For self-driving tech, weather is snow easy job

NASHVILLE, TN – Trucking is rolling into the future, and the trucks may be getting there all by themselves.

Autonomous trucking was the topic of the moment this week at the 2017 In.Sight user conference and expo in Nashville, with panelists disagreeing on what is possible and in what timeline.

While experts and industry insiders across multiple sessions debated whether trucks would be truly self-driving in just years, or advanced driver-assisted technology rolled out over a decade would be the reality instead, it was clear that we’re not quite there yet.

Development of autonomous technology still faces a lot of obstacles including, one panel noted, snowbanks.

Trimble executives spoke to conference attendees about successful testing and future plans the company is a part of in the autonomous vehicle space, but highlighted that much of those tests are being conducted in environments with typically stable weather conditions, such as deserts in California and Nevada.

The company is now working with the minds at the University of Waterloo’s center for automotive research on the challenges of over-coming driving conditions Canadians may be all too familiar with.

While human drivers already operating in snowy conditions may be used to the closures, slower traffic, slippery surfaces, and narrowed pathways the fluffy white stuff brings, it poses problems for autonomous vehicles which rely on layered technology including radar, sonar, inertial, and visual sensors to detect and adjust to road conditions.

Panelists said the vehicles use the technology to map and localize the vehicle, and snow buildup on the road surface can block sensors from determining something as simple as which lane it is travelling in, or how narrow the roadway is.

Working with a vehicle aptly named Autonomoose, Trimble – just one partner company on the project – and the team at Waterloo are using a Lincoln MKZ hybrid vehicle to learn how to navigate in poor weather conditions.

Although the testing vehicle is a car, Trimble’s team says the results will affect how trucks operate in inclement conditions, and will help to improve even level one or two driver-assisted technology, including assisted breaking and rear sensors that help with reversing.

The team at Waterloo has been working with the car, whose license plates read “UW Moose”, for three years, and presented some of their technology with Renesas Electronics America at the consumer electronics show in Las Vegas, in January 2017.

The autonomous truck discussions continued over the four-day conference, highlighting research looking at remote driver access to vehicles, platooning, and fully-autonomous highway vehicles, as well as various levels of driver-assisted technology.

Discussions were also focused on concerns about security and driver job loss. Multiple panelists reassured listeners that security is being taken very seriously, and would be taken into consideration at every stage of development. While some panelists said the technology will someday exist to make trucks driver-free, others reminded drivers that government regulations and public trust would keep them in the cab for many years to come.

Have your say

This is a moderated forum. Comments will no longer be published unless they are accompanied by a first and last name and a verifiable email address. (Today's Trucking will not publish or share the email address.) Profane language and content deemed to be libelous, racist, or threatening in nature will not be published under any circumstances.