Keeping autonomous vehicles on track

Jeremy Johnson presents three steps to ensure autonomous vehicle development remains nimble and innovative 

Never before have so many automotive engineers been tasked with bringing increasingly complex machines to market as they have with autonomous vehicles (AVs). Not to mention, it’s not simply a game of speed – elevating only the companies which manufacture products quickly – but a matter of those organisations taking these steps, while pushing the envelope on innovation and prioritising consumer safety. 

While important advancements are being made daily to bring fully autonomous vehicles into commercial availability, we are still a ways off from seeing the Level 5 autonomy we hope for. We’ve witnessed this in the significant challenges and shortcomings reported in the news in recent years, even amongst some of the biggest names in autonomous vehicle production. There are, however, a few actions engineers can take to stay nimble, innovative, and reduce the number of safety-critical mistakes throughout the development process. 

Focus on your core business: advancing technology

This is particularly important for start-ups or companies looking to apply their technology to the automotive space for the first time. If you’ve heard the saying “don’t reinvent the wheel,” it comes into play when considering what processes and procedures for meeting industry best practices must be in place. Seek a consulting or technology partner that can enable your business to continuously practice requirements, risk, and test management in alignment with market standards such as ISO 26262 and ISO/PAS 21448. By starting with a proven framework that can be applied and moulded to a particular business, engineers can focus their innovation and organisational energy on delivering new technology to customers. 

This focus is noticed in how Tesla drives its business, where the organisation leans more heavily into internal development to drive technology advancement and differentiation in the marketplace. Audi, which attempted to meet Level 3 autonomy with its 2019 A8, sought outside suppliers such as Aptiv, Intel, Infineon and NVIDIA– a different approach to Tesla. Although Audi ultimately pulled back stating the car was too far into the lifecycle, it was realised through properly executed business and safety procedures. 

Support collaboration internally and externally

Rapid innovation requires tight collaboration, often occurring across various hardware and software teams, and increasingly with partners or traditional competitors. Whether a formal joint venture or targeted collaboration around specific technology development, this “co-competition” has become more common as companies look to drive innovation in AVs forward. 

Ensuring the tools and processes to enable this collaboration are in place, and capturing the critical output that comes with it, will guarantee that R&D efforts move quickly while maintaining strong focus on verification and validation. It’s especially important engineering teams get out of their silos and work with adversaries on this front, because in the U.S. in particular, there’s no mandatory compliance enforced by the government to follow standards such as ISO 26262 or SOTIF. 

There have been promising signs of collaboration by some automotive companies to exchange learned information during AV development. One example is the collaboration among Aptiv, Audi, Baidu, BMW, Continental, Daimler, Fiat Chrysler Automobiles, Here, Infineon, Intel, and Volkswagen to develop a whitepaper, “Safety First for Automated Driving,” describing a potential framework for the development, testing, and validation of safe AVs. 

Maintain traceability of requirements, tests, and risks    

Developing complex, safety-critical systems that marry software and hardware requires a great deal of rigor and planning. Keeping track of each step of the development process with cumbersome documents and spreadsheets greatly hinders engineers’ ability to remain agile. By ensuring all of these steps are tightly managed, integrated into other product lifecycle phases and available for flexible reporting will enable organisations to innovate quickly. This also allows for prioritisation of safety and compliance and the ability to rapidly adapt as the regulatory landscape continues to evolve.. 

Most countries do not yet have specific regulations that govern autonomous vehicles, leading to uncertainty around requirements, reporting,and future regulatory compliance.  As previously mentioned, in the U.S. there is also an absence of tightly defined regulations which means states could implement differing standards that require specific nuances in technology and regulatory compliance. There’s also basic differences in infrastructure that makes development and safety a challenge – such as variability in road surfaces, lane markings, and signage.  

The key for teams in this fluid environment is to remain close to standard development and regulatory agencies, as well as supply chain partners, to define and influence these regulations. And they must be prepared to show traceability of requirements, risk, and testing information in multiple formats to support the various potential points of oversight – be it downstream customers or regulatory auditors. 

Ultimately, we’re likely a decade or more away from commercial availability of a Level 5 autonomous vehicle. As we inch closer, it’s vital engineering teams take advantage of the modern systems management tools at their disposal in order to get it right now – before it’s too late.

Jeremy Johnson is VP of product management at Jama Software

Recent Issues