A Deterministic Algorithm for Distributed Processing in a Vehicle Network

Paper #:
  • 2013-01-1181

Published:
  • 2013-04-08
DOI:
  • 10.4271/2013-01-1181
Citation:
Sundaram, S., Malagi, S., and Ahmed, S., "A Deterministic Algorithm for Distributed Processing in a Vehicle Network," SAE Technical Paper 2013-01-1181, 2013, https://doi.org/10.4271/2013-01-1181.
Pages:
10
Abstract:
Present day automobiles make use of a large number of Electronic control Units (ECU), for a varied purposes such as engine control, power steering control, transmission control etc. Each of these ECU's has its own processors, the selection of which is made on the basis of the particular task to be handled. All the more, these processors are interconnected by the use of one or more communication protocol such as CAN, LIN etc. This gives us a platform for the implementation of distributed processing in the system. Implementation of distributed processing has a number of advantages such as increase in processing power, reduction in wire complexity, reuse of same hardware module for different tasks, decrease in the processor size required as compared to when the entire processing was to be carried out by an individual processor and better and more complex problem solving tools can now be implemented.This paper describes a deterministic algorithm using LIN bus for a distributed processing. The algorithm talks of an approach where each of the slave processors provide their throughput to the master processor at predefined intervals, which in turn divides up a major processing task in to sub tasks on the basis of basic arithmetic operations, and distributes the same to one or more processors, starting from the one with the lowest throughput at that moment. Simulation of such a system interconnected with LIN was carried out and advantages are described. Finally, it is also shown that downsizing of processor in ECUs is possible using this algorithm.
Access
Now
SAE MOBILUS Subscriber? You may already have access.
Buy
Select
Price
List
Download
$27.00
Mail
$27.00
Members save up to 40% off list price.
Share
HTML for Linking to Page
Page URL

Related Items

Article
2016-11-15