“At the turn of the 21st century, transportation professionals face challenges of increasing complexity. Transportation professionals are asked to meet the goals of providing safe, efficient, and reliable transportation while minimizing the impact on the environment and communities. This has turned out to be quite difficult given the constant increase in travel demand, fuelled by economic development, and the ever-growing demands to do more with less.” — Adel W. Sadek (“Artificial Intelligence Applications in Transportation”, Transportation Research Circular, 2007)
In 2016, according to the National Safety Council, approximately 40,000 people died due to vehicle crashes in the US alone with property damage having a hefty price tag of more than $430 billion. Those costs include losses in wages and productivity, medical expenses, property damage, employer costs, and administrative expenses. Unsurprisingly, a vast majority of these collisions can be attributed to driver error, which is why companies are investing a significant amount of capital into exploring the possibility of driver-less vehicular transport. An experimental driverless truck made by Otto, an Uber-owned company, delivered 50,000 cans of beer after traveling 120 miles at 55 mph from Fort Collins, Colorado to Colorado Springs is strong evidence that if implemented, AI and machine learning can drastically reduce both financial costs and bodily harm related with vehicle crashes.
The video above shows how Tesla’s Autopilot predicts collisions and potential collisions with accuracy, prompting the driver while activating the vehicles safety brakes. AI technologies such as this are quickly ushering in a new era of vehicular travel, one that is safe and cutting edge. According to Business Insider, 10 million self-driving cars will be on the roads in the very near future, by 2020 in fact.
The idea of a self-driving car is not a new one. Major organizations like BMW, Mercedes, and Tesla have been hard at work developing cars, either possessing features that allow the car to behave autonomously without the input of a driver (but within a set of defined parameters) or are yet to release them. Google made headlines in 2009 with Waymo, an ambitious project to develop a fully autonomous car by 2020, though it is facing stiff competition from other organizations like Uber and Tesla in its quest to automate vehicular transportation.
Canadian start-up X-Matik is actively working in developing tools that will make any car self-driving. Its beta product LaneCruise is expected to roll out in Spring 2018 and allows any car to navigate busy roadways without a driver’s input, with a few exceptions that do require human input.
But what allows these self driving cars to navigate the ever changing landscape of major roadways? Machine learning, a component of artificial intelligence that allows computers to learn from the data it receives and respond accordingly, will allow self driving cars to respond to the variability of road travel.
According to a report published by Seattle-based INRIX and the Centre for Economics and Business Research, traffic congestion will cost Americans $186 billion a year by 2030. The figure was calculated using both direct and indirect costs, with direct costs being variables like time wasted and fuel consumed while idling, whereas indirect costs accounted for variables such as environmental cost and the cost of conducting business in a traffic congested city. Such costs can be minimized by implementing a city-wide AI system. This very system can employ a machine learning algorithm, which utilizes a city wide grid of surveillance cameras overlooking major roadways. This will allow cities to detect traffic patterns and provide drivers or automated vehicles with real-time information regarding traffic flow, pedestrian movement, and accidents, overall saving valuable time, fuel, and other attached costs.
The key to making commuting pain-free and providing drivers with an unparalleled driving experience is data, and massive amounts at that, crunching through 4000GB of data a day. This data can be pulled from cars itself, which learn about frequent destinations, times of those trips and other such variables offering drivers notifications on when to depart (due to traffic conditions), weather updates, and more. Furthermore, as witnessed countless times on social media, such cars prove to be life saving, detecting crashes split seconds before they happen, activating emergency brakes, and saving passengers who otherwise would not have the foresight nor the reaction time to activate said brakes.
Although fully automated vehicles are not yet commercially available, we can easily predict their widespread adoption in the very near future. The application of AI and machine learning, getting from Point A to Point B in a safe and timely fashion, is a new frontier not yet fully explored, but one that holds the potential to become a massive component of our lives.
Written by Osman Ansari.
