Ridesharing, which combines multiple passengers into a single vehicle, is not new. The first car-based taxi service was launched in the late 1900s. In fact, the first taxi service via horse and carriage was initiated in the early 1600s. Some estimates suggest that the cost can be 20 percent lower using a ridesharing service compared to the traditional automobile. Additional societal benefits of ridesharing include reduced traffic congestion, lower automobile emissions, the conservation of non-renewable energy resources, decreased oil imports, and less driver strain, which should make the passengers more productive upon arriving at work or school. Over the past decade, a new ridesharing business model has emerged. Pioneered by Uber, the scheme provides on-demand service to users via a smartphone app. Typically, in this business model, the contract drivers use their own cars when providing transportation services, and the ridesharing firm receives a percentage of the fare. The domestic annual revenues for this market sector, which presently total $3 billion, are expected to double to over $6 billion by 2020. This contrasts with the nearly $11 billion in revenues for the traditional taxi and limousine sector, which by some estimates have declined nearly 20 percent over the past several years. The on-demand ridesharing model is seeing widespread acceptance on a worldwide basis. Peer-to-peer car-sharing service is an important variation on this construct. Here, individual car owners can “rent” their vehicles out for limited periods via a third-party service provider, again via a smartphone app.
The automobile today is a rolling physics lab with a lot of computers in it.— David Sloan
But wait, there is more! On the near horizon is the autonomous vehicle (AV). An AV is a robotic-controlled automobile that is designed to travel between destinations without a human operator. Autonomous cars use various kinds of technologies, including GPS and radar-based sensors. The former is employed to navigate the vehicle, while the latter is designed to avoid collisions. Lane detection artificial intelligence (AI) based algorithms are used to ensure that the vehicle remains properly aligned. These systems, which employ lane information from previous encounters and lane marker imaging, must operate effectively under various weather, road, and illumination conditions.
So-called “self-parking” systems provide a glimmer into the potential of this technology. However, many drivers find the use of current self-parking technology somewhat problematic since they are called on to participate in the process, e.g., to specify parallel or perpendicular parking directions. Presently, several automotive manufacturers (e.g., Ford and Hyundai) are planning to offer AVs by the early 2020s. The deployment of autonomous vehicles will, no doubt, create many societal challenges concerning the efficacy of existing auto insurance models, police and emergency responders, and traffic control systems. To that end, the U.S. House of Representatives passed the Safely Ensuring Lives Future Deployment and Research in Vehicle Evolution Act, or “SELF DRIVE” Act, in late 2017. The act, which is currently being held up in the Senate, requires automakers to demonstrate that their AVs are as safe as cars driven by humans.
Significant research on autonomous vehicles is underway. One recent study involving the use of shared autonomous vehicles (SAV) in the Austin, Texas, metro area revealed that one SAV could replace nine conventional cars, on average, while maintaining a one-minute level of service response using state-of-the-art AI scheduling technology. These statistics are amazing! On the downside, the results also showed that the system would generate approximately 8 percent more vehicle miles due to journeying unoccupied to pick up the next customer or relocating to a more favorable position in anticipation of next-period demand. Either the AV or SAV models would provide the riders the opportunity to engage in productive work/study during the trip, which tends to average 15 miles and take nearly 30 minutes.
The search engine giant Google and Uber have taken lead roles in the deployment of autonomous vehicles. Google states that they have logged over 1.5 million miles of AV operations with a human on board, as required by current law. At the other end of the size spectrum is Aurora, which has teamed up with several of the major automotive manufacturers (e.g., Volkswagen and Hyundai) to provide self-driving technology. As with any new technology, there will be setbacks—tragically, a driverless Uber vehicle recently struck and killed a pedestrian in Arizona. This incident appears to be the first pedestrian fatality involving an AV. To this end, some additional technological challenges that lie ahead include the ability of the AV to detect traffic enforcement personnel giving hand directions, take appropriate actions in construction zones, and make ethical decisions.
Moral dilemma regarding crash logic—An AV is heading for a dog that is on the street,
and if the car turns to avoid the dog, the car may run into humans, including children.
A variation on the self-driving car is the hybrid model, which could have considerable appeal to those motorists that will not feel comfortable riding in a fully autonomous vehicle. The idea is that in some driving modes, such as freeways or country roads, the onboard computers would drive the vehicle. In more densely populated areas, the human driver could be called on to perform some basic tasks, such as steering.
To this end, the National Highway Traffic Safety Administration has defined five distinct levels for self-driving AVs, which range from controlling specific tasks such as parking and braking (level 1) to where the vehicle is fully autonomous with no steering wheel or other human controls (level 5). The automobile manufacturers will no doubt use a phased approach, segueing from level 1 toward level 5, as a way to “break the ice” with the motoring public in much the same way the hybrid-electric vehicle paved the way for the all-electric automobile. However the fact that there are approximately 130 million passenger vehicles in the United States alone suggests that the transition to hybrid or completely self-driving cars will not occur any time soon.
Today, the average price for a new car is around $30,000. The initial costs for a modest-sized AV could add multiple $10,000 increments to that price point—depending on the specific options—once the development costs have been recovered. Since we are talking about reducing the cost of transportation to the motoring public, this is where the SAV model comes into play. The on-demand ridesharing firms could absorb the additional costs associated with the AV, which would be more than offset by a significant decrease in labor cost—no more human drivers. Not surprisingly, artificial intelligence will play an important role as these systems continue to roll out. To that end, while employment for professional drivers (e.g., taxi and delivery) and those associated with the vehicle service and supply industries will continue to decline, the opportunities for analytics-based jobs will grow dramatically. Recent evidence suggests that there are over one million analytics and big data job openings that are presently going unfilled, many of which are associated with the AV industry.
These same AI-based technologies are expected to transform the trucking industry in the coming years. Since freight trucks mostly travel predictable long-distance routes, complete autonomy—no driver at all—may become more commonplace in 18-wheelers well before it comes to the vehicle in your garage. One of the first innovations the public is likely to notice is called “platooning.” This control strategy allows trucks to stay closer to each other on long journeys with the lead truck surveying the road ahead and controlling the throttle and brakes of the other vehicles. As distances between trucks are reduced, fuel consumption—composing about 40 percent of the total costs associated with many trucking operations—drops by as much as 10 percent. Highway safety is also enhanced using this integrated driving approach. The Netherlands successfully tested platooning in 2016. Several of the world’s leading truck manufacturers plan to demonstrate platooning in the United States over the next several years with control systems being provided by US-based Peloton Technology. These technological developments lead to the inevitable question—What about robotic pilots for commercial aircraft? In many respects, the technology required to fly a plane is considerably less demanding than for driving a car! With the money saved by replacing human pilots with robots, will the airlines bring back peanuts?
You’ve got to think about big things while you’re doing small things,
so that the small things go in the right direction! ― Alvin Tofler