CAN-based Distributed Systems Research Articles

Collaborative distributed sun-tracking control system for building integration with minimal plant area and maximum energy-conversion efficiency

This paper presents a collaborative distributed sun-tracking control system for a novel Fixed Mirror Solar Concentrator (FMSC) structure, which increases the energy-conversion efficiency of the FMSC and reduces the space between solar collectors units, a positive aspect for in-building integration. The improved FMSC uses solar concentration collectors suited for mid-range thermal applications (90–200 °C) and is designed for easy installation in buildings because of its relatively small extension. The proposed solar orientation system (ORSYS) relies on a two-step algorithm to increase the energy captured by the receiver, which provides tolerance to common logical and mechanical errors in the estimation of the receiver position. ORSYS is implemented as a CAN-based distributed system, extended with web-interface features for supervision and configuration of the overall system. ORSYS also includes a coordination algorithm that allows adjacent collectors to share the physical space between them, thus reducing the total plant area. Experimental evaluation has been performed using an industrial-scale solar collector prototype, showing its feasibility and efficiency in terms of energy conversion in real environments.

A Total Ordering Scheme for Real-Time Multicasts in CAN

The Controller Area Network (CAN) is a broadcast medium providing advanced features, which make it suitable for many real-time applications. Common application layer protocols designed for CAN (e.g. CAL, SDS, DeviceNet) exploit these features in order to provide reliable real-time communication. However, they do not provide a consistent global message ordering in certain fault situations, nor do they consider temporal properties of the application.This paper presents a multicast protocol which supports timely delivery of messages and guarantees atomic order, under anticipated fault conditions. In order to support causal ordering of events, the paper provides a discussion on establishing Lamport's precedence relation between events by appropriate usage of real-time multicasts.In a CAN-based distributed system, the restricted communication bandwidth constitutes a serious bottle-neck. Therefore an important feature of this multicast protocol is to achieve optimal protocol termination time while requiring minimum communication overhead.