Autonomous vehicles are artificial intelligence-based mobility devices. They are not a new technological development as the concept has been a work-in-progress for centuries. However, this century has the unique opportunity to witness the emergence of self-driving cars on roads all over the world. This, thus, makes it imperative to quickly address how the world would change with robotaxis on our roads and how we ought to prepare for them. The advantages and disadvantages of autonomous driving machines abound; however, the onus is on governments to minimize the threats and harness the benefits. This chapter, therefore, is about how governments can effectively leverage autonomous vehicles to promote humans' quality of life and the necessary considerations on how to minimize the risks associated with self-driving vehicles.
TopThe Technology
In the 1500s, Leonardo da Vinci designed an autonomous vehicle (AV) - a self-driving cart that would move along preset paths (Stanchev & Geske, 2016). This is believed to be the first AV design. Several centuries later, the city of Manhattan was home to the demonstration of a radio-controlled car in 1925. However, this demonstration was discontinued when the car crashed into other vehicles. Further in the 20th century, General Motors created a self-driving car controlled by an electromagnetic field on a road with implanted magnets (Bimbraw, 2015). The road was modified to aid the car using sensors mounted on the car. By the 1960s, cameras were attached to AVs to aid their navigation. Japan improved on this idea to gather enough data with the cameras, enabling the cars to go as fast as twenty miles per hour. By 1995, Carnegie Mellon University had made a self-driving car that could travel from Pittsburgh to San Diego through advanced image processing using neural networks. Advanced image processing helped the car to steer, but the speed and braking were human-controlled (Stanchev & Geske, 2016). In the 2000s, the United States Department of Defense became involved in AV challenges and testing. In the 2010s, ridesharing companies tried to push the current research limit of AVs, seeing as AVs are critical to their profitability and bottom line. AV research has come a long way, even though commercial AVs have proved elusive thus far.
To further understand AV research and appreciate how AVs should be regulated, let’s explore the levels of vehicle automation as defined by the International Society of Automotive Engineers (SAE) (Rivard, 2018). At level zero, we have no automation. Humans, at the wheel, make navigation, brakes, and speed decisions. At level one, we have the driver assistance phase where the car can control either its speed or steering, but not the two simultaneously. At this phase, a driver is at the wheel to take total responsibility for the car. Level 2 has partial automation in which the car can accelerate, brake, and steer itself in some circumstances. Tesla’s autopilot is an excellent example of level 2 vehicles in which drivers must remain at the wheel to respond to traffic signals and hazards. Conditional automation is achieved at level 3. Here, the car can drive itself while also scanning its environment. Nevertheless, a driver must be at the wheel if the car needs help navigating unfamiliar scenarios. Level 4 is a high automation level where mere human oversight is needed. The car still has a steering wheel and pedals, but can essentially drive itself. Level 5 is the peak of autonomous vehicles. Every human in the car is a passenger as there are no pedals or steering wheel (Stanchev & Geske, 2016). Thus, AVs at level 5 are considered the future of mobility and the holy grail of automation, the standard on which this chapter is based.