Feature: Software and tools


soſtware compliance with MISRA, AUTOSAR C++ 14, or similar coding standards is essential to product certification and market admission, which makes it a complex and resource-intensive endeavour. Traditional static analysis plays a pivotal role in validating code


for safety compliance, by systematically scanning for defects and deviations from established coding guidelines. However, this process is fraught with challenges: It oſten requires extensive manual review to siſt through false positives, demands specialised expertise, and necessitates iterative refinements to align with evolving standards. Te substantial effort and time required for these compliance activities can impede rapid development and increase overall costs. With the advent of AI, static analysis can be transformed. Te


AI-augmented approach uses large language models (LLMs) to improve code understanding and accelerate the identification and remediation of compliance issues. By integrating AI capabilities, such methods enhance the precision of issue detection and facilitate quicker code fixes, offering a more intuitive and context-aware mechanism compared to conventional practices.


Accelerating MISRA


compliance using AI- augmented static analysis By Arthur Hicken, Evangelist, Parasoft


well above 100 million in a single vehicle; see Figure 1. Soſtware enhances functionality, safety as well as user experience,


A


but at the same time it creates a larger burden on cars. Any soſtware used in vehicles must be safe to use and comply to strict automotive standards, such as ISO 26262. For safety-related products, embedded


34 March 2026 www.electronicsworld.co.uk


utomotive applications like advanced driver assistance systems (ADAS), infotainment, navigation, soſtware-defined vehicles, autonomous driving and safety have exponentially increased the soſtware in vehicles over the past few decades. Te code has grown from under 100,000 lines to


Safety and reliability first Te goal of the ISO 26262 standard is to help make safe and reliable vehicles. To test vehicle soſtware, the standard recommends using static code analysis with guidelines like MISRA to identify and prevent issues from entering production. Te benefits of static analysis include early detection of errors, reduction in ambiguous behaviour and improved reliability. However, whilst a critical component for ensuring soſtware safety


and compliance, static code analysis also comes with challenges and bigger costs. One of the primary issues is the overwhelming volume of results generated by these tools. Oſten, a single run of a static analysis tool can produce thousands of warnings and errors. Many are false positives or minor issues that do not directly affect system safety or are of unknown risk and importance. However, it is this flood of information that makes it difficult


for engineers to discern which problems are truly critical, thereby increasing the time and effort required to triage and prioritise them. In addition to managing a large number of results, understanding the root cause of identified issues and determining the appropriate fixes is oſten complex. Te analysis outputs may not provide enough context or actionable guidance, requiring deep domain expertise and manual code review to interpret the findings correctly. Tis process is resource intensive and can lead to increased development costs, as resolving these issues may necessitate extensive re-work and additional testing. ML-based solutions have proven very effective in automatic


violations pre-classification, clusterings and prioritisation. AI models help developers focus on the critical items and easily identify those that can be safely ignored or handled later.


Experiment to determine Many are seeing interesting results using LLM to generate code and tests, but here at Parasoſt we wanted to see how LLMs can improve the workflow for static analysis in a compliance industry like


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32  |  Page 33  |  Page 34  |  Page 35  |  Page 36  |  Page 37  |  Page 38  |  Page 39  |  Page 40  |  Page 41  |  Page 42  |  Page 43  |  Page 44  |  Page 45  |  Page 46  |  Page 47  |  Page 48