Incorporation of System Support and Diagnostics Into Time-Triggered CAN Protocol

摘要

To avoid non-predictable communication delays in the distributed control systems, time-triggered protocols were introduced. By this approach it is required that all communication in the system is planed in advance. That represents a problem when fault-tolerance must be considered and reconfiguration of the system is needed. The paper proposes how support for fault-tolerance, system management and diagnostic could be incorporated into time-triggered CAN protocol. To reduce the complexity of dealing with defects in the system and to allow for short reconfiguration times in the case of faults, a distributed shared memory scheme is proposed as a middleware layer between the control application and communication subsystems. This way, design and implementation of safety related system issues is possible independent from, and in parallel with, the functional part.rnFurther, the fault-tolerant mechanisms provided by TTCAN were exploited that allow for centralized system monitoring and management.rnTo support testing and diagnostics from a single node in the distributed control system, additional communication protocol is transparently integrated within the time-triggered communication.rnThe presented solutions were implemented and tested on an experimental distributed hardware platform built as a part of the IFATIS (Fault Tolerant Control in Integrated System) project.