Navigating SDVs 

Applying aviation’s proven strategies to the development of software-defined vehicles

Digital technology is transforming the way we move. Gone are the days when vehicles were merely a means of getting from one place to another. Today, automotive customers are looking for smart, sustainable, over-the-air upgradable connected vehicles that offer enhanced convenience and experiences. Over 90% of automotive innovation is now powered by software. This has led to a paradigm shift in the automotive industry, as manufacturers shift their focus from hardware-focused to software-defined vehicles.  

Fortunately, the automotive industry doesn’t have to travel unaided. The aviation sector—a forerunner in technology creation, adaptation, and commercialization within the transportation industry—has already navigated this journey. Despite facing challenges in transformation akin to those of SDVs, the aviation industry has maintained strong safety records. Hence, the automotive sector can gain insights by understanding how the aviation sector has adapted to these challenges.

Excessive code

Lesson 1: Understand that more code is not always better.

Today’s cars have around 100 million lines of code in their software. This number is staggering, considering that a modern commercial aircraft runs its complete avionics, control, and online support systems with less than 10 million lines of code.

Despite being less expensive (0.04%) and carrying a lesser payload (0.4%) than a commercial aircraft, a car has about 10 times more lines of code. This huge amount of code in vehicles is necessitated by the large number of car variants and user configurations, and are often related to navigation, infotainment, and comfort. 

As per the Society of Automotive Engineers (SAE), there are six levels of driving automation: level 0 (fully manual driving) to level 5 (fully autonomous driving). The automotive world is progressing through these levels. The number of lines of code in the software of these cars is likely to increase too, possibly to as much as 300 million by 2030.

With decades of research, the aviation industry has leveraged advanced technology concepts such as service-oriented architecture to optimize the size of its software. Learning from the aviation industry, SDV manufacturers should adopt smarter ways to reduce the lines of code before they reach unmanageable levels.

Automation bias

Lesson 2: Embrace a human-centric approach to automation deployment.

Decades of innovation, design, development, testing, validation, and strict regulations have contributed to high-level automation in the aviation industry. Making autonomous SDVs, or SDVs with autonomous technologies, will be no mean feat either. During the design stage, the SDV industry should keep in mind what the aviation industry has learned about human response to automation.

One significant learning involves automation bias, which is the tendency for humans to favor suggestions from automated decision-making systems while ignoring contradictory cues from non-automated systems. 

Pilots are trained rigorously for approximately 1,500 hours to reduce the automation bias and equip them with the skills required to handle safety incidents. Even though automobiles are more widely available and require comparatively less training, we should expect that automation in SDVs will lead to increasing automation bias. Drivers may tend to become complacent, and their driving skills may degrade over time. This could prove to be unsafe in critical situations where the vehicle gives control back to the driver. 

The airworthiness of every aspect of an airplane from tangible components to software undergoes robust testing. Crews operate the aircraft in meticulously controlled environments. In contrast, drivers must navigate unpredictable scenarios, such as weather, road conditions, and erratic behavior of other drivers and pedestrians.

Taking a cue from aviation, SDV manufacturers might aim for full automation in controlled environments, such as on highways, on roads with good infrastructure, and on days with clear weather. However, in crowded and chaotic areas, SDVs need to work at lower levels of automation, to reduce automation bias and improve safety.

Cybersecurity threats

Lesson 3: Emulate how the aviation industry has built-in security.

As the amount of software and IoT content in automobiles increases, so too does the automotive industry become more vulnerable to cybersecurity threats. An unprotected SDV ecosystem exposes itself to vulnerabilities that could provide ways for unethical players to access, enter, or control vehicles remotely. The number of cyberattacks on cars has soared to 225% in the last three years.

Fortunately, there are a few best practices that the SDV industry can adopt from the aviation world. Those include:

• Adopt a zero-trust approach: SDV manufacturers should design their cybersecurity systems with an underlying assumption that communication networks are inherently compromised, creating a need to focus on validation at every stage of digital interaction.

• Make security a priority in the design phase: The three traditional automotive design pillars are cost, quality, and schedule. SDV manufacturers should now consider security as the fourth pillar. 

• Consider ‘ethical hacking’: SDV manufacturers can hire security experts or ethical hackers to conduct penetration testing and detect vulnerabilities at an early stage before products go to market.

• Take an ecosystem approach: Analogous to the aviation industry, closely-knit security ecosystem partners and partnerships (including governments, cab aggregators, regulatory bodies, and certification agencies) can work together to create robust cybersecurity systems for SDVs.

Big data conundrum

Lesson 4: Get ready to manage big data

In aviation, data is collected from flights, fliers, third parties such as ground support in airports and aircraft control rooms, and social platforms. This data is analyzed to predict maintenance needs, improve the passenger experience, optimize route planning, maximize fleet optimization, and set effective pricing strategies. From an automotive perspective, despite an ever-increasing number of onboard sensors and devices, the quantum of data has not yet reached the levels of the aviation sector.

SDV manufacturers should look holistically at the data available to them, which could include data from onboard and offboard sensors, passengers, third parties such as cab aggregators and weather reports, and social platforms. Intelligent techniques for collecting and analyzing these varied data sets is the ask of the day. As in aviation, big data analytics can be applied to the aggregated data in SDVs for a multitude of applications. Examples include:

• Supply chain: Efficient supply chain management could allow organizations to pick the best components in the market, increasing brand value, revenue, and profitability.

• Vehicle maintenance: Smart predictive analysis could improve vehicle health and reduce maintenance costs.

• Personalization: By incorporating weather, road, and traffic conditions, SDVs could create a customized and optimized driving experience. 

• Seamless experience: A connected ecosystem could allow the SDVs to communicate bi-directionally with other systems to monitor vehicle health and offer enhanced safety and an enriched commuting experience.

The way forward

Leverage aviation’s expertise to drive SDV progress

The prospect of realizing a perpetually upgradable vehicle in tandem with increased levels of comfort, safety, security, and a never-seen-before user experience makes SDVs a promising proposition. It could be the automotive industry’s best bet to meet ever-rising customer expectations.

And fortunately for the industry, aviation has already witnessed most of the challenges it is going through now in one form or another and has managed to offer mankind one of the safest modes of transportation. By taking cues from the learnings in aviation, SDVs can set themselves up to be the future of transportation.