On the Development of Optimal Hardware and Software Architectures for Propulsion Domain Control

Abstract

<div class="section abstract"><div class="htmlview paragraph">The modern automobile consists of a series of functional systems which implement the features of the vehicle and which are often connected via electronic networks. The function of each block in a system is often controlled or implemented by an electronic processor. In practice the existing architectures of both the network and processors is a consequence of the gradual introduction of electronics to a previously mechanical implementation and therefore tend to favour a flat network of cooperating peers and processors with hardware functions that are highly specific to the function required. New architectural approaches are now being evaluated to remove the limitations brought by the existing approach and in particular to address the changes in the propulsion system caused by the introduction of electric drive components. In this paper we describe the historical evolution of vehicle architectures towards domain control topologies and how this evolution has affected the functionality that is required in each functional block. From this starting point we consider the specific requirements that apply to the propulsion domain and consider the effect that these have on the software and hardware implementation of new modules. Finally, the paper considers how the internal architecture of a microprocessor can be developed to provide an optimal environment for software features and development flow. Elements considered in this discussion include leveraging existing software, the impact of electrification, scalable processing capability, facilitation of independent development environments and the implementation features that allow assurance of the safety, security and functional needs of modules operating in the propulsion domain.</div></div>