Skip to main content
Start of main content

What’s that car’s IQ?

October 15, 2019

Understanding autonomous vehicle performance is key to building public trust in smart mobility

On June 21 of this year, an EasyMile EZ10 autonomous vehicle (AV) rolled out on the streets of Montreal for a six-week pilot project. Starting at the city’s famed Olympic Stadium and ending at the Maisonneuve Market 1.4 kilometers away, the shuttle pilot ran through a dense neighborhood collecting data and assessing the vehicle’s performance in car traffic, through intersections, and near pedestrians and cyclists. The City of Montreal is hoping to one day add driverless shuttles as a complement to their transit services. The number of autonomous shuttle projects identified or planned has grown significantly over the past year, to more than 250 in the United States alone.

By standardizing and creating quantified measures, we can establish more realistic expectations as to what an AV can do and what the driver will still need to do.

Convenience is a key part of the picture, but transportation planners are focused on a more important goal: public safety. According to the World Health Organization, road accidents are the eighth leading cause of death globally. That adds up to 1.35 million lives lost every year, along with 50 million injuries. Sadly, the majority of those accidents are caused by human error. If we can increase safety by removing or reducing the human error factor through new artificial intelligence (AI) algorithms, it is something we need to explore.

Connected autonomous vehicles (CAVs)—AVs that are connected to other vehicles or infrastructure by signals—have the power to change our transportation system for the better, and utilizing these pilots to gather data are crucial to understanding how to make that happen. The main barrier to adopting CAVs, like any technology that changes the shape of our society, is public trust.

Public trust is difficult to gain and easy to lose. Trust in smart mobility solutions will come through understanding—how CAVs work, what they can and can’t do, and how they will make life better. There are things we can do right now to build the understanding that leads to public trust: establishing vehicle IQ assessment, good codes of practice, and professional training.

Testing and creating standards for a car’s IQ can help to build public trust around CAV technologies. 

Assessment: measuring a car’s IQ

Not everything in life is created equal. You probably wouldn’t take a Chevy Cruze off-roading, and a Hummer might not be the best rental choice if you’re staying in Le Marais in Paris. Similarly, not all CAVs are the same, and some will perform differently dependent on conditions. While they all use some combination of technology in order to operate (Light-Detecting and Ranging, radar, software, GPS, RTK radio, and Dedicated Short-Range Communication), there are important variances.

International SAE Standards currently have six levels of driving automation. While this offers important guidelines, there is a lack of nuance that relates to the performance of different vehicles—or their “driving IQ.” To help address this, we can focus on tangible tests that can be observed and measured:

  • Traffic infractions: Ideally logged by the vehicle itself and not dependent on outside intervention, there is a statistically significant relationship between infractions and crashes for human drivers. This will likely translate to CAVs.
  • Roadmanship: Captures the ability to drive on the road safely without creating hazards and responding well to the hazards created by others. The concept centers on whether the vehicle “plays well with others.” There are some challenges to creating unified definitions for this, but it is an attractive metric to aspire to.
  • Disengagements: When a human needs to take control of the vehicle. This can be initiated by the CAV itself, by in-vehicle or remote safety drivers, or by passengers. These are currently a widely used, though non-standardized measure of safety.

By standardizing and creating quantified measures, we can establish more realistic expectations as to what a CAV can do, and what the driver will still need to do. Imagine an industry-approved sticker on the window of a car that clearly explains the capabilities and limitations of the technology in the vehicle. This could also help the government regulation process. If we boost expectation and understanding, we are assembling the building blocks needed for increased public trust.

Regulations and codes of practice

The importance of having definitive regulations in place to help guide the socialization of CAVs cannot be understated. Every jurisdiction should look to develop their own code of practice—a written set of rules that explains how a profession should behave. For example, codes exist for engineers to create certainty, predictability, and understanding among the general public.

Any framework of this type needs to be robust, yet flexible. Manufacturers and technology providers should be able to follow the guidelines in order to obtain trial permits from governing bodies, and eventually introduce their CAVs to market. The benefits of such frameworks include safer travel, increased comfort for drivers and passengers, and a more sustainable ecosystem for prototyping and engineering. This all helps contribute to more structured tests and trials, risk assessment and review, approvals, and acceptance.

Dubai’s Road and Transport Authority engaged Stantec to help them create a code of practice on how to implement self-driving vehicles.

Professional training

Any code of practice should also be supplemented by a strong standardized training program, which contributes to widespread understanding, builds a strong stable of expertise, and creates a foundation for an industry to thrive. The world of CAVs is extremely fast-moving. From updates in technology, deployments and policies, it’s critical to stay on top of the trends. While different jurisdictions might prioritize different things in a training program, some important basics to consider are:

  • The technological elements of CAVs
  • Lessons learned from present and past CAV deployments
  • The use of testbeds in CAV testing
  • The role of artificial intelligence and machine learning

The challenge of training is to reach a broad spectrum of stakeholders involved in making these vehicles work—everyone from software developers writing code to traffic safety experts planning routes. Many people have expertise in one part of the larger pie, but knowledge of the big picture should be shared across industry, government, and academia. There are few broad scope training opportunities currently available. The public will justifiably be cautious of driverless vehicles, so providing education is just as important as perfecting the technology itself.

Stantec’s Transportation team performed the engineering work for the City of Montreal’s first autonomous electric shuttle.

Bringing it all together

Public perception of CAVs and smart mobility technology is evolving. The fear of change isn’t going away, but actions taken now will shape how well or how poorly this evolution is received.

Generating a level of awareness, understanding, and social acceptance of new ideas among the general public is critical. Transparency in testing and evaluation will help make AV deployments possible. Of course, this needs to be backed up by strong standards and educational practices. And by informing consumers exactly how smart our cars are and what they can do, we can form a level of trust.

Throughout history, disruptive technologies have always needed validation. The same is true for CAVs. By laying the right foundation for acceptance, public trust can be slowly earned over time. Only then can the societal benefits of CAVs be realized.

End of main content
To top