MDH research improves safety in cars and medical equipment

Dec 03, 2014 | Research/Cooperation

Ensuring that a computer’s embedded systems select correct information while maintaining their critical function is a big challenge. Now Leo Hatvani, doctoral student at Mälardalen University (MDH), is presenting a new model to improve safety. In his licentiate thesis he proposes a method by which the computer can select what tasks are to be performed and when, in a way that ensures that everything is done within the given time frame.

- Modern computers process a series of tasks. Just as a person when ticking off the items on a to-do list when given a complex task, a computer needs to perform a number of tasks from a list, in order to complete a larger task that a user has requested, Leo Hatvani explains.

Computers that help us run other machines such as cars, industrial production lines, or medical equipment, often need to carry out the tasks before a certain time (deadline).

Drastic consequences for a missed deadline

A missed deadline can have drastic consequences, and therefore choosing which task to complete next is of the utmost importance. For example, if the computer that regulates the braking system in a car takes just a couple of seconds longer to translate the push of a pedal into a brake plate activation, the car can already be involved in an accident.

There are a lot of existing techniques that aim to ensure that a computer task is completed within its time frame. Unfortunately these techniques are not always enough, since industry is constantly adding new features that need to be completed before the deadline. This problem can often be solved by using a faster computer, but in certain cases this is not an option, and in other cases not even the fastest available computer can perform everything we ask of it.

Adapt the system according to the circumstances

- But all functionalities are not always critical. In such cases we can build computers that can select their actions depending on the circumstances. For example the computer can choose between reducing and stopping non-critical functions to ensure that the critical functionality is completed in time. Thus the system is adapted to the circumstances.

In his licentiate thesis, Leo Hatvani presents an approach to address this challenge. He presents a model to describe specified circumstances, and proposes a method for the computer to choose what tasks to execute and when, such that the required timing constraints are enforced.

This model is valuable in that it allows the user to determine whether it will always maintain its critical functionality. 

About Leo Hatvani

Leo Hatvani grew up in Sarajevo in Bosnia and Herzegovina. His childhood came to be marked by the war which went on from 1992-1996, but in his teens, after the war, he began to compete in programming/computer science, in which he won several competitions.

- It was these competitions that aroused my interest in researching into computer science, Leo Hatvani tells us.

After graduating from the Faculty of Electrical Engineering in Sarajevo in January 2009 he was recommended to apply for a doctoral studentship in Computer Science (Embedded Systems) at MDH. This doctoral studentship became a natural continuation of the studies he pursued in Sarajevo. Leo Hatvani commenced his doctoral studies at MDH in October 2009.

On Friday 28 November Leo Hatvani defended his Licentiate thesis “Formal Verification of Adaptive Real-Time Systems by Extending Task Automata” at MDH in Västerås.