The world is currently experiencing a technological revolution almost beyond comprehension. Every industry is in the process of an ongoing digital evolution, and this change is becoming especially evident in the transportation sector. Transportation is no longer just the physical infrastructure of streets and signs – today, the transportation system includes layers of digital infrastructure that link with the physical. Now, with the recent technological breakthrough of artificial intelligence, or AI, the digitalization process is quickening. AI has gained popularity but also faces concerns, and as the transportation industry works to keep up with the demands of a constantly growing and moving population, experts look to harness the benefits of this technology.  

History 

Alan Turing first published a paper titled Computing Machinery and Intelligence in 1950, earning him the title of “The Father of Artificial Intelligence.” At the time of publication, computers could execute commands, but they were not able to remember the information and therefore learn from it. Now, AI has the ability to gather and analyze mass amounts of information with speed and efficiency never seen before, while constantly learning and adapting to new information. The early 2000s was a seminal stage for AI in transportation, as laboratories were producing autonomous car prototypes. The period between 2004-2012 then brought extreme advances in sensing technology and machine learning, ringing in the current era in which AI can be found in almost every modern transportation technology company’s toolkit.  

Current Usage 

One of the most visible developments in AI is the advent of Advanced Driver Assistance Systems (ADAS). These tools provide a second layer of safety as they monitor items that require driver attention and alert the driver when they may miss something. AI continues to advance in the movement of people as well, as transportation system users can now catch a ride in a fully autonomous Waymo taxi in San Francisco, Los Angeles, and Phoenix. However, though much excitement has been generated about the more tangible implementation of AI in transportation, there has been a strong undercurrent to engage with this technology to achieve industry goals. Organizations have been utilizing AI as a tool for a few major priorities, mainly to increase safety, reduce the harmful impact of transportation on the environment, decrease traffic congestion, and lower resource strains for companies, consumers, and public agencies.  

Safety 

Safety has always been a fundamental motivation of the transportation sector. Now armed with the technical toolkit of AI, organizations are deploying AI to decrease the number of fatalities on our roads, both for vehicle users and pedestrians. One area of concern within road safety is the length of time it takes emergency personnel to respond to crashes. The U.S. Department of Transportation (USDOT) has noted that the amount of time it takes emergency responders to arrive on scene directly impacts the survival of an injured individual. Rekor has utilized AI to assist in streamlining the emergency response for roadway crashes. They have created tools that detect crashes faster than the traditional method of receiving a 9-1-1 call, automatically alerting first responders and cutting response time by an average of 9 to 10 minutes in one case study in Nevada. 

However, even before a crash happens, AI is being proactively utilized to prevent crashes. An AI platform created by Derq produces predictive analytics to help cities appropriately address safety concerns. This tool is currently in use at intersections in Osceola County, Florida. Using footage from traffic cameras installed at intersections, AI algorithms provide insight on movement and dangerous scenarios, giving officials actionable data to implement safety countermeasures and data-driven infrastructure solutions.  

Traffic Efficiency 

AI has the capability to increase general traffic efficiency by reducing congestion, decreasing commute times, and streamlining travel efforts. This technology can be harnessed to better analyze and manage traffic patterns, reducing the time, effort, and cost that currently burdens public agencies. AI is being integrated into vehicle-to-everything, or V2X, systems that are the fundamental basis for connected transportation and infrastructure. The company Flow Labs has created a tool that can optimize traffic signal timing by generating efficient timing plans at lightning speed, integrating directly with traffic signal controllers for updates. The speed at which this tool operates offers a heightened safety environment. Traffic efficiency and safety spans both urban and rural areas, as the Yakama Nation in rural Washington has implemented AI-powered roadside units (RSU) at an intersection in which the highway meets a local road. The RSUs are equipped with multi-sensing, computing, and communication tools, allowing for ideal traffic monitoring and event detection, improving safety in a community that faces a disproportionate traffic fatality rate. 

Energy Efficiency  

With AI powered tools reducing congestion on busy roads and intersections, internal combustion vehicles may spend less time on the road, therefore reducing tailpipe carbon emissions. Since the transportation industry is the largest contributor to greenhouse gases in America, at about 28%, there is a need to mitigate the industry’s harmful impact on the environment and consumers.  

However, AI itself can be a major strain on energy resources and have a carbon footprint. AI models need to be trained so they can build their “intelligence” framework, and training for even just one AI model can consume thousands of megawatt hours of electricity and emit hundreds of tons of carbon. Experts expect these environmental impacts will continue to escalate in correlation with the boom of AI and will need be addressed as the technology becomes more integrated into major industries. 

Economic Efficiency 

Reducing costs for consumers is often an industry priority when developing and deploying any modern technology, and this has been a major part of the AI discussion. Training a technology to be able to engage in human-like intelligent tasks can have many resource-related benefits, such as reducing the amount of work time that real humans commit to various tasks, therefore increasing productivity in critical areas and allowing the workforce to focus on exciting advances. For example, data indicates that it takes about 70 manual human hours to retime one intersection. AI is able to consume and analyze data to create actionable insight in almost the blink of an eye, reducing the need for long manual hours retiming traffic lights. These resources, both human and economic, can be used more efficiently elsewhere in safety-critical areas. Additionally, with cars spending less time on congested roads, it can put money back in drivers’ pockets through fuel savings. Companies are also looking into automated shipping systems to reduce general costs for drivers, shipping time, and vehicle wear and tear.  

However, there are workforce challenges as AI continues to move into the automation space and concerns that AI will replace the human workforce.  In industries across the nation, we see jobs go unfilled because the needs of the workforce outpace our ability to grow hire, retain, train, and educate the workforce. AI tools can help us better plan for the future and the growth of our workforce, while we upskill workers for new, higher-paying jobs and improve our technical workforce capacity to keep our transportation system running safely.  

State of Progress 

In general, the goals of AI in transportation typically revolve around matters of efficiency, whether in time, cost, effort, or capability. However, though the technology has already taken off in the industry, hurdles to progress still exist. The time, cost, and effort needed to reach some of these large-scale goals, such as entirely “smart” cities, is no small matter. The existing need to update and maintain physical infrastructure outpaces available resources, but that should not prevent DOTs from also investing in technology. While this technology does require an initial investment, we see that leads to increased roadway safety and addresses the fatality crisis, and states are also seeing greater returns on investment as they increase efficiency, expand mobility, access, and opportunity all while reducing our environmental impact. And, even if economically feasible, the public may harbor some natural caution about trusting the new technology, especially one that is especially complex and futuristic. Consumers may not feel comfortable with an increased number of sensors and information sharing in smart cities and cars, as this allows for higher levels of visibility and tracking, potentially threatening privacy. These concerns are the reason ITS America has created a set of AI Principles and an AI Decoded document, as the industry can see great advancements from the introduction of more AI technology, but to do that in the right way puts a duty on all of us to ensure that along with the technology we build a sense of trust, transparency, privacy, and security that builds public faith in our AI-enabled future. 

AI continues to have a significant impact in transportation and could develop in unpredictable ways, offering exciting opportunities for the next generation of ITS.