Jobs Button PSU Button        

Communication, Information & Navigation (CIN) | intelligent control (ic)

ic.gif Intelligent Control (IC) researches and develops innovative control mechanisms, strategies and related technologies for non-traditional, distributed, and highly nonlinear systems.

Intelligent Control is multidisciplinary, incorporating research in software engineering, formal control theory, formal language theory, information and complexity theory, and signal analysis. Applications include computational languages for distributed autonomous control of robotic devices or nanoparticles, behavior discovery for situation awareness and control adaptation, and multiobjective mission planning for dynamic undersea scenarios.

IC is able to combine internal resources with resources across the ARL and Penn State University to form multi-disciplined research and engineering teams supporting a wide range of programs.

Key Areas of Expertise:

  • Hierarchical Distributed Control Strategies
  • Cooperation, Coordination, and Control Mechanisms
  • Control Theory
  • Formal Mathematical Models
  • Intelligent Controller Architecture
  • Multiagent Interlingua
  • Distributed Sensor Networks
  • Autonomous Robotics
  • Unmanned Underwater Vehicles (UUVs)
  • Unmanned Aerial Vehicles (UAVs)
  • Wireless Communication Devices and Protocols
  • Automated Image Understanding
  • Adaptive Network Control
  • Emergent Engineering

IC is committed to the advancement of distributed control and sensing technology through applied research, knowledge base development, studies and analyses. The Department has performed work for Department of Defense sponsors such as DARPA, ONR, and ARO.

IC has special expertise in heterogeneous systems operating in complex time-critical mission scenarios such as unmanned undersea vehicles (UUVs) and remote distributed sensor networks. Distributed, hierarchical control strategies allow collections of vehicles to self-organize into hierarchical teams capable of both coordinated collective behavior and robust autonomous activity. A language has been developed, called the Command Control & Communications Language (C3L), facilitating command and control-oriented communication among distributed autonomous systems. C3L also gives human mission planners the ability to script missions, command vehicles, and share behaviors. An intelligent controller architecture, the SAMON architecture, has been developed for UUVs and is being extended into other operational platforms such as unmanned aerial vehicles (UAVs) and industrial process control.

IC explores other forms of coordination as well such as semi-selfish multiagent cooperation. Often, in real-world scenarios, it is not possible for a central authority to have enough knowledge of the system to make proper control decisions. The semi-selfish control mechanism allows agents to make non-optimum local decisions in the form of peer-to-peer negotiation, which result in an emergent optimum global solution. In most cases, it would not have been possible to program the solution that emerges naturally from such a system.

Essential to intelligent control is the ability to sense and interpret the environment. In this area, the Department has developed methods to convert data streams from distributed sensor platforms to symbolic strings that can be compared to previous readings and exemplars for classification and semantic description. Work in this area includes video understanding, target tracking, and threat analysis. Extension of this work into the nanotechnology arena facilitates data acquisition and analysis from nanosensor grids acting as “artificial noses” in detecting toxic and hazardous substances.

IC has also developed adaptive distributed sensor networks involving re-configuration of, and collaborative sensor fusion from, diverse networked sensing elements and technologies.

Current research projects include:

  • Dynamic Planning
    • Multi Asset Dynamic planning for Early Localization of Mines

    The purpose of this project is to develop advanced S&T for making mine hunting missions more efficient, minimizing lag times and enabling greater search efficiencies through parallelism in searching. This is done in two ways: initial asset allocation to exploit parallelism and enable adaptive assets to give greater benefit and dynamic re-planning for cooperative mine hunting and the use of data fusion within and across sorties to find more mine-like objects and to find them earlier in the search, enabling follow-up mine identification assets to enter the operation sooner and to better utilize time and resources.

    First Priority Map  Second Priority Map 
    Third Priority Map  Fourth Priority Map.
    Evolving Priority Map of an MCM mission

    As part of this program we develop software algorithms and protocols to enable a group of mine-hunting assets to adaptively search for mines thus finding more mines faster than existing search protocols. This includes initial asset allocation and later reallocation, due to mission progress and in response to asset and environmental changes. The wide variety of assets available at different water depth levels and kill chain steps presents a complex, highly interdependent planning problem across multiple objectives. Re-planning exploits the solution space mapped out during planning, as well as models of asset adaptation.

    • Science of Autonomy (ONR funded)
  • Statistical Analysis Tools

    • NSF Geographic Hotspot Detection
    • Digital Governance and Hotspot Geoinformatics for Monitoring, Etiology, Early Warning  and Management