We know that AI is playing a role in cutting-edge transportation systems, but AI's role today is likely not the same as its role in five, ten, or twenty years. These industry insiders shared the trends they see most shaping AI's use in the transportation sector. Here's what they said:
1. Brendan P. Keegan, CEO of Merchants Fleet
“Facial recognition technology is starting to take off. The 2019 Subaru Forester offers what they call DriverFocus safety technology, which uses facial recognition to alert drivers if they are distracted or drowsy. Distracted driving is a significant safety issue, and this is one of the first vehicles to proactively address it. Beyond just cars, we’re starting to hear more about how facial recognition is going to become increasingly important to safety and rider authentication for mass transportation methods as well.”
2. Tina Quigley, CEO of the Regional Transportation Commission of Southern Nevada (RTC)
“The #1 trend shaping AI in transportation this year is how AI technology providers will continue to partner with transportation agencies across the country that are installing smart, connected infrastructure to prepare for autonomous and connected vehicles. This trend is a testament to collaboration equaling innovation.”
3. Bryce Johnstone, Automotive Segment Director at Imagination Technologies
“The two main challenges to AI adoption in transportation are government legislation and acceptance. There are still a lot of answered questions when it comes to legislation. For example, what are the laws around issues like non-autonomous and autonomous vehicle crashes? What will the laws be on insurance? How will governments handle the transition between non-autonomous and autonomous vehicles?
When it comes to acceptance, Millennials will most likely be comfortable with using unmanned Robotaxis where all driving decision is removed from the passenger. Older generations will probably not be as accepting of such AI-based technology and could potentially slow down the adoption.”
4. Hongmo Je, CTO of StradVision
“Cost. More than half (53%) of global business and IT leaders cited the high costs associated with AI technology as a major deterrent to adoption, according to a survey conducted by MIT Technology Review. This is why StradVision is sticking to the fundamentals and drives down on powerful deep learning-based embedded perception algorithms that can be used with platforms cheaper than a smartphone.
Take truck platooning, for example, it requires a higher level of automation, yet most recent cases are only up to level 2, limited to steering, acceleration, and deceleration by the vehicle itself. However, there’s a growing desire for highly automated trucks (level 4) for safer roads and more sustainable transport solution – StradVision can deliver this safely, and in an affordable manner.”
5. Alex Shartsis, founder and CEO of Perfect Price
“This year the focus is on the promise of the internet of things (IoT) and sensor data. IoT has the potential to reduce latency to next to nothing, making connected cars and autonomous vehicles safer and laying the foundation for concepts like connected highways. Eventually, transportation companies will be able to allocate resources and plan routes using AI.”
6. Matt Hill, Chief Science Officer at Rekor Systems
“The smart city initiative is trending across metropolitan cities. Traffic and emissions are at an all-time high and municipalities are focused on leveraging technology to reduce these common issues. Congestion management is an area where governments can intervene to deter vehicle travel during peak hours, as well as encourage carpooling and other means of transportation, e.g. a city implementing high-resolution cameras to monitor vehicle traffic within the city. By pairing AI-based software with this hardware, cities can identify vehicles in real-time and enforce traffic management fees based on vehicle type, size, and number of occupants.”
7. Joseph Notaro, Vice President, WW Automotive Strategy & Business Development, ON Semiconductor
“By definition AI is not a deterministic function. Any function in the car needs to go through rigorous validation. These systems have traditionally been deterministic: given a series of inputs the functions were designed to generate a series of outputs. I.e. deploy an airbag if a certain level of acceleration is detected; turn the front wheels right if the steering wheel is rotated clockwise; etc. Validation is a very long process to ensure that in all conditions (cold, hot, humid, dry, high battery voltage, low battery voltage, short-circuit, etc.) given a series of inputs, the output is the expected one.
AI-based systems are not deterministic. AI-based systems are ‘trained’ to perform certain functions. Similar to humans, though, AI-based systems need to react to unknown situations. The decisions they come to are based on how that particular system was trained. So the validation process needs to be reconsidered.
There are also ethical, legal and liability ramifications linked to AI-based vehicles (Level 4 -5 Autonomous Driving). Who is responsible for validating these vehicles (OEM? States? Federal Government?)? Who is responsible for the decisions the vehicle makes?”
8. John Barrus, Director of Business Development at Groq
“The #1 challenge for AI adoption in transportation will be security. AI hardware and software systems in transportation are very complex and depend on remote connectivity. It’s very difficult to lock down the mosaic of complicated hardware and software in a way that prevents network infiltration and sabotage.”
9. Juan Rodriguez, co-founder and CEO of FlashParking
“Gathering real-time data from different types of IoT connected devices in and around the garage asset allows for the repurposing of parking spots according to the mobility needs of the city. This allows you to drive Yield Management decision making and adjust variable pricing based on the spots repurposed use.”
10. Sean Pour, co-founder of SellMax
“In my opinion, it’s the morality issue with AI in cars. If there are three people walking and you can crash into them, or injure the driver badly what should we do? These are all considerations the programmer must consider. So, we are getting into the ethics of AI and people are thinking about this more and more.”
Share your feedback and we'll consider adding it to the piece!
Have expert insights to add to this article?
Share your feedback and we'll consider adding it to the piece!ADD YOUR INSIGHTS