Building more trustworthy software

From autonomous vehicles to telecommunication satellites, understanding the potential safety, security and reliability issues and regulations surrounding rapidly emerging technologies is paramount in getting products to market quickly and efficiently.

As the number of Ottawa companies working in these disruptive fields multiplies, a pair of new researchers at University of Ottawa’s Faculty of Engineering are working to bring their expertise and research to the city’s tech community.

A driverless future

After working with various automotive and software companies in Europe, Prof. Shiva Nejati saw a growing opportunity to apply her research in Canada’s capital and its growing autonomous vehicle cluster.

Her primary focus is on improving the reliability and dependability of automated driving systems by developing software to test for any bugs or issues – a potentially perfect fit for Ottawa’s growing L5 autonomous vehicle testing facility. 

“As we make software to do what humans do, the information input to this software becomes a lot larger and harder to test,” she says. “We develop software tools that help engineers detect errors or false marks in their autonomous system more effectively, which also ensures confidence in that system.”

Nejati, who previously worked at the University of Luxembourg, developed her research and techniques using driver simulation technology to see how the various components of autonomous software react in different test cases. By running the driver simulation in bad weather, in traffic and with distractions, Nejati could, for example, foresee when the autonomous system would run a false positive – stopping the car without reason. This testing is imperative to the industry as autonomous technology cannot be tested on public roads – with pedestrians or other motorists – without potential safety risks, she says. 

By partnering with Delphi, an automotive technology manufacturer, as well as sensor development firm IEE to test the latest AV technology, Nejati says she realized the opportunity to deepen the relationships between industry leaders and academia.

“There are a lot of problems that the industry faces and it requires them to invest in research to be able to produce better systems,” she says. “The goal for us is to build the same kind of partnership programs or projects here, in particular in collaboration with Kanata North.”

AI-powered compliance automation

In the School of Electrical Engineering and Computer Science, another recent hire is also keen on using his expertise in early stage software development to help ensure companies are meeting the proper requirements and standards in their projects.

“Software systems are increasingly subject to laws and regulations in regard to safety, security, privacy and data protection,” says Prof. Mehrdad Sabetzadeh.

One way he’s tackled this challenge while helping businesses comply with such rules is by applying artificial intelligence software to transform jargon-heavy regulatory documents into easy-to-read digital versions to enable engineers to better understand the legal issues affecting their work. 

Sabetzadeh spent years working with law firms and banks in Europe ensuring their IT systems were up to code. From these partnerships, he says his team was able to develop new projects such as simulating the outcome of abolishing income-splitting tax policies in Luxembourg, which proved to be an efficient way of determining the impact of a government decision on the public. 

“I found this to be extremely useful and valuable, both for the industry side to see how they could benefit from all the work that researchers do and also for us to be able to ground what we do in practical needs,” he says. “Proximity to industry is key and we have a very lively ecosystem around regulations and regulatory compliance within both the public and private sectors in downtown Ottawa.”