Elon Musk informed the world in late 2017 that Tesla was taking its automotive know-how and making use of it to a completely new problem: self-driving big rigs. However one 12 months later, he positioned the Tesla Semi fourth on a list of priorities for the company, behind the upcoming Mannequin Y compact SUV and an electrical pickup truck. This week, Daimler executed a transfer a few years within the making by saying its personal large rig (albeit diesel-powered) outfitted with semi-autonomous expertise. And others are following go well with.
The German automaker additionally dedicated to manufacturing the truck this summer season, with deliveries scheduled for later this 12 months. It pledged 500 million euros over the subsequent few years to the continued growth of an autonomous large rig, and stated it has employed lots of of staff to maneuver the tech ahead. And similar to it did when it unveiled the prototype version in 2015, Daimler gave us a experience within the truck to get a style of what the close to way forward for trucking will appear to be.
Whereas there are just a few Tesla Semi prototypes on the street now, and a dozen or so big name companies have placed preorders for the vehicles, it doesn’t appear to be a manufacturing model is coming any time quickly. Tesla nonetheless hasn’t stated where or exactly when it will build the trucks, and would probably have to raise more money (or promote a hell of much more Mannequin 3s) to fund the mission.
This has left the door large open for corporations like Daimler, the guardian firm of Mercedes-Benz. Daimler introduced it was working by itself self-driving big rig in 2015 when it confirmed off a working prototype referred to as the Freightliner Inspiration Truck. The automaker went large, debuting the truck on the Hoover Dam and providing check rides at Las Vegas Motor Speedway. This week on the Client Electronics Present, Daimler returned to Las Vegas to make good on its promise with a manufacturing model of that prototype truck.
The brand new Cascadia is just not way more superior than the prototype was in 2015. In reality, the expertise continues to be fairly restricted. Daimler says it’s the primary Class eight business truck with Degree 2 autonomy (referring to the Society of Automotive Engineers’ scale for self-driving definitions), that means the motive force is in management, however is supported closely by the truck’s expertise in sure conditions. In that sense, the brand new Cascadia primarily has the identical fundamental driver help expertise many fashionable vehicles supply, together with automated lane centering, adaptive cruise management, and emergency braking.
However the brand new Cascadia is doing this with a restricted set of sensors. There’s a forward-facing digital camera, a forward-facing radar, a second radar sensor on the correct facet of the truck. That bundle pales compared to the handfuls of cameras, ultrasonic sensors, and radars you’d discover powering Autopilot, not to mention the Tesla Semi, which is meant to have a beefed-up model of this similar sensor suite.
This helps preserve prices down, however means the expertise is extra according to what you’d discover powering one thing like Nissan’s ProPilot driver assistance feature versus Autopilot, and even Audi’s supposedly Level 3 system, which makes use of comparable tech, however depends on LIDAR as properly.
Protecting with a theme of much less is extra, there’s additionally no camera-based monitoring system within the truck to verify the motive force pays consideration whereas utilizing the Degree 2 options. As a substitute, the Cascadia makes use of a system just like the one present in Tesla’s vehicles.
A sensor within the steering column measures resistance utilized to the steering wheel. If the motive force takes their arms off the wheel whereas utilizing the lane centering function, the instrument cluster will, after about 15 seconds, floor a warning that tells them to put their arms again on the wheel. If the motive force doesn’t do this, the warning adjustments from yellow to crimson. After one other 60 seconds, if the motive force nonetheless hasn’t put their arms again on the wheel, the truck will come to a cease on the facet of the street.
The brand new Cascadia is a far cry from a totally autonomous truck, however primarily based on my temporary experience, Daimler has refined the expertise in comparison with the prototype model. The prototype swayed on the freeway throughout my two-mile demo experience in 2015, ping-ponging between the lane markers. The brand new truck, in the meantime, felt locked to the middle of the lane throughout this week’s experience, which adopted the identical precise route from just a few years in the past.
A Daimler consultant additionally informed me that, whereas lane centering is on, the motive force may even select the place the system locations them within the lane. (For instance, if a driver is on a good one-lane freeway and desires to keep away from clipping oncoming visitors, they may inform the truck to hug the correct lane line.) That is one other signal that system is maturing from what debuted in 2015, although it’s a small one.
Daimler promised another fashionable applied sciences are coming the brand new Cascadia, although none of it was on show within the preproduction vehicles getting used for the demonstration. The corporate plans to supply an non-compulsory 10-inch touchscreen within the dashboard, and a 12-inch digital cluster behind the steering wheel. The truck will be capable to obtain over-the-air software program updates, too.
The Cascadia received’t be as filled with tech because the Tesla Semi, neither is it as modern. However it will likely be out there later this 12 months. Daimler has argued that bringing automation to trucking will assist squeeze higher gas effectivity out of the tens of millions of miles that its large rigs cowl yearly. It will lower the toll these miles tackle the drivers. Most significantly, it might assist scale back the some four,000 fatalities that outcome from crashes involving these huge hunks of equipment. If all goes properly, we would have a way by the top of this 12 months of whether or not any of that’s true.
Daimler wasn’t the one firm showcasing a truck with autonomous tech this week at CES. TuSimple, a startup headquartered in each China and California and backed partly by Nvidia, introduced a Navistar truck outfitted with its personal tech that founder Xiaodi Hou says can fully drive itself in restricted, rigorously mapped geographic areas with out human intervention. That is what the SAE refers to as Degree four autonomy, and it’s the identical degree of self-driving that Waymo is attempting in Arizona right now.
Whereas TuSimple didn’t show its truck at CES this week, the startup has been making Degree four check runs for over a 12 months in several areas throughout the US. It’s additionally carrying cargo on a few of these check runs for round 12 business companions — although Hou received’t say which of them simply but.
Daimler could be taking a gradual method to growing the autonomy of its vehicles, however Hou says Degree four is feasible within the close to future. “We’re assured of launching this factor each day, and it’s been operating each day. I feel that may be very large information to distinguish us from the remainder of the opposite gamers,” he says.
The important thing, in response to Hou, is just making an attempt to function in these rigorously plotted domains. “We solely function inside that area we outline. Principally we now have pre-mapped the situation and we’ve been run by way of the map and that there’s no surprises to us,” he says.
Moreover, Hou says, whereas Degree 2 techniques can have advantages, they received’t resolve one of many largest issues dealing with trucking proper now: driver shortages. If TuSimple can get a Degree four truck on the street someday within the subsequent few years, even when it’s restricted to sure areas, Hou believes eradicating the necessity for a licensed security driver will assist the business sort out this drawback. He additionally thinks that change will probably be sufficient to make TuSimple money move constructive.
“The distinction between the price of constructing a extra dependable system versus a mediocre assisted driving system is that this excessive,” Hou says, spreading his arms just a few toes aside. He then pinches his fingers collectively. “Nonetheless the worth of your return [on Level 2] is that this small.”
For what it’s price, Daimler’s head of trucking Martin Daum stated this previous week that his firm stays centered on growing extremely automated vehicles, regardless of the deal with partial automation. “We have now a transparent aim to stay pioneers,” he stated. “And we wish to put a Degree four truck on the street right here in the USA this 12 months. Keep tuned.”
What’s clear from the presence of those corporations at this 12 months’s present is that this: an enormous German automaker is able to begin augmenting how people transfer items round our nation. “We’re prepared to maneuver boldly ahead,” Daum stated. Others, like TuSimple, are nonetheless testing the outer bounds of what’s potential, whereas attempting to make a sustainable, if restricted, enterprise out of it within the quick time period.
Tesla’s identified for being a frontrunner in bringing new applied sciences to passenger vehicles. It sparked a transfer to electrical propulsion, triggered a landslide of large touchscreens, and impressed automakers to develop their very own partially automated driver help techniques.
However by the point it enters the trucking business with a business product, the corporate will probably be taking part in a little bit of catch-up, at the very least with respect to autonomy. If partial or full autonomy turns into a distinguishing issue for corporations that promote vehicles, Tesla should persuade clients that Autopilot is healthier than tech that’s already on the street (all whereas combating a separate battle: convincing clients that electrical propulsion is a viable choice on the planet of hauling freight). That’s potential, even perhaps probably. It’s simply not precisely typical for Tesla.
Images by Sean O’Kane