Driver Assist Systems Proved Their Worth In Largest Study Yet
If you’re friends with anyone on “Tesla Twitter,” you’ll see the occasional video of absolutely terrible driving, often tagged as #HumanPilot. Why? Because Tesla fans believe that the future Full Self Driving software will drive safer than humans, make fewer mistakes, and save lives. They even sometimes go as far as to say that FSD skeptics have blood on their hands, because they’re slowing down the deployment and costing lives that self-driving cars should have saved.
But, what if there’s a middle ground here, something other than full self driving and #HumanPilot? Are human drivers, backup up by assistance from computers when we fall short, a safer alternative that’s available today and not in “two weeks?” Do systems like automatic emergency braking, pedestrian braking, lane keep assist, and adaptive cruise control make human drivers safer to the point where we don’t need mandatory autonomous vehicles in the future for safety?
Studying This Question
In 2018, NHTSA partnered with a variety of automotive manufacturers to conduct one of the biggest safety studies ever. By partnering not only with one or two automakers, but with automakers representing a vast majority of the industry, real data could be gathered that could tell a very complete truth about Advanced Driver Assist Systems (ADAS), such as automatic emergency braking and lane keep assist. This program is known as the Partnership for Analytics Research in Traffic Safety (PARTS).
Naturally, Americans are skeptical of government data collection, and the team behind PARTS knew that would be a concern that could drive vehicle manufacturers away. After all, if customers think an automaker is spying on drivers, it could greatly impact sales. So, NHTSA relies on a third-part company that aggregates and anonymizes the data, giving government officials usable data without being able to identify any individual drivers.
They started by looking only at automatic emergency braking (AEB), and then started looking at more advanced systems like adaptive cruise control, lane keep assist, and others. This is an ongoing process.
“MITRE has managed the Aviation Safety Information Analysis and Sharing [ASIAS] initiative on behalf of the FAA and aviation industry for over a decade,” said Gregg Leone, vice president and director of MITRE’s Center for Advanced Aviation System Development, in a press release last year. “The ASIAS collaboration model, which has been critical to improving aviation safety, can also advance traffic safety, and we are proud to serve as the independent third party in this partnership.”
“Although ‘data sharing’ has been a discussion topic in the automotive safety landscape for years, there are few — if any — demonstrable efforts other than PARTS,” said Tim Czapp, industry co-chair of the PARTS Governance Board, and a senior safety manager at FCA.
“No other projects have the scope or scale of PARTS, nor are they derived from such a proven and successful model. We all want to cooperate and lay groundwork needed for our autonomous future, and believe that adapting the ASIAS [aviation] model is the best way to shape and benefit from a collaborative safety analysis.”
How These Driver Assist Systems Stack Up
PARTS and MITRE (including the automakers) have spent the last several years studying the effectiveness and safety benefits of ADAS systems. The purpose of this analysis was to explore how effective ADAS features are in avoiding crashes that would be relevant to the system. To do this, they used police-reported crash data and vehicle equipment data contributed by PARTS partners. They looked at data from 93 different vehicle models from model years 2015 to 2020 that had crashed in 13 states between January 2016 and August 2021. The fact that the data is relatively recent allowed them to keep up with new ADAS features as they are deployed into the marketplace.
The purpose of this study was to assess the effectiveness of various safety features including forward collision warning (FCW), automatic emergency braking (AEB), pedestrian automatic emergency braking (PAEB), lane departure warning (LDW), lane keeping assistance (LKA) and lane centering assistance(LCA). crashes that were relevant to each system were defined as front-to-rear crashes for FCW and AEB; frontal crashes involving pedestrians for PAEB, single vehicle road -departure crashes for LDW, LKA, and LCA.
In this study, three crash severity groupings were defined and measured for ADAS effectiveness: All Crashes (property damage only, unknown injury level, or any injury severity); Injury Crashes (any injury severity including fatality); Serious Crashes (known serious injury or fatality).
The Results
The analysis showed that ADAS features like FCW and AEB provide massive safety benefits in different scenarios; LDW and LKA, however, are only mildly effective. Conversely, PAEB requires more data to make a judgement of its efficacy.
For braking, the study found that when a vehicle is equipped with FCW + AEB, it reduces the chance of all front-to-rear crashes by 49%.2 In addition, it was found that injury front-to-rear crashes were reduced by 53% and serious front-to rear crashes were reduced by 42%. This study found that equipping vehicles with front collision warning (FCW) systems reduces the overall number of front-to-rear collisions by 16%. When these FCW systems are combined with active braking, the reduction in collisions jumps to 19%.
Another very interesting thing the study found was that AEB is successful in a variety of conditions, even when the roadway, weather, and lighting are not ideal. ADAS can still help by making the crash less severe and reducing injuries.
For lane departure avoidance/warning tech, they found that LDW + LKA reduced all single-vehicle road-departure crashes by 8% and injury single-vehicle road-departure crashes by 7%. Adding LCA reduces crash rates by about the same amount (9%). There were no other significant results when analyzing injury or serious crashes with these features together, nor was there a significant reduction for LDW alone.
The findings weren’t as noticeable when investigating pedestrian braking. This study looked into the effectiveness of PAEB on non-motorist crashes but was unable to detect an effect because of (e.g., the limited number of these incidents in crash reports and the lower level of market penetration for PAEB, particularly in older model years).
What We Can Learn From This
It’s pretty clear that some assistance is far better than none, and that systems that provide this assistance are available and useful today. While these systems could improve (especially for detecting pedestrians and staying in lanes), reducing rear-end collisions by half is a big deal.
It’s also important to look to the future with assist systems. Instead of advocating for an all-or-nothing approach to autonomous systems, and demanding full adoption of them when they’re ready for primetime, we should look at a continuum of technologies that save life today while giving us options tomorrow.
Have a tip for CleanTechnica? Want to advertise? Want to suggest a guest for our CleanTech Talk podcast? Contact us here.
Latest CleanTechnica.TV Video
CleanTechnica uses affiliate links. See our policy here.
