John James

The state of computer-aided control system design (CACSD)

Introduction to a special issue consisting of a sample of the broad range of research activities, seeking to provide theoretical foundations and software engineering tools to help control engineers understand and overcome challenges in building computer-controlled devices. >

A hybrid systems approach to computer-aided control engineering

In this article we provide a tutorial-style overview of ideas for a new computer-aided control engineering (CACE) environment. We discuss these ideas through repetition of two themes. The first theme is that the CACE environments should support a thread from requirements development to implementation. While some CACE environments come close to this today, none provides the end-to-end thread we envision. The second theme is that the environment should support automatic generation of automata that simultaneously comply with discrete and continuous constraints. Success in this theme would reduce the need to conduct extensive simulations and build prototypes in order to deliver and incrementally change computer-controlled systems. An example shows how such an environment can assist in addressing a recurring problem in manufacturing systems: repair of production schedules required by changes in logical or continuum constraints on factory operation. We argue that a declarative, hybrid-systems approach to off-line design and online generation of reactive control is needed for achieving synchronization, scalability, integration, and incremental construction of large-scale, computer-controlled systems. >

Multiple agent hybrid control for manufacturing systems

We discuss how our generic hybrid control architecture, the mulitple agent hybrid control architecture (MAHCA), can be implemented to capture the complexity of the manufacturing process in an incremental fashion. We provide an example of how a MAHCA network of agents can be used to solve a planning, scheduling and control problem. MAHCA allows for the online reactive matching of current logic and continuum constraints from plant sensors with user-defined logic. The MAHCA architecture supports incremental expansion of new components with greatly reduced requirements for expensive experimentation validation.

A New Approach to Generating Finite-State Control Programs for Hybrid Systems

Abstract We provide an abstract of a new approach for generating finite-state control programs for hybrid systems. We illustrate these ideas in the context of a cost-based Computer-Aided Control Engineering (CACE) environment. We discuss our methodology for ensuring that the environment should support automatic generation of automata which simultaneously comply with logic and evolution constraints. We argue that a declarative, hybrid systems approach to off-line design and online generation of reactive control automata is needed for achieving synchronization, scalability, integration, and incremental construction of large-scale, computer-controlled systems. The methodology supports a multiple-game strategy for distributed systems optimization whereby different components of the systems can employ different local optimization strategies. We conclude by comparing this constructive approach for on-line selection of actions that are near-optimal with respect to economic cost to the simulation-based approach for heuristic and exhaustive trade-off analysis.

Research needs in economic analysis of control design projects

Outlines a new methodology for generating economic plans and budgets for design projects. This approach is based on an extension of the Kohn-Nerode hybrid system models and algorithms. Given a hybrid system model of a distributed system and a performance specification for that system (expressed as a relaxed variational problem of minimizing the integral of a Lagrangian over a trajectory on a suitable manifold), the authors extract a nearly optimal chattering control program which forces the system to obey its performance specification.

Hybrid systems and distributed interactive simulations [military]

The Army has set a goal of enhancing battlefield effectiveness by fielding a digital division by 1998 and has started a sequence of field exercises to investigate how new applications of digital technology will affect military operations. These transition activities are part of preparing the Army to fight third-wave warfare-information-age warfare. Distributed interactive simulation (DIS) is a key technology in determining and analyzing alternatives for digitizing the battlefield. Advanced information systems can create a shared situational awareness (visualization) of the battlefield. Realization of shared awareness will support faster-paced operations through real-time force synchronization. Given the rapid changes in computer capabilities, communications bandwidth, and software complexity, it has been unclear what the mid- and far-term technical opportunities and challenges are in applying results from the ongoing information systems revolution to improve battlefield effectiveness. Substantial improvements in current DIS technologies are needed to enable professionals at widely distributed sites to interact simultaneously through simulators, simulations, and deployed systems in a common joint synthetic operational environment. In this article we discuss how one of the foundational technologies supported by the US Army Research Office, hybrid systems technology, can support closing some of the DIS technological gaps.<<ETX>>