This paper discusses ongoing research on automatically configurable distributed control systems. We are investigating the use of networks of relatively small controllers that contain intelligence and communication capabilities to perform factory automation. These systems are expected to be more robust, adaptable and flexible than conventional designs. To achieve these benefits, fundamental changes in the underlying control system architecture are needed. The system architecture is divided into two portions: the control infrastructure and the control algorithms. Requirements for the control infrastructure are described that implement basic services to support distributed control, including the ability for a new control node to announce its presence and capabilities to the system without disrupting ongoing activities. The control algorithms for this system must be designed with certain properties: autonomy and cooperativeness; that are able to respond to changes in the physical plant and its usage. A laboratory for studying autoconjigurable control systems is under development using prototype distributed controllers on a CAN-based network. Two experiments are described that demonstrate autoconfiguration capability. The potential applications and the future directions of this research are discussed.

Systems in which many data sources are multiplexed over a single communication channel are considered. Data from all the sources are generated in fixed-length packets and are stored in a common buffer with finite capacity. Packets that overflowed or were removed from the buffer prior to transmission are lost. The system performance measure is the set of packet loss probabilities associated with the sources. Queueing disciplines vary depending on the stringency of prioritization and the utilization of the system resources. The set of all possible performances is characterized as we span the set of all queueing disciplines. Whether a given performance is possible can be deduced. Strategies that achieve the minimum overall loss probability are identified. The extreme disciplines are specified, and their performances are calculable by means of a given algorithm.

A class of slotted ALOHA dynamic control strategies is considered. These strategies are simple to implement and can yield lossless and stable operation for arbitrarily large user populations with aggregate arrival rates below e-1 packets/slot. An ergodicity analysis is given that provides conditions on the system parameters, such that any specified set of control parameters that satisfies the given conditions is guaranteed to yield stable performance. The system state is modelled as a two-dimensional Markov chain that incorporates the backlog (the number of packets awaiting retransmission) and the estimate of the backlog. The geometrical concepts are illustrated by figures corresponding to an example case. Simulation results are presented that compare alternative control schemes.

This paper describes an ALOHA random access protocol for fading communications channels. A two-state Markov model is used for the channel error process to account for the channel fading memory. The ALOHA protocol is modified to send multiple contiguous copies of a message at each transmission attempt. Both pure and slotted ALOHA channels are considered. The analysis is applicable to fading environments where the channel memory is .short compared to the propagation delay. It is shown that smaller delay may be achieved using replications and, in noisy conditions, can also improve throughput.