Are Self-Driving Cars Safe Enough?

With their embedded software and data-driven applications, the built-in safety features of self-driving cars are rock solid! While this statement may mean something to engineers, for average consumers, the mere thought of cars roaming highways—without drivers—will still raise little red flags.

Let’s face it, it’s difficult to imagine technology carrying the same decision-making abilities as us, no matter how advanced it is. If our human brains had to study DMV handbooks diligently to get our driving privilege, how can cars drive themselves without learning a similar set of guidelines?

They do―in a different, hard-coded way―through the Advanced Driver Assistance System (ADAS); this framework is what enables self-driving cars to pass safety regulations with flying colors. Yet, however, robust the system may be, if consumers can’t fully comprehend safety mechanisms behind objects moving 80 MPH, the majority of them won’t be biting into the autonomous automotive offering anytime soon. So let’s drive into it.

ADAS―An Affirmative Safety Feature

Self-driving cars are built with sensors, actuators, and microprocessors that enable them to receive, process and utilize data to make critical moves such as stop, slow, or go. The ADAS framework, extended by vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I), is what makes safe self-driving possible.

V2V―Keeping Vehicles Safe

The moment connected cars enter another’s circulatory radar range of 300 meters, they will “shake hands” and exchange pre-packaged information about their speed, direction, positioning, condition, and more. V2V utilizes this data in algorithms driving applications such as:

  • Intersection and Left Turn Assistance
  • Emergency Electronic Brake Light Detection
  • Forward Collision Detection
  • Blind Spot and Lane Change Assistance

While human drivers may err at blind spot detection, autonomous cars are much less likely to glitch, as Google and Delphi’s autonomous cars recently demonstrated during their test drive. Just as the Delphi Audi Q5 was getting ready to change lanes, it sensed Google’s autonomous prototype moving into the lane. Instead of territorially bumping heads, the Audi chose to delay its own lane change, thus ensuring the safety of both autonomous cars (and everybody else on the road). Delphi’s safety division praised their car’s response:

It was a good example of how our car leverages complex software algorithms to assess its situation—much like a human driver would…[...]...automated vehicles which need to work through a decision via a highly complex network of technology and data processing to calculate the correct move. Our car did exactly what it was supposed to.

And what a courteous move it was. If only human drivers could interact with positive attitudes, it would make driving less stressful and roads much safer. But the development of autonomous cars has cut us some slack; intricate software rules are written into automotive programs to keep cars from collisions and passengers relaxed—with little emotional management required.

V2I―Keeping Roads Safe

Just like how V2V commands inter-vehicular activities for collision avoidance, vehicle-to-infrastructure technology directs autonomous cars to adhere to road regulations. Software-enabled roadside infrastructures such as stop signs and traffic signals will relay critical traffic information so autonomous cars can respond appropriately. Some safety applications of V2I include:

  • Curve Speed Warning: Signaling vehicles to slow down if its present speed is too fast for an upcoming curve
  • Red Light and Stop Sign Gap Assist: Telling vehicles when it is safe to move through an intersection
  • Reduced Speed Zone Warning: Directing vehicles to decelerate or change lanes when approaching work zones
  • Spot Weather Information Warning: Providing vehicles with real-time weather updates
  • Railroad Crossing Violation Warning: Assisting vehicles during railroad crossings
  • Oversize Vehicle Warning: Directing oversized vehicles to alternative routes if oncoming road is too narrow or tunnel too low

In short, V2I mimics human driver’s need to keep a certain level of environmental awareness―only autonomous cars do it in a much more comprehensive and non-distractive manner.

Safe Enough? Most Likely, Yes.

Otherwise, they wouldn’t be released onto the roads in growing numbers. The National Highway Traffic Safety Administration predicts that V2V-enabled autonomous vehicles will reduce driver-related accidents by 70-80%, preventing thousands of deaths and injuries each year. Of course, ADAS functionalities still require much more testing, but current developments are pointing in a hopeful direction.

What may end up happening when self-driving cars dominate the road is that their speed, distance, and mobility will converge somewhat robotically especially during peak hours. Picture lines of autonomous cars driving under 40 MPH in a highly predictable fashion going to and from work. Younger generations might never experience the heightened adrenaline of racing past others―but perhaps this is the exact type of regulation needed for a safety-first highway utopia.

What are your thoughts on self-driving cars? Would you invest in one yourself?

National Highway Traffic Safety Administration: Readiness of V2V Technology For Application