Three challenges in performing reliable software maintenance on automotive control systems

Modern automotive control systems are incredibly intricate hybrids of software components and hardware mechanisms. Gaining a sufficient understanding of such a system in order to perform reliable maintenance with predictable results can be quite difficult. This paper discusses three specific challenges related to this task: (1) having tools that provide sufficiently robust analysis to support the decision-making process of engineers performing the maintenance; (2) managing the complexity of mixed hardware/software analysis (by both the user and the tool); and (3) constructing tools that are suitable for adoption by the practitioner community so that they can provide an effective automated aid. These three challenges were identified as part of our own experience in working on projects in this application area.