Peloton CEO Talks Level 4 ‘Automated Following’

August 1, 2019

Ryan Gehm

Demonstration of Peloton's Level 4 "automated following" platooning system. (Peloton)

Daimler Trucks abruptly hit the brakes on its platooning activities, announcing at CES 2019 in January that analysis of its years-long testing on U.S. roads showed no business case for truck platoons due to less-than-expected fuel savings and difficulty in establishing and maintaining platoon formations. Silicon Valley-based Peloton Technology, on the other hand, is hitting the accelerator, unveiling in July the development of a new SAE Level 4 (L4) Automated Following solution.

The advanced platooning system uses vehicle-to-vehicle (V2V) communications and radar-based active braking systems, combined with vehicle control algorithms, to connect a fully-automated follow truck with a driver-controlled lead truck. The company’s current Level 1 system, PlatoonPro, has a driver in control in both the lead and follow trucks. Peloton’s CEO Josh Switkes recently spoke with SAE Truck & Off-Highway Engineering magazine to share details of its new approach, which he believes will obviate many of the technical challenges facing “standalone” (i.e., single vehicle) L4 systems, and to explain why platooning does indeed have a promising future.

Can you explain Peloton’s approach to Level 4 automation?

Inside the following truck of a two-truck platoon demonstrating Peloton's SAE Level 4 automated system. (Peloton)

Most of the other companies out there are going for what we call ‘standalone driverless,’ to differentiate between following another vehicle [in a platoon]. But more and more people in the industry and the public are realizing that it’s not going to be like flipping a light switch when you can pull the safety drivers out. If you look at Waymo as an example on the passenger car side, they’re starting in Phoenix at low speeds during the day in a geofenced area in good weather. When they can pull the safety drivers out there, it doesn’t mean suddenly they’ll be able to do so at night and outside that area and in bad weather and at higher speeds. I think everyone’s acknowledging now that L4 is going to be a long, gradual rollout to more operating conditions.

At Peloton, we’ve taken a very different approach. Our Level 1 platooning system has built a foundation for us to develop L4 automated following, which is now in prototype form. The L4 we need for that following truck is dramatically simpler than what you need for a standalone L4 truck. What this means is we can deploy it earlier, so pull the safety drivers out sooner, but maybe more importantly, it means we can much more rapidly broaden those deployments to a broader set of operating conditions.

Our goal—the same as the L4 standalone companies—is to provide as many driverless miles as possible. But our basic strategy is, rather than trying to get 100% driverless miles on a single truck, we get 50% on a pair of trucks. We’re saying we can provide you safety and fuel savings today, labor savings tomorrow, and increasing labor savings over time.

What’s needed to go from Level 1 to the L4 system?

The way we’ve approached this is let’s leverage what we can from the Level 1 system, so we’ve based the hardware on the PlatoonPro hardware. We’re starting with the same sensor suite, but we’re adding some additional sensors, some additional hardware. On the software side, again we started from the same software base, the same communications, and go from there...In terms of steering, a non-redundant steering system is fine for lane-keeping assistance where if it fails, we just don’t have assistance anymore. But for L4, you need the right level of redundancy. So that is increasingly important for us.

What are some of the main challenges moving forward?

Josh Switkes, CEO of Peloton Technology. (Peloton)

On the regulatory side, there’s been a lot of good progress at the federal level and the state level in allowing at least testing and in some cases deployment of automated trucks. Initially running with safety drivers, every state where we have platooning approval [22 states currently, with 8 more authorizing testing and initial deployments] we can test L4 following. As we roll out the no-driver-in-the-follow-truck version, there will be a lot of scrutiny from safety groups, state regulators and federal regulators. One of our approaches here is to show that we know how to do safety engineering, we know how to safely test and deploy—show all that with the Level 1 system and then leverage that as we move on to Level 4.

On the technical side, you need to have the right functionality. We’ve found using the ISO 26262 approach [for functional safety] that we used for PlatoonPro has already been powerful for Level 4. We’re certainly not done with it, but it’s been a valuable start. Level 4 is much more difficult than Level 1—you need more redundancy and more reliability. There are more situations the system has to handle and so there’s certainly more work to do there.

What’s your response to Daimler saying there’s no business case for platooning?

Of course, they were testing their system, not ours. We’ve been seeing virtually identical fuel savings results in the field, in customer’s hands, to what we’ve seen on the test track [which is often more than 7% fuel savings]. We don’t know the details of their systems, but it’s clear if they’re not getting the results there was some gap in the system they were testing.

One of the possible reasons for that may be that a little over a year ago, we had a number of patents issued that are very strong blocking nets for platooning. It’s not that we want to keep platooning for ourselves—we want to bring it to the industry—but we want to protect what we’ve developed and built.

Somewhat similarly on the [challenge of] getting two trucks together to platoon, early on we identified the need to coordinate trucks to help them find platooning opportunities—what we call our network operations cloud. If trucks are out on the road and they’re a mile apart, it identifies whether they should speed up or slow down to rendezvous and start platooning, or it could be [determined] ahead of time. We’ve also found a lot of trucking operations that already have multiple trucks together…So we’ve got plenty of excited customers for platooning, many of which are at least in part Daimler customers, and we’ll continue to deploy with them and show the good results that we’ve been seeing.