Ph.D. Candidate Qi Zhu

From the EECS Department at University of California, Berkeley, CA, USA.

Authors

Abhijit Davare, Qi Zhu, Marco Di Natale, Claudio Pinello, Sri Kanajan, Alberto Sangiovanni-Vincentelli - Univ. of California, Berkeley, CA

Abstract

The complexity and physical distribution of modern active-safety automotive applications requires the use of distributed architectures. These architectures consist of multiple electronic control units (ECUs) connected with standardized buses. When deploying the applications onto the distributed architectures, we need to decide the allocation of functional blocks to tasks and tasks to ECUs, the packing of signals to messages and the allocation of messages to buses, as well as the priorities, activation models and periods of tasks and messages.

In this work, we focus on the period assignment of tasks and messages in Controller Area Network (CAN) based systems. Given the allocation and priority assignment of tasks and messages, our approach automatically assigns periods for all tasks and messages in order to satisfy the end-to-end latency constraints in the system. Based on worst case analysis, the problem is formulated as mixed-integer geometric programming (MIGP). To efficiently solve the problem, MIGP is further approximated as a geometric programming (GP) formulation, and the approximation error between the two formulations is reduced with an iterative procedure.