Five Major Software Challenges the Automotive Industry Must Conquer by 2030
March 15, 2022
This post is Part 2 of a 3-part series derived from TrustInSoft’s latest white paper, “On the Road to Zero-Bug Vehicles.” To obtain a FREE copy, click here.
Introduction
Software is set to drive the automotive industry for years to come. In our previous post, we looked at the key, software-driven trends expected to dominate the automotive sector over the next decade, and how increasing software complexity will likely become a make-or-break issue for many automotive OEMs and suppliers.
In this article, we’ll examine five more key challenges, including:
- The autonomous driving challenge
- The connectivity challenge
- The verification & validation challenge
- The software skills and toolchains challenge
- The ISO 26262 and ISO 21434 challenges
These will drive automotive software complexity and the need for more efficient and effective software verification methods in the sector.
The autonomous driving challenge
Much of the continued increase in automotive software complexity will arise from advancements in autonomous driving (AD), which is heavily dependent on software development.
“Growth in SW and sensors is largely driven by the development and adoption of AD, requiring advanced SW functionality (e.g., object detection and classification based on neural networks, raw data sensor fusion, and environmental modeling as well as algorithms for path planning), increased functional safety, and new sensor types (especially light detection and ranging, LiDAR),” says McKinsey.[i]
A large portion of this advanced AD SW functionality will be safety-critical. This software will face greater scrutiny to ensure safety and compliance with regulations like ISO 26262 and ISO 21343. This, in turn, will require efficient, effective defect detection and removal, as well as careful verification of strict adherence to requirements.
The connectivity challenge
Another major driver of software growth in the automotive industry will be the need for connectivity. Infotainment and other connected services, including the capability to update vehicle software components, will drive demand for greater cybersecurity in vehicles over the next ten years.
Therefore, any code related to connectivity must be free of software errors and other vulnerabilities. Such anomalies could allow hackers access to safety-critical functionality, exposing OEMs and suppliers to costly recalls, liability, and reputational damage.
As with safety-critical software, automotive players in the connectivity domain will need to adopt technologies and methodologies that provide guarantees that their software is impervious to cyberattacks.
The verification and validation challenge
McKinsey projects validation and verification will constitute approximately 29 percent (USD 24 billion) of the total automotive SW market by 2030.[i] This sizeable market share will be driven primarily by the need to verify safety-critical software (AD, advanced driver assistance systems (ADAS), security) and the operating systems deployed in the vehicles.
“The implication for automotive players is that they need to further invest in their capabilities to test and validate SW efficiently,” says McKinsey. “This translates to adoption of more specialized tools and restructuring of teams to foster efficient validation and collaboration with HW teams.” [ii]
In other words, along with the need to provide the guarantees of safety and security mentioned earlier, automotive OEMs and suppliers will be compelled to adopt tools that support modern methodologies like continuous integration and rapid debugging.
The software skills and toolchains challenge
Unfortunately, the vast majority of automotive companies are not yet equipped to meet the challenges being wrought by these emerging trends. Most face a challenge compared to the software market leaders in the acquisition and development of strategic software skills. Plus, as the demand for those skills mushrooms within the industry, automotive companies will be forced to compete more aggressively for development talent with the software industry at large.
To remain competitive, all players in the industry will have to cultivate in-house skills and strategic partnerships. They will need to build up competencies in coding, agile development, and specialty skills related to the ACES domains.
Finally, to unlock productivity, they will need to build up integrated toolchains that support state-of-the-art development, continuous integration, and test automation
“Overall, the introduction of a standardized, state-of-the-art development toolchain is a key enabler to unlock 30 to 40 percent of productivity potentials from automated testing and agile methods,” says McKinsey.
The ISO 26262 and ISO 21434 challenge
ISO 26262 is a functional safety standard used in the automotive industry, adapted from IEC 61508. It includes requirements for software development and design aimed at ensuring the safety of the vehicle at every stage in the vehicle’s lifecycle. As mentioned earlier, the volume of software having a direct impact on passenger safety is expected to grow enormously over the next decade and beyond. It’s important that automotive software complies with ISO 26262 and that automotive firms can provide guaranteed proof that their software does so.
ISO 21434, on the other hand, is a new standard (published in August 2021) that concerns the cybersecurity risk of the electronic systems of vehicles. As modern vehicles become more and more connected to the Internet, they also become more vulnerable to cybersecurity exploits.
ISO 21434 encompasses the entire vehicle production process, from design to software development. It introduces a structured approach to ensuring the security of vehicles. Developers of connected vehicles will need to demonstrate that their vehicle was designed with cybersecurity in mind, from the beginning to the end.
Although these standards cover a wide scope for developing safe and reliable code, exhaustive static analysis focuses on the most difficult part: identifying and helping developers eliminate undefined behaviors. It ensures that, whatever the inputs and the program conditions, the software will react in a deterministic way and is immune from security flaws.
Meeting these challenges…
The five challenges we’ve just looked at have one thing in common: they will all require more efficient methods and tools for software verification. In our next post, we’ll look at an emerging technology that meets this requirement. Most importantly, we’ll see how it can help automotive companies guarantee their software is safe, reliable, and secure.
This post is Part 2 of a 3-part series derived from TrustInSoft’s latest white paper, “On the Road to Zero-Bug Vehicles by 2030.”