HOW ARTIFICIAL INTELLIGENCE MAY BREAK OR REMAKE THE TRUCKING INDUSTRY
(originally published Jan. 13/2009 in the Aquinian)
The engine sputtered to a slow stop, but its echo lingered in the raw air as Lloyd Carter pocketed the keys of his tractor trailer. He gingerly slipped from behind the wheel, out the door and down the steps, in the measured manner of a man that has done it countless times.
“We were a different breed, way back when,” he said with a grunt, hopping from the last step onto his icy driveway. “I’ve been doing this for 40 years, and I still work 70 hours a week, for maybe $40 thousand annually. But if you ask a young fella today, they’re only at it for three or four weeks before they tell me ‘You guys work awfully hard for the money you make.’”
“When you’re trying to start or raise a family and you have to take 13 days to do a round-trip to California for a load of fresh fruit, you start to realize that’s not much of a life.”
Howard Li thinks that truck drivers will not only continue to change in the coming years- they’ll be rendered obsolete. He and his team of fellow engineers at UNB are researching artificial intelligence- similar but more sophisticated than the technology used to automate assembly line robots- so that our vehicles will be able to drive themselves.
“Hopefully, one day, technologies will save drivers from the tedious work of driving,” he said. “Then they will be able to do more exciting work that they like.”
Li and his team are tinkering in their lab, using photos and sensors to simulate moose, deer, pedestrians, sharp turns, and the other rough rules of the road. He said vehicles with artificial intelligence will one day use this data to navigate our highways and avoid hazards all on their own. And he said that day isn’t too far off.
According to Li, the biggest obstacle is figuring out the right algorithms to make intelligent vehicles compatible, to keep them from constantly crashing into one another.
“There aren’t thousands of vehicles at our disposal to properly test artificial intelligence compatibility,” he said. “That means we’ll have to use computer simulations of multiple cars working together, in order to write the program. That way these intelligent vehicles can drive together on our roads as one big team.”
Li said that may take years, or even a few decades, to accomplish. But Carter said it won’t take long for those days to creep up on us, especially if they mean the end of an era for snow plow operators, taxi and bus drivers, and truckers.
In fact, he said that new age is practically upon us- with automatic transmissions truck drivers don’t have to loose concentration by shifting, and with cruise control they don’t even have to touch the gas on the hills.
“The trucks are almost to the point of being driver-less already,” he said with a soft chuckle. “You just sit there now and watch everything. When you start the truck up and set your cruising speed, you really don’t do anything except watch what’s going on around you.”
“Many my age think it’s the best thing since sliced bread,” he added. “But younger guys, who like to horse around and speed up, they’re bored. And the last thing you want is for that to lead to more fatigue. We need something to alarm the drivers, because we simply don’t have allot to do anymore.”
Li said that even though fully automated cars may be years out of reach, researching them now can bring about many of their advantages today. Alarms and sensors could warn truckers of hazards on their path quicker, for instance, giving them more time to react.
“Thousands of radar based collision warning systems are already operating on heavy trucks, and lane-departure warning systems have also become available,” Li said. “But more technologies are needed for collision warning. Because fatigue plays a key role in truck accidents, we need to develop drowsy driver countermeasures for the trucking industry. We need in-vehicle systems to monitor the condition of the driver, so that we will be able to help him regain alertness.”
Carter popped the hood, and it opened with a sharp groan. He said not long ago the main mechanism underneath was the engine. But now it has to make room for what truckers call the Brain, a foot wide box of circuitry fastened to the engine’s side that monitors its every function, regulates speed, and where the drowsy driver alarms that Li described will one day be installed.
“So much technology, it’s really come along way. Us older fellows used to be allot more knowledgeable, if our truck broke down we could usually get it going again,” Carter said. But he said old school mechanics can do little to repair all the software entwined with most engines these days. And that has turned what used to be a last resort into a first.
“It makes some of us older guys embarrassed, because when your truck stops now there’s nothing you can do. Now it has to be towed to a dealership for what you could once fix yourself. It doesn’t seem like long ago that you would tinker for hours under that old hood, because the last thing you wanted was to have it towed.”
Carter said the trend of cutting edge technology edging out old jobs is nothing new. Today’s snowploughs haul two more blades than when he was young man, so now one driver can clear the width of what used to take three. Robots are now used in the mill he once hauled from, and they eliminated the jobs of at least seven workers who made their livelihood inside.
“Those robots never lost any time, never went for break , all you had to do was change the oil in them every five years,” he said. “So it looks pretty good to a big company that wants save $35 an hour for labour. The robot never complains, takes a sick day or anything. And it could be the same thing with any vehicle and any driver.”
Li said his research is not designed to put workers out of a job, but make that job easier and safer. Snowplow drivers, he said as an example, need better indications to clear the roads at low or zero visibility .
“The drivers will remain in control of the steering, and use a display indicating the lane edge to accurately guide the snowplow,” he said, adding that universities in California and Minnesota are testing these kinds of systems.
He added that workers should see technology as a solution, and not a threat.
“I think one day we will be able to develop fully automated snowploughs as well. But technologies will provide solutions to socio-economic problems, they always give us more options. Today even though we have snow blowers, there are still people using shovels.”
Carter is a little less enthusiastic, but he said the notion shouldn’t be dismissed.
“I’m fascinated with this idea for big cities, for taxis to drive on their own and buses to run almost like old street cars,” he said. “But how much time would it know to give you, so you could get on and off again? It’d be pretty scary for anyone with one foot out the door if it automatically starts up again.”
He said a computer could miss similar subtleties on a truck.
“You could probably program an automatic route for a truck right now, and it would know which exits to take and how fast to go,” he said. “But using a computer to back a 53 footer into a crowded warehouse? That might be a little hairy, I can’t see how it could do that. But if it ever could, well, that’s something I’d like to see.”
Photo by Kyle Mullin