Letting go of the wheel: Can humans do it?

May 3, 2017 OpenSystems Media

A look at the future vehicle through the eyes of SAE International

This article is the fourth of a six-article series from SAE International providing a practical look into the feasibility of connected vehicles and autonomous driving. Read the first, second, and third articles.

Letting go of the wheel – it’s a daring and foreign concept for even the most trusting individual. The idea of hopping into a vehicle, inputting a destination, and letting the vehicle lead the way without any intervention seems straight out of a sci-fi movie, yet numerous automakers and technology developers are showcasing autonomous and connected car systems that prove it is possible. The technology will eventually be here, developing more day to day, but will humans be ready to make the transition?

Humans are used to controlling their means of personal transport. Whether it’s a steering wheel in an automobile, handlebars on a bike, or even reins on a horse, humans have relied solely on themselves for navigation and managing the motion of their transportation. Within the last 50 years or so, there has been an implementation of automated transportation systems, such as low-speed trams and trains, but even they have a driver on-board or remote monitoring, ready to take control in an emergency situation.

Technologies have steadily developed to change the driving experience, sans human. From on-board sensor fusion, to advanced data processing and external systems that allow vehicles to communicate with each other and the infrastructure around them, the human driver’s role is quickly changing and perhaps eventually becoming obsolete. However, there are several factors holding back the transition – and misunderstanding may be the biggest hurdle.

According to Bobbie Seppelt, research scientist at Touchstone Evaluations and at MIT’s AgeLab, and an SAE technical standards committee member, “There are several consequences of misunderstanding automated driving. The first is mistrust resulting in disuse. If drivers do not understand how the automated driving system works and how to properly interact with it, they are likely to instead rely on their own control. Potentially more concerning, in terms of its implications for safety, are those drivers who place too much trust in an automated system and overuse the technology in situations for which it was not designed. This is how accidents happen. We’ve already seen situations where a semi-automated driving system is relied on beyond its capabilities and the results can be deadly. We’re working to educate drivers so a negative connotation is not formed about current advanced vehicle technologies that subtend self-driving ones to come. The potential benefits in relieving fatigue, overall safety, and productivity far outweigh the negatives, provided we are able to support drivers in these early stages in forming calibrated expectations of system behavior.”

The cornerstone of educating drivers about an autonomous vehicle’s capability is through its designation on SAE Levels of Automation. Formally known as SAE J3016, automated driving systems are rated in this SAE Recommended Practice from 0-5 based on the level of human and system responsibility. In short, level 0 relies on full human driver control, levels 1-2 rely on humans to monitor the driving environment, and levels 3-5 rely on the system to perform the driving function. Level 5 is full automation that requires zero human intervention in all driving situations.

Right now the industry is focusing on bringing level 3 and 4 systems to market, meaning they have a need for a human driver. Even a driver on standby will not be going away anytime soon. Additionally, fully automated vehicles rely on a connected car infrastructure to completely remove the need for a dedicated driver. While systems are being built in special urban zones around the world, the cost of installation and upkeep will slow expansion to rural roads, possibly taking decades to complete widespread implementation. In the meantime, it is essential to address challenges facing drivers. These core challenges include the ability to understand and operate the automated driving vehicle, remaining engaged with the vehicle and driving environment while it is operating, and defining the responsibilities and liabilities of the driver and vehicle.

One of the most recent examples of these challenges not being met involved a fatal crash that occurred with a Tesla vehicle fitted with the company’s Autopilot feature, a level 2 automation system. With such a system, the lines between who is at fault in an accident can become blurred. Following this particular incident, lawmakers and insurance providers are now scrambling to update laws and policies that address advanced driving technologies.

Automakers are scrambling, too, to develop ways to overcome these challenges of keeping a driver informed and alert. SAE has been a leader in developing industry standards for an effective human-machine interface, trailblazing with best practices in developing effective vehicle-to-driver communication methods, assessing driver activities and workloads, and developing common terms and icons so drivers have a consistent experience between manufacturers and their vehicles.

Taking this to the next step, SAE is currently developing a new set of standards specifically tailored for automated driving. These include J3114, which defined key terms in relation to human interaction with automated systems and identified future standards needs focused on human factors, ergonomics and human performances related to automated driving systems. The end goal is to develop a safety-based human-machine interface that can be replicated in its basic form across all automated vehicles, and which both assesses a driver’s state and communicates effectively to keep the driver engaged.

“We’re currently evaluating level 2 systems and are finding the toughest challenge to overcome is awareness while driving. It’s important to collect data both in controlled and naturalistic environments to test and advance learning on what tools and approaches are best for keeping drivers alert and safe,” said Seppelt. “As the basic framework is laid out, we will see more of these vehicles on the road in the next several years. Data collected from these vehicles will be critical in further refining standards and developing a working framework for introducing more advanced systems.”

As we look forward, it appears the human driver will be behind the wheel for the foreseeable future. As automakers develop their own advanced automated driving systems or partner with technology providers to bring them to market, the driver’s seat and steering wheel will remain in place for a varying degree of human intervention. As systems become more complex, the reliance on developing a solid foundation of standards and adequate education will determine whether the technologies take off in mainstream applications, perhaps reaching a day in the distant future in which level 5 driving is feasible.

A safer, more efficient method of transportation is well on its way, but don’t hang up your driving gloves just yet.

Shawn Andreassi is Manager of Corporate Communications at SAE International.

SAE International, a global association of more than 128,000 engineers and related technical experts, has published more than 1,600 technical standards and recommended practices for passenger cars. The organization, founded in 1905, leads the industry with advanced, unbiased knowledge to benefit society.

Society of Automotive Engineers (SAE) International

www.sae.org

sandreassisae
Previous Article
Cross-industry semantic interoperability: Glossary

"When we define a word we are merely inviting others to use it as we would like it to be used; that the pur...

Next Article
2017 Top Embedded Innovator: Oliver Winzenried, CEO and Co-Founder, Wibu-Systems
2017 Top Embedded Innovator: Oliver Winzenried, CEO and Co-Founder, Wibu-Systems

Since co-founding Wibu-Systems in 1989, Oliver Winzenried has dedicated his career to securing embedded sys...

How to Develop Cross-Industry IoT Interoperability

Multi-Part Series