Visiting Italian graduate student Alberto Zuccollo aids Kettering professor with FlexCAN research project.
Alberto Zuccollo, a graduate student currently completing his Ph.D. in safety critical systems at the University of Padova in Padova, Italy, recently completed a five-month visiting scholar appointment at Kettering for a project headed by Electrical and Computer Engineering Professor Juan Pimentel.
Over the last several years, Pimentel has worked in the area of safety critical systems for automotive applications. Specifically, his research into controller area networks (CAN) based mechanisms may lead to the creation of more dependable safety systems in vehicles. Dependable systems work in conjunction with other electronic systems inside vehicles for operations, such as braking and steering. But according to Pimentel, no simple and inexpensive dependable systems currently exist on the market, which means automotive companies must spend an exorbitant amount to develop these systems from scratch. This situation led Pimentel to create an architecture and protocol titled FlexCAN, which he based on controller area networks suitable for highly dependable systems. Pimentel also integrated SafeCAN into his system, which deals with error detection and fault management of electronic communication nodes that interact and communicate with other system components inside a vehicle.
In other words, Pimentel's high dependable system is cheaper and simpler for automakers to incorporate into vehicles while sustaining the high level of research necessary to create systems to protect drivers and occupants in emergencies. His architecture and protocol rely on commercially over the shelf components (COTS), thus making the system cost effective.
Zuccollo's work on this project focused on an in-depth analysis of dependable mechanisms of the FlexCAN architecture, including research on new mechanisms to make FlexCAN more tolerant to an array of hardware and software faults. "One specific mechanism that he suggested, implemented and verified experimentally involves a 'bus-guardian' that protects against babbling-idiot faults that can occur in safety critical applications," Pimentel explained. These "babbling-idiot" faults happen when one of the electronic nodes in the vehicle continuously denies communication between other nodes, thereby making communication between electronic systems in the vehicle difficult.
Zuccollo performed a theoretical study to determine how many nodes might compensate for one node mired in a "babbling-idiot" situation. Currently, when one node does not respond correctly in this system, the other nodes receive no communication and thus do not operate correctly. Zuccollo focused on identifying a way other nodes could detect when another is inoperable, and thus use emergency routines to compensate for the problem node. An example would be a brake system on a car. If a communication node fails, other nodes would detect the failure and compensate for the lost node by slowing the car down.
This, according to Zuccollo, would make vehicles much safer. However, safety critical system networks must be "perfect," he said. "The issue currently regarding this idea is that the electronic components must be very reliable. There is a tiny threshold for failure," he added.
One day Zuccollo believes vehicles will achieve this fail-safe mark. Until then, he will continue his Ph.D. work in Italy. He said he enjoyed his stay in the U.S. and at Kettering, and found he learned many things working with Professor Pimentel. Zuccollo noted that Pimentel was very helpful in overseeing research and in allowing the Ph.D. candidate to be free in expressing his ideas.
While here, Zuccollo visited Chicago, Detroit and other areas of the U.S. He found coffee in this country to be a little "weak," he said. "In Italy, we drink very strong coffee," he added. One of the first things he said he would do uponhisreturn home was to buy a cup of espresso.
Written by Gary J. Erwin