Common Tread

Highly automated vehicles and motorcycles: Tech moves faster than regulations

Oct 20, 2017

In the less than two months since my previous stories on automated vehicles ran, the U.S. House of Representatives has passed H.R.3388, aka the “Self-Drive Act.” The bill’s primary purpose is to override the patchwork of laws passed by various states over the last few years, and ensure that the federal government, in the guise of Department of Transportation (DOT) and the National Highway Traffic Safety Administration (NHTSA) will set national standards for Highly Automated Vehicles (HAV).

The first installment of this series examined the current state of vehicle automation and asked whether it placed motorcyclists at increased risk. Part two looked forward to a future in which motorcycles will make a great leap forward in safety through the use vehicle-to-vehicle (V2V) communications technology.

Mostly, the bill just lists a series of rules the feds promise to write later. Meanwhile, manufacturers can put from 25,000 vehicles (in 2018) to 100,000 (2020) on the road as long as they’re able to show that any automated feature offers safety comparable to the average of all vehicles currently operating.

So, for example, an automated turn system would have to turn left across the path of an oncoming motorcycle no more frequently than human drivers do it. (The bill mandates systems to warn drivers that they’re leaving a child in the back seat, which results in less than 40 fatalities a year in the United States, but makes no mention of motorcycle safety, which results in over 4,000.)

To be clear: According to H.R.3388, a manufacturer could put 75,000 autonomous vehicles on the road before NHTSA even finishes writing an applicable safety standard. Then there’s the fact that H.R. 3388 has been referred to a committee in the Senate, which must also pass it and then it must be signed by the president before it becomes law.

Glossary
HAV Highly Automated Vehicle
DSRC Dedicated Short Range Channel
V2V Vehicle to Vehicle
V2I Vehicle to Infrastructure
V2X Vehicle to Whatever
C-V2V Cellular Vehicle to Vehicle
C-ITS Connected Intelligent Transportation Systems
ARAS Advanced Rider Assistance Systems
ADAS Advanced Driver Assistance Systems
BSM Basic Safety Messages

We (the United States and everyone else) are rushing forward into a brave new world of HAVs operating in Connected-Intelligent Transportation Systems (C-ITS), and the people writing the rules have pretty much told auto and truck makers, and tech companies like Google, Uber and Lyft, “You guys go ahead. We’ll write rules and standards as fast as we can. Until then, promise us that your automated systems will be at least as safe as ordinary human drivers, OK?”

I’d feel better about that if ordinary human drivers didn’t kill a dozen motorcyclists every damn day in the United States alone.

I’d feel worse about that except that in the last few months I’ve interviewed several people working to ensure that motorcycles will be able to operate safely on the roads of the future. The motorcycle industry as a whole is still largely in reaction mode, but I’m still optimistic about a future in which we’ll safely share the roads with robot cars, based on a series of conversations with these people.

  • Mark Gilbert, Electronics Engineer, and Francine Romine, Director of Communications, at University of Michigan’s Transportation Research institute
  • Yaniv Sulke, VP Marketing and head of motorcycle projects, with Autotalks
  • Rok Upelj, COO, and Miha Ernstshcheider, Director of Sales, at Smart Turn Systems
  • João Barros, CEO, of Veniam
  • Dr. Karl-Viktor Schaller, Head of Development at BMW Motorrad

They’re all actually testing and designing hardware and software, and building motorcycles that will function in a world where cars, trucks, and buses may not have drivers at all. While they’re not in complete agreement about exactly when and how we’ll safely interact with HAVs, they all agree on this much: Motorcyclists and motorcycle manufacturers should definitely not just rely on the ability of HAVs to passively detect us and avoid us. We need to proactively communicate with traffic around us using V2V systems.

UMTRI test
Testing a motorcycle V2V system at UMTRI in Ann Arbor. The car at left is warned about an oncoming motorcycle which is concealed by the van. UMTRI photo.

The world’s largest V2V research project includes motorcycles

Building functional V2V safety systems — especially systems that will work in a traffic mix of ordinary human drivers, connected vehicles with Advanced Driver Assistance Systems, and autonomous vehicles — first depends on understanding the way drivers interact with each other now, and how behaviors will change in C-ITS traffic, when drivers have better information about what other vehicles are doing around them and what road conditions prevail ahead.

The largest and longest-running real-world V2V research project was funded by the U.S. DOT, and run by the University of Michigan Transportation Research Institute, which Mark Gilbert and Francine Romine refer to by its initials, UMTRI (pronounced “Um-Tree”).

The program, Safety Pilot, equipped 3,000 vehicles in the Ann Arbor, Michigan area with Dedicated Short-Range Channel (DSRC) V2V communications systems, multi-channel data-logging including logging of all operator inputs, and multi-angle cameras. The fleet included a number of V2V-equipped motorcycles. The vehicles were loaned to Ann Arbor residents, who then drove them in daily use; the vehicles were kept deliberately unobtrusive, in order to ensure they were treated like ordinary traffic by other drivers.

Thanks to Safety Pilot, millions of real-world interactions between vehicles were logged, and that data’s available to car makers and other stakeholders. Gilbert reassured me that, despite the absence of references to motorcycles in government policy, the fact that they had motorcycles in the Safety Pilot fleet was proof that “There are people at the DOT who care about motorcycles as a part of the transportation system.”

There are already cars on the road (albeit only a few in the United States) that send and receive DSRC safety messages. So, when will motorcycles be communicating with other vehicles around them?

UMTRI test
Engineer Mark Gilbert (standing) briefs a motorcycle test rider. In this test, the motorcycle is equipped with a beacon only. The car driver gets an active warning. UMTRI photo.

Mark Gilbert
University of Michigan Transportation Research Institute engineer Mark Gilbert. UMTRI photo.
“Less than 10 years. And just like with cars, retrofit is going to be a big piece of it,” Gilbert told me. “If I’m a motorcyclist, and I know that I can spend fifty bucks or whatever it is, and mount a device on my motorcycle that broadcasts direction, speed, and heading, so that if there’s a guy up ahead looking at his phone, it will make his car aware of me, I’m going to spend that fifty bucks. Will everyone do it? No, but the potential is there to make bikes DSRC-capable. There’s no question that’s coming.”

Ready to build a V2V-capable motorcycle? There’s a chip for that.

Until recently, auto makers were certain that critical safety messages would be transmitted between vehicles on the 5.9 Gigaherz DSRC channel reserved for that purpose by the Federal Communications Commission (in the United States) way back in 1999.

At least one company, Autotalks, based in Israel, has already developed a motorcycle specific chip that will enable motorcycles to send and receive Basic Safety Messages (BSMs) on the 5.9 GHz band. Yaniv Sulkes heads that company’s motorcycle project, and like most of the people I’ve spoken to in the HAV/C-ITS field, he would much rather see the whole transportation industry standardize on that radio frequency. However, he told me, Qualcomm is one of the huge companies now pushing an alternative approach, often called Cellular-V2X (C-V2X). Rival providers from the mobile phone world see a new business opportunity in V2X and also covet that reserved bandwidth; they successfully lobbied against mandating DSRC in the recent House bill.

Sulkes explained that “cellular” is a misnomer. “It’s actually a peer-to-peer network using the same frequency as mobile phones. Yes, there’s a preliminary standard for C-V2X, but that’s all it is – there’s not even a real standard,” he said. “Meanwhile we’ve had a DSRC standard for 10 years; companies have spent a billion dollars in R&D. Cadillac already has vehicles on the road using it. Toyota’s been selling cars equipped with it since 2015 and Volkswagen has promised a mass deployment of DSRC beginning in 2019.”

Autotalks feel that highly mobile vehicles like motorcycles place an even greater emphasis on latency in communications — basically the lag time between one vehicle sending a message and another nearby one receiving and processing it. The existing 5.9 GHz technology updates every tenth of a second, and processes BSMs almost instantly.

The thing is, companies pushing C-V2X are describing vaporware. All their systems would hinge on 5G (“Fifth Generation”) technology that doesn’t even exist yet. In case you’re thinking, “Well my phone’s a 4G phone, so 5G will be coming along any day,” you should bear in mind that almost all so-called 4G phones and services are really only delivering 4G LTE (“Long-Term Evolution”). That’s the telecom industry’s way of saying, “We promise to upgrade it to 4G standards, soon.”

Autotalks system
Autotalks is an Israeli company with a motorcycle-ready V2V chipset. One likely customer: Bosch is developing a V2V system that will include a beacon transmitting location and vector information to nearby cars, and which will provide an audible warning when nearby cars present a threat. Autotalks photo.

One of the world’s largest motorcycle makers recently approached this small Israeli start-up to ask how motorcycles can be made safer in traffic that, it knows, will come to be dominated by HAVs. Motorcycle companies and major aftermarket suppliers are evaluating Autotalks’ chip.

“Yes, motorcycle companies tend to be conservative,” Sulkes told me. “But if I compare them to car makers it’s not that motorcycle makers are less innovative, but they have less room to innovate. Car makers think in terms of systems, which are often independent of the driver. In motorcycles it’s all about how it will affect the rider.”

It’s reassuring that so many of the people working on this technology recognize that what matters to us is: What will it be like to ride with it?

One of Autotalks’ key partners right now is Bosch, which has already used its motorcycle chip in a demonstration system. Bosch, of course, was one of the first companies to port automotive ABS over to the motorcycle industry, and they faced similar resistance from riders then, too.

Smart Turn Systems team
Smart Turn Systems is one of the few tech players in this area that is a pure-motorcycle play. Their mission is to ensure a place for existing motorcycles in the connected traffic ecosystem of the future. Smart Turn photo.

At EICMA next month: An aftermarket system to upload data from your existing bike to the cloud

Not long after talking to Mark Gilbert, I had a conversation that proved his “less than 10 years” timeline was quite a bit too conservative. The first aftermarket V2X system will be shown — by a Slovenian start-up called Smart Turn System — at EICMA next month.

Smart Turn System was founded by a group of very tech-savvy motorcycle nuts. They initially set out to solve a single, seemingly simple motorcycle safety problem: Riders inadvertently leaving their turn signals on, which confuses drivers.

Rok Upelj — an engineer with the name of an action-film villain — told me that Smart Turn quickly realized there was a reason every major manufacturer has tried to make a self-canceling motorcycle turn signal, and they all suck: it’s a daunting engineering problem that they only managed to solve by developing a digital device that measures, among other things, motorcycle position, heading, speed, and inclination. That hardware was matched to software — a proprietary algorithm — that allows the system to correctly identify all the normal uses of turn signals, and turn them off appropriately.

What do self-canceling signals have to do with Connected-Intelligent Transportation Systems?

It’s a two-part answer: First, the avid riders at Smart Turn System, being the tech and vehicle nerds they are, are well aware of that there will soon be a lot of HAVs in the traffic mix. They fear the day may come when motorcycles are prevented from using certain roads — or banned altogether — unless they have (at least) some kind of V2V “safety beacon.” They especially want to ensure that the motorcycles they currently own and love can be retrofit to ensure compatibility in future HAV/C-ITS traffic.

Smart Turn Systems product
Smart Turn’s first product is the world’s first really good self-canceling turn signal. As it turns out, it’s also the foundation for what will likely be the world’s first motorcycle V2V safety beacon. Smart Turn photo.

Then, they realized that their self-canceling turn signal could be turned into a safety beacon. After all, the hard part is determining the vector information. Broadcasting it to nearby vehicles with a low-power radio is easy. So, before they’d even taken their first product to market, they applied for an EU road-safety research grant, and set out to build a “black box” capable of communicating with compatible cars and trucks.

The product Smart Turn will show at EICMA will enable any modern motorcycle to connect to other equipped motorcycles and transfer data over the cellular network. The company will be analyzing data from those beta units to study the potential of systems that rely on mobile phone technology to transmit and receive critical safety messages. Like a lot of companies, Smart Turn’s hoping that a critical mass of players will settle on one standard — basically choosing between DSRC in the 5.9 GHz band or C-VTX.

Obviously, right now, when only a minute percentage of all vehicles are equipped to send and receive safety messages of any kind, there’s little direct benefit safety-wise, and little incentive to pay extra for V2V capability in a new vehicle, or to retrofit an existing one.

Virtually everyone I spoke to had ideas about what value-added services could be packaged with enhanced safety. Most of the companies involved are honest about selling the sizzle of added features, and basically giving away the steak of added safety — at least until a critical mass of V2V-capable vehicles is reached.

Smart Turn System has a bunch of ideas for functions they’ll build into the black box, to make us want it right now. It could flash your brake light when you use the engine to slow down; have an alarm system, and if that doesn’t prevent thefts, a “trace” function similar to LoJack; it could include a system that will allow you to share data with your friends so people on group rides can keep track of each other, and a function similar to BMW’s “e-call” system that will call emergency services when it realizes you’ve had a crash.

Connecting cars, trucks, buses, and motorcycles into a smart network

The outgoing Obama administration suggested a mandated phase-in of V2V safety systems in automobiles, beginning in 2019 and including all cars and light trucks by the early 2020s. But, the current government’s somewhat hamstrung by a promise to cut two existing regulations for every new one.

Even if all new vehicles do include the ability to broadcast and receive Basic Safety Messages — whether by regulation or some combination of corporate responsibility and consumer demand — the rate of fleet replacement ensures that it will be years before most vehicles are V2V-capable. But people working in the field are confident that aftermarket systems to give existing vehicles V2V functionality will be very common.

Veniam is a Portuguese company that is a thought leader in the area of peer-to-peer “mesh networks” of vehicles. It has already developed and deployed V2X systems for a number of fleet users. It came to my attention because Yamaha was an early investor — presumably because Yamaha envisions a future in which motorcycles, too, are connected.

I reached the CEO, João Barros, in Singapore. He told me that a lot of road use there is tolled, and that all vehicles carry a device functionally similar to the E-ZPass familiar to Americans who commute over toll roads and bridges. Singapore’s in the process of updating the specifications on those devices, and all the new ones have built-in V2V capability.

João Barros
Veniam’s CEO João Barros told me that his company’s work with Yamaha was confidential, but he hinted at an interesting future in which something like a fully autonomous motorcycle might play a role in a Connected-Intelligent Traffic System. What if you used a ride-hailing service and a self-driving motorcycle picked you up? Veniam photo.
Barros told me that company-owned fleet vehicles — everything from garbage trucks to UPS delivery vehicles to taxis — represent about a quarter of vehicles worldwide, but accumulate 50 percent of the total mileage.

His company is betting that fleet managers will realize that V2X capability, which connects fleet vehicles to each other, to infrastructure, the internet, vehicle manufacturers, etc., can deliver an immediate payoff in improved efficiency and cost savings. If you’re UPS, you might not pay much for a safety feature that only interacted with a few of the vehicles on the road, but imagine the fuel savings if your trucks could communicate with traffic lights along their routes, allowing drivers to adjust their speeds to avoid stopping at red lights.

Like most (though not all) the people I’ve talked to in the HAV/C-ITS field, Barros would welcome a more pro-active government and a regulatory framework. Obviously, Veniam has a financial reason to want mandatory V2V rules, but that’s not all of it; Barros also wants clarity on technical standards.

“One of the big societal challenges we face,” Barros told me, “is that the pace of regulation is not as fast as the pace of technological development.”

Dr. Karl-Viktor Schaller rides his BMW
BMW Motorrad’s R&D chief Karl-Viktor Schaller obviously enjoys controlling motorcycles on his own, but he admitted that if motorcycles are to find a place in the future’s autonomous traffic systems, riders are going to have to accept some degree of robotic control. BMW photo.

BMW envisions motorcycles that operate in an ‘electronic safety cage’

Dr. Karl-Viktor Schaller, BMW Motorrad’s head of R&D, is certainly qualified to talk about the history of technology like ABS and stability control being developed for cars and then refined and adapted for motorcycles. Because it can swap engineers between automobile and motorcycle divisions, that cross pollination happens faster at BMW than most motorcycle-only companies.  

In order to ensure that BMW’s own HAVs are able to identify motorcycles, Motorrad gave the auto division a gift of motorcycle-shaped test targets for use when testing cars’ robotic guidance systems. Schaller told me that he was personally confident that BMW’s Advanced Driver Assistance Systems reliably spotted motorcycles. (He noted that BMW had also tested all the other German manufacturers’ cars and they’d all worked well when it came to avoiding motorcycles.)

Since existing BMWs do a good job of “seeing” motorcycles, Schaller argues that V2V may not be essential for motorcycles in the medium term. However, BMW’s long-term vision is of a future in which motorcycles operate in an ‘electronic safety cage’ that makes crashing obsolete. Even he admits that’s going to take V2V capability.

When I asked him how long it would be before BMW sold a motorcycle with a built-in V2V system he paused (perhaps exchanging a glance with Motorrad’s head of communications, who was sitting with him) before saying, “Not in the next two years, but I don’t think you’ll have to wait 10 years.”

Schaller dropped a particularly interesting hint at the end of our chat, when I asked him about the extent to which V2V capability could be retrofit onto existing motorcycles, to allow them to operate seamlessly in autonomous, connected traffic.

“That could be a partial solution,” he told me, “but you only get the full benefit if the system controls the engine and brakes, and maybe even steers.”

At that point I interrupted to ask, “Are you telling me you’ve built a self-riding motorcycle?”

Long pause: “Not as the prototype for a future product.”

Foresight is 2020

By 2020, many new cars and an increasing number of fleet vehicles will have V2X capability. Although there’s not complete agreement on exactly how those vehicles will communicate with each other, it seems likely that the safety part of that communication – the part that matters most to motorcyclists – will rely on DSRC. At least one motorcycle-specific chip is already available, and Bosch is developing a system based on it, for OEM use. The first aftermarket version is only a couple of years away, too.

Although the thought of some future BMW with Advanced Rider Assistance Systems that will take over steering in order to avoid an accident sends a chill through old-school bikers, BMW’s Schaller told me that in Europe, 51 percent of all motorcycle accidents are the fault of car drivers. The most common driver excuse is the usual, “I didn’t see him.”

Some time around 2020 — if not sooner – this will happen.

A distracted car driver will attempt to turn left directly into the path of an oncoming motorcyclist. But, the car itself will know that the motorcyclist is coming — even around a bend or if, for some other reason, there’s no line of sight — because the car will process a constant stream of data from the motorcycle.

The rider will get a warning, but even if the car driver attempts to cross the rider’s path, the car will simply not obey.

The crash won’t happen. I think we can all agree that’s a good thing