6:00 p.m. The Albany Club - 91 King Street East, Toronto, Ontario
Speaker: Dr. Akramul Azim, Assistant Professor in software engineering at Ontario Tech University
We want embedded software to be verifiable. However, if all the interactions of the software with its associated system are not known, we may not be able to provide evidence of predictable and deterministic system behavior. The evidence is crucial for any safety-critical functionality, such as autonomous driving. The behavior of a system represents not only functionality but also architectural and execution properties such as resource consumption and timing. Non-determinism might arise for unpredictable input or output, which makes a system uncontrollable.
Automated verification is useful for verifying properties, even for large programs. However, validating complex functions such as embedded machine learning algorithms often tend to be difficult because of the large software size of the current embedded system (size can be measured in lines of code, number of system inputs and outputs, and/or number of classes/modules). The code size also constrains the functional safety certification process, such as IEC 61508 (generic), ISO 26262 (automotive), and IEC 62304 (medical). This talk will cover some guidelines to test software systems ranging from automotive to medical device software to accelerate the functional safety certification process.Register