ONE HUNDRED. On average, that is the number of people that die every day in motor vehicle crashes in the U.S. alone.
To put this into perspective, every month, more Americans are killed in crashes than the total number of lives tragically lost as a result of the terrorist attacks on 9/11. For those of us in the transportation engineering discipline, whose profession is founded primarily on the principle of safe movement of people, this statistic should send chills down our spine. With approximately 94 % of crashes being caused by human error, and with cases of distracted driving on the rise, we desperately need to find a safer alternative to a human driver.
The secondary principle of our profession is the efficient movement of people. Over the past century, as traffic density has increased, our solution has been to build more roads and highways, or to widen existing roads and highways to add more lanes, in an attempt to maintain an acceptable level of service. While it seems logical to supply more roads and highways as the traffic demand increases, this practice is not sustainable because we will eventually reach a point in time where widening a road or highway will not be possible, due to finite right-of-way availability. To be proactive, there has to be a paradigm shift in our profession, where our focus changes from adding more lanes to maximizing the efficiency of the lanes that we already have.
One emerging technology that holds the promise of improving the safe and efficient movement of people is the autonomous vehicle.
One emerging technology that holds the promise of improving the safe and efficient movement of people is the autonomous vehicle. An autonomous vehicle (also known as a self-driving vehicle, driverless vehicle, or automated vehicle) is one that is capable of performing the complex task of driving without a human driver. Autonomous vehicles accomplish this by pairing sophisticated computer algorithms with sensing systems such as visible and infrared (IR) cameras, light detection and ranging (LIDAR), radio detection and ranging (RADAR), and dedicated short-range communications (DSRC) technology. Some of the systems are purposefully redundant, such as those used for obstacle detection, so that if one system malfunctions, another system acts as a “back-up” system. Recent advances in technology have enabled autonomous vehicles to go from dream to reality, with several automakers having ongoing programs to develop and launch autonomous vehicles into the market.
With the arrival of autonomous vehicles, the Society of Automotive Engineers (SAE) International developed a widely-accepted classification system consisting of five levels of driving autonomy. The classification system uses the numbers zero through five (inclusive) to identify the level of driving automation. The majority of vehicles on the road today fall into the level zero class, meaning no vehicle automation (all human control). Newer vehicles on the road are equipped with semi-autonomous features, such as automatic braking for collision avoidance, and adaptive cruise control with lane centering. As the numbers on the classification system increase, human control starts to gradually fade out, and automated control starts to fade in. A level five vehicle has full automation, with no option for human driving (i.e., no steering wheel and pedals). The National Highway Traffic Safety Administration (NHTSA) adopted SAE International’s levels of driving autonomy in their first ever “Federal Automated Vehicles Policy”, issued in September 2016. Currently, Texas is one of only three states that allow testing of autonomous vehicles without a human in the vehicle. Historically, transportation engineers have designed roads and highways based on human parameters. From posted speed limits to the curvature on a road, every design component takes into account the limitations of a human driver, specifically our perception-reaction time (PRT). Our PRT is the main reason we are supposed to leave a progressively larger gap from the vehicle in front of us as our speed increases. Because of limitations on speed and large gaps between vehicles, current maximum highway capacity is limited to approximately 2,400 vehicles per hour per lane. Humans also make mistakes, so roadways have to be designed with certain safety features in place, all which contribute to higher construction and maintenance costs. As an example, the width of a typical highway lane is 12’. With the average width of a passenger vehicle being approximately 6’, the remaining 6’ of lane width serve as a buffer space to make the highway more forgiving.
Autonomous vehicles will not be bound by the same limitations as humans because they will have virtually no reaction time, and they will be immune to distractions and fatigue. Autonomous vehicles may one day have the ability to operate in a high-speed (detached) train formation and self-detour around potholes, unexpected roadway debris, disabled vehicles, and through work zones. With higher speeds and negligible gaps between vehicles, some projections indicate that highway capacity could ultimately reach approximately 12,000 vehicles per hour per lane (limited mainly by the mechanical performance of vehicles). Lane widths may shrink since autonomous vehicles will operate with higher levels of precision. Drastic changes in vehicle performance and operation will lead to the revision of roadway design standards and guidelines. Crashes may become a rare occurrence, and will likely result in lesser injuries and damages when they do occur, as safety systems continue to improve into the future.
As autonomous vehicles become more prevalent on public roadways, there will inevitably be a transition period during which manually-driven vehicles and autonomous vehicles will have to co-exist. This will be a critical time for autonomous vehicles to prove themselves worthy of sharing the road with traditional vehicles. On highways, autonomous vehicles may initially be restricted to their own designated lane that may potentially have a higher posted speed limit than ordinary lanes. At some point during this period, insurance companies may have to modify their policies from covering human drivers to covering technological failures. Government legislation will need to evolve to keep up with the dynamic growth of autonomous vehicle technology.
Some experts believe that manual driving may one day be outlawed (or restricted to geo-fenced areas, such as a race or test track), based on the premise that a human-controlled vehicle may pose too large of a safety risk to others. In a future with 100 % autonomous vehicles, traffic signs and traffic signals may possibly become obsolete. Professor Peter Stone, with the Department of Computer Science at UT, developed a conceptual Autonomous Intersection Management (AIM) reservation system. The AIM reservation system would use a controller at each intersection that would wirelessly communicate with approaching autonomous vehicles and assign the r ight-of-way, allowing vehicles to travel through the intersection without stopping.
A future of autonomous vehicles is difficult to imagine for many, and rejected in full by others. The single biggest challenge for autonomous vehicles will be to earn widespread public acceptance, as has been the case w ith any new technology in its introductory stage. Some people may claim that they will never let a computer drive for them. This is completely expected. Not long ago, people were uncomfortable with the concept of a computer deciding when to deploy an airbag in their face, and felt apprehensive about trusting a computer to pump the brake pedal (several times per second) in emergency braking situations. Supplemental restraint system (SRS) airbags and anti-lock braking systems (ABS) are now standard on all new vehicles sold in this country. Both systems have proven to be highly effective in saving lives and preventing more serious injuries.
There are still many questions to be answered about the future of autonomous vehicles, ranging from moral and ethical dilemmas to concerns about potential hacking or malfunctioning of the vehicles’ automated control systems. However, we cannot ignore the large-scale investments being made by automakers to get autonomous vehicle technology to a trusted level. Earning trust is an incremental journey; it may take several years, possibly decades, but once autonomous vehicles have the ability to establish a proven track record of being safer than human drivers, I have faith that we will be well
on our way to ZERO.