becomes available.

We recently (August 2008) extended this methodology to the case where mutliple different maps (e.g. radiation and temperature) are concurrently constructed, by combining the information content of all sensing modalities.

A Khepera II mobile robot interfaced with a small CsI (Cesium Iodide) radiation sensor and the associated electronics for signal processing.

Khepera II robots simulate radiation mapping by filtering IR readings (from the red light source) through a Poisson process. Robots localize by triangulating their Cricket distance measurements.

Publications

R.A. Cortez, X. Papageorgiou, H.G. Tanner, A.V. Klimenko, K.N. Borozdin, R. Lumia, J. Wood, and W.C. Priedhorsky, "Smart Radiation Sensor Management; Nuclear Search and Mapping using Mobile Robots," IEEE Robotics & Automation Magazine, Vol 15, Issue 3, pp. 85-93, 2008

R. A. Cortez, H. G. Tanner and R. Lumia, “The Entropy of Cooperative Radiation Sensing by Distributed Sensors,” 2nd ANS International Joint Topical Meeting on Emergency Preparedness & Response and Robotic & Remote Systems, March 2008 (in print).

R. A. Cortez and H. G. Tanner, “Radiation Mapping Using Multiple Robots,” 2nd ANS International Joint Topical Meeting on Emergency Preparedness & Response and Robotic & Remote Systems, March 2008 (in print).

R.A. Cortez, X. Papageorgiou, H.G. Tanner, A.V. Klimenko, K.N. Borozdin and W.C. Priedhorsky, "Experimental Implementation of Robotic Sequential Nuclear Search," IEEE Mediterranean Conference on Control and Automation, 2007 pp 1-6

A. Kumar, H.G. Tanner, A.V. Klimenko, and W.C. Priedhorsky, "Automated Sequential Search for Weak Radiation Sources," 14th IEEE Mediterranean Conference on Control and Automation, June 28-30, 2006, Università Politecnica delle Marche, Ancona, Italy, pp 1-6.

A.V. Klimenko, W.C. Priedhorsky, H. Tanner, K.N. Borozdin, and N. Hengartner, “Intelligent Sensor Management in Nuclear Searches and Radiological Surveys,” ANS Transactions, vol 95, pp 21-22, 2006.

K. N. Borozdin, A. V. Klimenko, W. C. Priedhorsky, N. W. Hengartner, C. C. Alexander, R. A. Cortez, and H. G. Tanner, "Optimized Strategies for Smart Nuclear Search," 2006 IEEE Nuclear Science Symposium and Medical Imaging Conference, San Diego CA, pp 926-928

Task-driven multi-formation control for coordinated UAV/UGV ISR missions

Funded by: Sandia National Labs

Overview

In this project we developed cooperative control algorithms that coordinate formations of Unmanned Aerial and Ground Robotic Vehicles for autonomous intelligence gathering, surveillance and reconnaissance (ISR). We devised a search strategy that is based on the different capabilities of these heterogeneous units, and exploits the size of the groups to execute formation coverage and coordinated parallel scanning of an area of interest. Within this context we proposed new types of formation and flocking control algorithms, as long as a combination of a consensus protocol with a particular TDMA-like communication scheme that makes the flocking algorithm robust to communication delays.

A group of four UAVs orbit over a moving UGV formation. They move on spiral orbits while their scanning footprints (marked by the conical regions) keep track of the formation perimeter.

Groups of UGVs secure an area of interest by falling into formation. Their formation decomposes the area into rectangular segments and their sensors can detect target crossings over the boundaries. Then the area is swept by a formation of UAVs.

Watch the video here.

Pubications

Herbert G. Tanner and Dimitrios Christodoulakis, “Decentralized Cooperative Control of Heterogeneous Vehicle Groups,” Robotics and Autonomous Systems, vol 55, no 11, pp 811-823, 2007

Herbert G. Tanner, “Switched UAV-UGV Cooperation Scheme for Target Detection,” IEEE International Conference on Robotics and Automation, 2007 pp. 3457-3462

Herbert Tanner and Dimitrios Christodoulakis, "Cooperation between Aerial and Ground vehicle groups for Reconnaissance missions," 45th IEEE Conference on Decision and Control, 2006 pp. 5918-5923

Herbert G. Tanner and Dimitrios K. Christodoulakis, "The stability of synchronization in local-interaction networks is robust with respect to time delays," 44th IEEE Conference on Decision and Control, 2005, pp 4945–4950.