On December 13, 2016, the National Highway Traffic Safety Administration published a vehicle-to-vehicle communication Notice of Proposed Rulemaking (NPRM) to require wireless communication technology on all new vehicles.
“Using V2V technology, vehicles ranging from cars to trucks and buses to trains could one day be able to communicate important safety and mobility information to one another that can help save lives, prevent injuries, ease traffic congestion, and improve the environment,” NHTSA said in a release.
The proposed rule would require automakers to include V2V technologies in all new light-duty vehicles. The rule proposes requiring V2V devices to “speak the same language” through standardized messaging. NHTSA estimates that safety applications enabled by V2V and V2I could eliminate or mitigate the severity of up to 80% of non-impaired crashes, including crashes at intersections or while changing lanes.
V2V devices would use the dedicated short range communications (DSRC) to transmit data, such as location, direction, and speed, to nearby vehicles. That data would be updated and broadcast up to 10 times per second to nearby vehicles, and using that information, V2V-equipped vehicles can identify risks and provide warnings to drivers to avoid imminent crashes. Vehicles that contain automated driving functions—such as automatic emergency braking and adaptive cruise control—could also benefit from the use of V2V data to better avoid or reduce the consequences of crashes.
“After NHTSA handles all comments, we’ll see whether or not this happens,” said Mike Lukuc, a research scientist for the System Reliability Division of Connected and Automated Transportation for the Texas A&M Transportation Institute (TTI). “If it happens, this will be followed by an NPRM for heavy vehicles and buses and commercial vehicles—which is in your wheelhouse.”
Texas A&M’s research focus areas include intelligent infrastructure, connected and automated vehicles, and human factors. TTI is predominantly an infrastructure-focused organization.
Lukuc previously worked with NHTSA as a leader for the V2V Communications Safety Program that fed the NPRM that came out last December.
In his presentation, “The Evolution of Freight Movement Through Connectivity and Automation,” Lukuc addressed a number of myths:
Myth: The roads need to be changed to make automated driving possible.
Lukuc: “Most of the OEMs say they can handle everything on the vehicles. What that means to you folks in the commercial vehicle industry and trailers is that they’re expecting sensors all around that trailer to be able to sense all the vehicles and surrounding conditions. The industry thinks roads don’t need to be changed. TTI says ‘not necessarily,’ but we could provide a better solution and relieve the burden on OEMs.”
Myth: All vehicles need to be connected to enable automation.
Lukuc: “It’s not really true. You see Google’s Waymo running around. You see Uber with automated vehicles. But they can’t handle all situations and they’re running in some pretty controlled environments. All of us would like to believe it’s wrong. Connected vehicles allow us to handle some of the conditions. It’s just another piece of information. You get a 360-degree view—about 300 meters of coverage. It’s a pretty powerful sensor. It enables a lot of capability.”
Myth: Automated vehicles are prohibitively expensive.
Lukuc: “Some say it’s wrong. The interesting thing is that just recently, 18 automakers joined with NHTSA to decide to put up radar bays for collision avoidance on all new vehicles. And that’s without NHTSA rulemaking. We’ll see on light vehicles if that technology starts to proliferate on all makes and models and lower end vehicles. As more sensors are out there, the technology improves and the cost comes down and becomes more cost competitive. So it’s not really necessarily going to be cost prohibitive.”
Myth: New laws and regulations are needed to enable automated vehicles.
Lukuc: “Not necessarily. You already have automated vehicles running around today. There’s been control radar-based adaptive cruise control on light vehicles since the 1990s. Now you see camera-based technologies bringing the cost down. Until you see fully automated vehicles hit the street and a lot of publicity on it, the regulators really didn’t get involved. The industry has been testing stuff on public roads since there’s been an industry.”
He said the promise of connected and automated truck crashes is seen through the potential of a greatly reduced number of crashes for trucks and cars.
“Reducing congestion is a big thing,” he said. “The reality is, the number of cars is increasing. The roads are not being expanded. So you need to somehow improve throughput.
“Human error is really the cause of most crashes: inattention, drowsiness, having the wrong reaction to some sort of circumstance. Automated driving fixes the major design flaw in trucks: human operation. It frees truck drivers to become something a little different—maybe ‘logistics managers.’ It enhances productivity and vastly reduces the labor component of road transport. One of the interesting things is the discussion about whether or not automated driving and the ability for the driver to disengage could somehow factor into some sort of relief of hours of service.
“Automated driving is different from crash avoidance. Automated driving is the evolution of the technology of crash avoidance, but it’s not the same. Crash avoidance technology intervenes when things go wrong. Automated driving technology automates the things we do right.”
He said the Society of Automotive Engineers (SAE) and NHTSA have agreed on the levels of automation. Here’s a simplified look:
• Level 0: hands, feet, eyes, and brain ON.
• Level 1: hands or feet OFF.
• Level 2: hands and feet OFF, eyes/brain ON.
• Level 3: hands, feet, eyes OFF, brain ON. “But the driver is starting to become disengaged. This is where it really becomes a problem from a human factor perspective as to how to bring that driver back into the loop, if something’s going wrong, in a timely manner. You hear a lot of this. Ford has decided to skip this level and go straight to a highway automated driver. Google and Waymo did that same thing. It’s a challenge.”
• Level 4: hands, feet, eyes, brain OFF.
“Constrained environments. This is where the vehicle is doing everything. Envision a managed lane that’s dedicated to automated vehicles or allows automated vehicles mixed with other human-driven vehicles. In this condition, the driver is completely disengaged. If something goes wrong, the vehicle can either handle the situation or pulls over. It does not make a decision that’s an ethical decision. You hear that a lot in discussions. The industry will never decide between hitting a car and a person or a dog. It’s going to handle all those conditions.”
• Level 5: hands, feet, eyes, brain OFF.
“Unconstrained. The vehicle does everything everywhere. This is what Google started out doing. It’s tough to do that, especially when you’re doing it with human-driven vehicles. The sensor technology and processing technology are not there to handle everything at this point in time.”
Use cases for heavy trucks:
“On-road, there’s a benefit for fuel economy with driver assistive truck platooning,” he said. “This is not commercialized yet, but it’s on the way. Driver productivity is helped by traffic jam assist, highway pilot, automated movement in queue, automated trailer backing, and parcel delivery automation.
“In constrained environments, this is where automation becomes more of a reality to most of us. It’s pretty easy for us to make technology work in a constrained environment and make it work safely. You don’t have to deal with as many institutional policy issues, at least at the state and federal level. We see this as probably the likely area for highly automated technologies to be rolled out in relation to things like mine hauling, drayage, and dispersed local sites for manufacturing and distribution.”
He said trailers as AVs are “not a very practical near-term solution, particularly with the number of trailers out there—the burden on the trailers and keeping things up to par from a compliance and certification and performance standpoint. I would think most of you don’t want to see that path in the near term due to cost.”
He said state-of-the-art technology has been developed in tandem by the Volvo Group and California PATH (Partners for Advanced Transportation Technology) at the University of California-Berkeley. Volvo provides the vehicles and an on-site engineer to work with PATH researchers on the technology.
On-road testing is underway in California. Three Volvo trucks have traveled along Southern California’s busy 110 Freeway hauling cargo containers. They typically are just 50 feet apart, controlled by robots in two of the vehicles. The pack is controlled by a “captain” in the cab of the front vehicle, and the others follow with their own drivers—or autonomously eventually.
“It’s more related to corridors near ports, which makes a lot of sense,” he said.
In the near term, he said the elements of driver assistive truck platooning are:
• Two trucks following safely at close following distances (50 to 80 feet).
• Combines vehicle-to-vehicle communications with radar. “That allows the following truck to brake simultaneous with the lead truck. They’ll have to sense the truck slowing down before the second vehicle starts to slow down. You know from a vehicle communication if the brakes had been applied pretty much in real time—about 1/100 of a second.”
• Research has touted and fed lab tests have shown significant fuel savings due to aerodynamics: 4% to 5% for the front truck and 8% to 10% for the rear truck in two-truck platooning. “This is really without all the aerodynamic packages you have available for trailers. Our intent is to eventually assess not just a standard cab and trailer without aero for two vehicles for the same thing, but go to iterations of testing. What’s the benefit with aerodynamics on there? Is it additive? But right now, we know you can see 10% for the trailing vehicle.”
• The following truck driver is still responsible for steering. “We see that as a problem because now you have two trucks following a lot more closely. The field of view of the trailing driver is really impeded by the truck in front of him. He’s staring at the back end of that truck and isn’t really aware of the surroundings. We have some research that looks at whether that creates strain on the driver. We haven’t finished that yet but that’s our hypothesis.”
• Driver selectable gaps could also vary with frequent entry/exit, traffic, and weather.
Other platooning research/pilots/demo’s:
• Peloton Technology has been working on Level 1 on-road testing since 2014, with commercial introduction expected in 2017. “They try to broker the cost savings between the trailing vehicle and the lead vehicle. That means they expect the trailing vehicle to cut a check to the lead vehicle to make up for getting better fuel economy out of this platoon, assuming they’re not from the same commercial vehicle operator. It’s a hard sell to the industry, I’m sure.”
• FHWA-funded Auburn University testing. “There are demo’s this year. They were doing a lot of fuel testing and understanding the aerodynamics and fuel specs for Level 1.”
• United Kingdom. There is a platooning pilot starting this year.
• Florida. There is a platooning study done by the DOT and State Police, with pilot testing in 2017-18. “They’re looking for a partner that has the technology.”
He said the heavy trucks’ AV estimated timing is “ambitious.”
The timeline for truck platooning:
• 2017: Level 1.
• 2020: Level 2.
• 2025: No rear driver.
The timeline for independent operation, with one vehicle working alone with some sort of automation features:
• 2018: traffic jam assist.
• 2020: Level 2 highway pilot.
• 2020: Level 4 highway pilot.
• 2025: Level 5 highway pilot.
“Level 4 is probably more of a reality because it can be in a managed lane and in a controlled environment,” he said. “But it’s going to be close to other vehicles and it’s going to be heavily loaded commercial vehicles, so there are some safety risks there. I’d say 2025 is too soon to tap all of the federal and state policy and institutional issues for Level 5.” ♦