With electric vehicles yet to hit critical mass worldwide, the road to level 5 AV ubiquity lies beyond the horizon yet.

Much of the appeal of autonomous vehicles (AVs) is their potential to make the roads safer by reducing traffic buildups, crashes, injuries and fatalities. However, before level 5 autonomy can be achieved, it is vital to develop trust in autonomous driving, but this is a complex problem with no margin for error.

Adding to the challenge, the ecosystem involved in building and developing AVs is more complex and distributed than the established auto industry.

To that end, various pilots are underway in Singapore and Malaysia, including self-driving buses and taxis within a dedicated testing area. In Malaysia, an AV pilot on a 5G highway is taking place, and Japan has recently allowed nearly autonomous vehicles on public roads in a limited capacity. All of these initiatives are helping accelerate the drive for more autonomy.

Hwee Yng Yeo, Automotive & Energy Solutions Manager, Keysight Technologies

Strategies for overcoming AV hurdles

Unless trust is built within the industry and society by overcoming the web of safety concerns, the drive toward fully autonomous vehicles will remain in the slow lane. What are some of the current constraints and coping strategies?

    • The industry needs to collaborate to create a shared data framework supporting more intelligent modeling to help address the complex safety concerns with driverless cars. To address that, vehicles today have an increasingly wider array of sensors and in-vehicle networks, and all of the data collected is helping make AVs more predictable. Digital twin solutions enable simulations to evaluate the performance in realistic scenarios, fostering an environment where AVs can make accurate, predictable and safe real-time decisions. This approach ensures that AVs meet safety and quality standards before hitting the assembly line. In addition to saving money and time, this is the only realistic way to test and evaluate each model under the multitude of environments and situations the vehicle may be in.
    • From traversing snow-covered tarmac on the Hokkaido Expressway to the congested streets of Manila, each potential permutation requires evaluation before autonomous driving becomes ubiquitous. This is no small feat. AV technology’s sheer complexity is part of the problem and the solution of accelerating market realization. Software-defined vehicles are an essential destination on the road to full autonomy. And the technology race in vehicles is creating vast data streams that can aid development by improving deterministic decision-making. As intelligent software becomes commonplace in vehicle and driver assistance systems, this will help minimize human error — creating safer vehicles. However, more technology is required to test that the AV performs exactly as expected, meeting the necessary safety standards under every scenario.
    • The wider attack surface and vulnerabilities of making AVs more connected and safe invites hackers and fraud. The consequences range from the loss of financial or personal data or, in the worst-case scenario, life threatening hazards. Therefore, ensuring every possible loophole is identified and addressed in the design phase is critical to delivering a safe and secure autonomous driving experience. In addition, once the car is on the road, every software or system update must be subject to rigorous testing to ensure it does not have security flaws.
    • With the majority of accidents a result of human error, AV legislation has the opportunity to help usher in an era of safer driving. However, the entire ecosystem and supply chain must be subject to these edicts to build trust. Advancing AVs requires collaboration and implementing various standards — for example, what is the minimal viable, safe vehicle? Work is underway to ensure autonomous driving is subject to regulations with initiatives, including the UN’s World Forum for Harmonization of Vehicle Regulations. In Asia, level 4 autonomous driving or fully automated driving is now permitted under certain conditions in Japan. Singapore was the first country to put in place regulations by amending its Road Traffic Act to recognize that vehicles do not require human drivers. As regulations expand, they will help address all safety concerns and build trust.

Developing strategies to address the areas outlined will make the journey to more autonomous vehicles less bumpy and more direct. The risks associated with AVs require all parties to work together to resolve them. And the level of collaboration will determine the pace at which they are overcome.

However, rather than suddenly pivoting to fully autonomous driving, the ecosystem will continue to make incremental pit stops toward the destination. Given the complexity of achieving complete autonomy, attempting to predict the exact timeline is futile. However, as the software arms race continues, by 2030, we should have a clearer understanding of when a secure and autonomous future will come into view.