The transportation industry has emerged as a primary focal point for artificial intelligence (AI) thanks to the all the research being poured into autonomous vehicles. But while millions of dollars continue to be poured into autonomous vehicle research, the day when billions of vehicles drive themselves on the streets of major cities is still very far off indeed. That doesn’t mean investments in AI as it applies transportation are a waste of time and money, though. In fact, just about every aspect of transportation will soon be transformed by digital assistants infused with machine and deep learning algorithms that will automate a wide range of tasks inside and out of the vehicle.
Just about every major IT vendor from Apple to Google is researching autonomous vehicles. Intel plopped down $15.3 billion to acquire MobileEye, a provider of automotive safety software that has been researching autonomous vehicles. MobileEye earlier this year announced it has signed a pact with a vehicle manufacturer in Europe to install software to automate a variety of driving tasks sometime after 2021 using EyeQ5 processors Intel has under development.
During an Intel Data Centric Innovation Summit earlier this month, Navin Shenoy, executive vice president and general manager of the Data Center Group at Intel, described how Intel is investing heavily in computer vision algorithms and high-definition mapping software to enable vehicles to understand where they are going. As part of that effort, Intel is employing two million vehicles to gather data.
“The car is the greatest data collector in the world,” says Shenoy.
Shenoy says Intel is investing in autonomous vehicle research in part because each vehicle, on average, generates 4TB of data per hour to be analyzed. Some estimates suggest vehicles may need to be able to process and share as much as 25 to 40 GB per second.
But while computer vision will play a key role in transportation, there may never be the level of intelligence required to process all that data fast enough to enable autonomous vehicles to safely run at scale, says Mike Redding, managing director for Accenture Ventures. That doesn’t mean Accenture isn’t keen to apply computer industry expertise within the transportation sector. Accenture just made a strategic investment in Malong Technologies to gain access to computer vision software based on machine learning algorithms. But replicating what can be achieved in a controlled environment is very different than the conditions human drivers deal with every day on the road, says Redding.
“There’s a real difference between a clear, sunny day and a cloudy day when it’s raining,” says Redding.
On a perfectly sunny day, it’s possible for algorithms to drive a car from one destination to another. But on a cloudy day with limited visibility, those same algorithms are not going to be able to correlate the fact that a bouncing soccer ball in the middle of the street is indicative of a child nearby, says Redding. Because of those issues, Redding says that for the foreseeable future, algorithms will employed to augment human drivers by, for example, enabling them to detect potential hazards such as the presence of gases that would otherwise be invisible to the naked eye.
Transportation companies are naturally tracking all these developments. For now, however, most of them are focused on how to insert digital assistant technologies such as Alexa from Amazon into their business processes. For example, P&S Transportation, a leading national trucking company, is now using Alexa to enable managers to employ voice commands to more easily access fleet data and track key performance indicators (KPIs) using a fleet management application developed by Omnitracs.
As managers become more comfortable with that capability, the expectation is that the company will want to provide its drivers with similar voice query capabilities, says Mauricio Paredes, vice president of technology at P&S Transportation.
Over time, the predictive analytics software embedded within the Omnitracs fleet management applications will be increasingly infused with machine learning algorithms that drivers will then be able to invoke using voice commands from within the cab of the truck to optimize their route, based on both traffic patterns and which customers are the most valuable to P&S Transportation, says Paredes. That capability will have to be developed by embedding digital assistants into the trucks.
“The driver can’t have a phone in their hand,” says Paredes.
P&S Transportation also expects to be able to monitor vital signs of the driver, adds Paredes.
In general, most companies will follow a similar path to embedding AI within their business processes, says Monte Zweben, CEO of Splice Machines, a provider of a data warehouse and analytics application infused with machine learning algorithms.
While it’s doubtful that truly autonomous vehicles will ever be achieved at scale, it’s already clear that the driving experience for humans will be significantly enhanced by relying on algorithms capable to processing massive amounts of data at scale, says Zweben. Those experiences will include everything from something as simple as being made aware of special deals being offered near the driver’s location to optimizing routes based on weather data gathered from external sources, says Zweben.
Once that becomes widely deployed, businesses will then move to reengineer a much wider variety of backend business processes because they will be able to correlate data from internal and external data sources in real time. The ultimate success of those efforts will have a lot more to do with how the organization employs that data rather than the underlying AI technology.
“The real value is not going to come from some fancy, schmancy algorithm,” says Zweben.
It’s not clear to what degree vehicles will become autonomous. Waymo, a unit of Alphabet, the parent company of Google, plans to soon launch a ride-hailing service using driverless minivans. Delphi, a provider of automotive parts, recently spent $450 million to acquire nuTonomy, a startup company focused on self-driving vehicles. Just about every automotive manufacturer CEO has made grand predictions concerning the future of autonomous vehicles. At the same time, however, Uber’s attempt to create a fleet of self-driving cars is in doubt following the death of a pedestrian in Arizona and there are concerns about the viability of Tesla’s Autopilot driver assistance system after a fatal crash.
Most transportation companies are likely to prefer to wait and see how all these autonomous vehicle ventures progress before making any firm decisions concerning how they employ drivers. But as far as relying on AI to enhance the productivity and safety of drivers is concerned, the future is right around the next proverbial corner.