Driver assistance technology could have unintended consequences. Could people end up losing their driving skills, and what might that mean?
All the advanced driver assistance systems (ADAS) are expected to help prevent crashes with some functions doing better than others. However, there are its unintended consequences, including potential drawbacks with long-term use and dependency of ADAS equipped vehicles.
Don’t get me wrong — ADAS functions are important advances in driving safety and rapid deployment is the right strategy for saving lives, lowering injuries, and reducing property damages. But this column looks at potential negative aspects of long-term ADAS usage. I am also trying to make a speculative estimate of the potential size of this issue by using two different methods to generate two forecast scenarios. One method is to look at how fast ADAS vehicles are growing and make speculations on what portion of ADAS users may see declining driving skills between 2020 and 2040.
The second method is to look at historical driver’s license statistics by driver ages and project these driver age mix between 2020 and 2040. I am only using three groups: seniors (65+ years), inexperienced drivers (under 19 years) and experienced drivers (19 to 64 years). For each group I make an estimate of what portion will have declining driving skills between 2020 and 2040.
Yes, this is rather ambitious. But think of it as what-if scenarios that deserves future research to determine if and how serious this problem could be. I could see this as thesis-topics for many masters and Ph.D. degrees and/or research grants from the U.S. Department of Transportation.
Feel free to do your own version of my speculative forecasts and estimates!
The next table is an overview of the issues for what I call “ADAS & Laws of Unintended Consequences.” The content of the table is explained below.
There are three parts to the above table. The rows in black are general comments about the ADAS issues that may lower long-term driving skills of ADAS and some AV users. The ADAS and AV mode issues have been covered lately—see this article by Junko Yoshida.
In summary, extended use of ADAS functions has the potential to lower driving skills for a significant portion of drivers as some level of ADAS features are now on every new car sold. At a minimum every new car has display-based backup driver assist. Many ADAS L0 warning features are proliferating such as forward collision warning (FCW), lane departure warning (LDW) and lane keep assist (LKA).
The second part of the table is information on the forecast scenario using ADAS sales and is in red color. This section has ADAS and AV data from IHS Markit that is used in the next forecast section. The data includes ADAS and AV sales share for the U.S. for 2020 and 2030. This section also has perspectives on the most important categories that is likely to lower the driving skills for users of each of the ADAS and AV levels.
The third table section provides similar information for the estimates that are based on driver’s license data. The senior or 65+ year category is increasing its share of U.S. population and share of total driver’s licenses. The experienced driver category is by far the largest category and accounts for 70% or more of all driver’s licenses.
The inexperienced driver category is the smallest category and consist of the young drivers under 19 years old. However, I am using a rolling sum of six years starting in 2020 for this category—in other words after six years they become experienced drivers.
This section also has assessment on the most important categories that is likely to lower the driving skills for the three driver’s license categories.
ADAS unintended consequences: Estimates using ADAS data
The next table is summary of a forecast spreadsheet from 2020 to 2040. I only included data for five years: 2020, 2025, 2030, 2035 and 2040. There are four ADAS-AV segments included: L0-L1, L2-L2+, L3 and L4-L5. At the bottom of the table is the sum of the four categories.
The methodology for each ADAS-AV segment is the same. ADAS-AV vehicle sales is based on the segment market share I got from IHS Markit for 2020 to 2030. I extended this forecast to 2040. The ADAS-AV vehicles in-use I estimated based on historical data and using eleven year as the vehicle lifetime. The parameter marked in red and called “portion with lower driving skills (%)” is the great unknown and is a speculative estimate. The red number is multiplied with “vehicles in-use” to get “drivers with lower driving skills.”
From the above table, the L2-L2+ segment has the highest “portion with lower driving skills” at 39% in 2040. This is due to lots of mode confusion, overconfidence (i.e. such as Tesla AutoPilot) and the potential for extra distraction issues.
L3 has lower “portion with lower driving skills” than L2 because the mode issues have fewer complexities and the L3 is driving by itself part of the time. L4 has the lowest “portion with lower driving skills” because the robotaxi segment will always be self-driving and the personal L4s will be self-driving part of the time.
When the four categories are added up, the “portion with lower driving skills” grows from about 0.5% in 2020 to 32.5% in 2040 or 0.67 million in 2020 to 71.5 million in 2040. The vast majority is drivers using ADAS vehicles.
ADAS unintended consequences: Estimates using driver’s license data
This section is a summary of a similar forecast based on looking at how driving skills may evolve by using historical driver’s license data by age groups and projecting this data from 2020 to 2040. There are detailed U.S. driver’s license data by state, age and gender going back to at least 1963.
I downloaded comprehensive U.S. driver’s license data from 1998 to 2018 and analyzed the age share trends. From this historical data on age trends, I projected how the three age groups would change from 2020 to 2040. The historical data shows that the three age groups change slowly, and I used similar slow changes in the 2020-2040 forecasts.
I also included the U.S. population forecast to 2040 and the share of population that has a drivers’ license at the top of the next table. The U.S. population grows from 333 million in 2020 to 374 million in 2040 according to the U.S. Census Bureau. The percentage of population with driver’s licenses is projected to peak between 2025 and 2030 and start declining as Mobility as a Service (MaaS) usage grows rapidly during this timeframe. The number of driver’s licenses peak in 2035 at around 235 million.
The changing share of three driver’s license age segments are shown in next table. The 65+ group increases its share every year—from 20.3% in 2020 to 22.3% in 2040. By comparison, seniors accounted for 16.1% of driver’s licenses in 2010 and 14.6% in 2005. The under 19 group had 1.9% share in 2007 and declined to 1.4% in 2015. The forecast is for 1.35% share in 2020 and 1.15% in 2040 for the under 19 group.
Lower driving skill estimates
The above table summarizes the estimates for lower driving skills for three groups. As in the ADAS-based forecast, the parameter called “portion with lower driving skills” determines the estimate and is again the great unknown and a speculative estimate. This parameter is marked in red.
The 65+ group is expected to see the most impact as their driving skills decline with age and ADAS functionality may hasten this trend. My speculative estimate is that the “portion with lower driving skills” will increase from 1% in 2020 to 38% in 2040. The equivalent number of impacted drivers increase from 0.47 million in 2020 to over 21 million in 2040. The factors that lowers driving skill among seniors are reaction time, vision and hearing issues, mode confusion and lack of HMI simplicity. ADAS functionality and implementation are major factors in these issues.
The group called inexperienced drivers is a sum of six years of under 19 years drivers. The inexperienced driver group start at 0.5% in 2020 for “portion with lower driving skills” and is estimated to grow to 32% in 2040. The factors that lowers driving skill among inexperienced drivers are overconfidence, mode issues, smartphone distractions and HMI clarity. One concern is that the inexperienced group will not reach the driving skill levels that were common before ADAS became important.
After six years of driving experience the inexperienced drivers move to the experienced driver group. The experienced group is calculated by subtracting the 65+ and inexperienced group from total driver’s licenses. It is the largest group at over 170 million through the forecast period. The experience group start at 0.1% in 2020 for “portion with lower driving skills” and is projected to grow to 26% in 2040. The number of impacted drivers increase from 0.18 million in 2020 to over 45 million in 2040. The factors that lowers driving skill among experienced drivers are distractions, mode issues, overconfidence, and HMI complexity.
The sum of the three groups are shown at the bottom of the previous table. The group sum for “portion with lower driving skills” is 0.1% in 2020 and is forecasted to grow to over 29% in 2040. The potential number of impacted drivers is 0.67 million in 2020 with growth to 76.6 million in 2040.
These results are certainly speculative forecast scenarios. To get better perspectives on the size of this potential problem, I used two different forecast methodologies. The forecast parameters with the most uncertainty is the “portion with lower driving skills” and my guesses need more information. Obviously, we need better research on the range and how this main forecast parameters changes over time.
The difference between the two forecast scenarios varies from 0.2% to 8.1% with the lower percentages coming at the end of the forecast years.