Fault-tolerant distributed processing architectures for automotive applications

Future automobiles will have various features that will require a structured design of the whole automotive multiplexing system. Fault tolerant techniques are used in computer systems to keep them running in case of one or more processor failures. The concept of fault tolerance is well known in many applications like airplanes, industrial automation and military. However, the question of fault tolerant design in automotive multiplexing has drawn very little attention. This dissertation presents fault tolerant distributed processing architectures for automotive applications in which fault tolerant schemes have been suggested to meet the cost constraints associated with the auto industry. Availability and reliability models have been developed to evaluate the performance of the proposed fault tolerant distributed processing architectures.; Data reduction techniques can be employed in automotive multiplexing systems to improve the information exchange rate among various intelligent modules. In the cited references, some off-the-shelf data reduction algorithms have been considered for automotive multiplexing. However, their applications have been limited to the text data class only. This dissertation proposes a data reduction algorithm, which can be generally applied to all data classes found in automotive multiplexing. A discrete event simulation model has been developed for the performance analysis of the proposed algorithm. Although this algorithm has been developed to fit in the automotive environment, however, it can also be applied to other non-automotive real time applications where extensive information exchange is performed among control modules via a multiplexing bus.