The makers of autonomous vehicles should incorporate effective motorcycle recognition into their fundamental design, not as an afterthought.
Like it or not, autonomous cars are hitting the road. One of their main selling points is that they will be safer than human drivers when it comes to avoiding accidents. It's true that human drivers aren't that great at avoiding motorcyclists. Can autonomous cars do better? Companies are only just starting to test them on open public roads, but so far the results are less than promising.
As we previously reported, there are at least two recorded cases so far of an autonomous car crashing into a motorcycle. On July 27, 2016, a motorcyclist in Norway was seriously injured when she was rear-ended by a Tesla Model S with its Autopilot self-driving system engaged. Additionally, on December 7, 2017, California motorcyclist Oscar Nilsson was hurt in a collision with a self-driving Chevy Bolt. The car had aborted a signaled lane change and was returning to its original lane while Nilsson was lane splitting past it, a maneuver that is legal in California. The car failed to detect Nilsson and lightly bumped him, causing his bike to fall over.
Another disturbing incident occurred when an autonomous Uber car struck and killed Elaine Herzberg, who was walking her bicycle across the road at night. One would think that the car's radar and infrared scanners would detect such an object far better than the human eye could. Indeed, the car's onboard computer detected the obstruction in the road six seconds before impact. Unfortunately, it failed to identify what type of object she was until just 1.3 seconds before impact, at which point it was too late to avoid hitting her. Other factors, such as an inattentive human driver who was supposed to be monitoring the car's surroundings, also contributed to the fatal crash, but a computer that fails to identify a target for nearly five seconds should be quite disturbing to motorcyclists who may suffer from this fate as well.
This trend isn't limited to autonomous cars, either. The example of the Tesla is a car that is not autonomous but is common on our roadways today. Adaptive Cruise Control is a radar guidance system that enables a car to change its speed relative to the vehicle in front of it based on traffic conditions. A study by RDW has determined that existing adaptive cruise control systems often do not do an adequate job of locking onto a motorcycle rather than a car. These systems seem to have difficulty detecting a motorcycle not riding in the center of its lane. This is a problem since riders tend to occupy either the left or right tire grooves of a lane, rather than the center, to avoid oil spills, road kill, and other hazards that car tires tend to push to the center of the lane.
Another study by Dynamic Research shows similar results when it comes to forward collision warnings detecting motorcycles. Throughout the tests, forward collision warning systems failed to detect the motorcycle adequately in 40 percent of trials. This 60 percent failure rate isn't too far off the 77 percent rate of human driver/motorcycle accidents occurring within a 60-degree arc of the front of the car cited in the famous Hurt Report. Similarly, 37 percent of simulated crashes in this test occurred because the car's onboard detection systems didn't see the motorcycle, which is also the number one cause of motorcycle accidents with human drivers.
As a result of this, the American Motorcyclist Association wants to get involved on the ground floor of autonomous driving system design. "The AMA wants the technology to be developed with us and with motorcyclists in mind, rather than deploying the technology, then trying to work motorcyclists into the picture," Michael Sayre, On-Highway Government Affairs Manager for AMA, told RideApart. "The AMA has reached out to manufacturers, systems developers, elected officials, and regulatory agencies with that request."
If autonomous driving systems are designed properly, keeping motorcyclists in mind, they could potentially deliver on their promise of making the roads safer for everyone. They could be on the lookout for motorcycles and other road hazards in all directions at all times, something even the most attentive human driver can never do. The trick is to design such systems to work in this way as a core functionality. Autonomous vehicles must be able to search, evaluate, and execute the same way the Motorcycle Safety Foundation teaches all riders to. This is especially true when it comes to detecting motorcycles, which don't have a steel cage around them to protect them from autonomous cars' mistakes.
Automated Vehicles and the Safety of Motorcyclists
The safety of motorcyclists is of singular importance to the mission of the American Motorcyclist Association. As technology allows vehicles to communicate with each other and with roadway infrastructure, the promise of improved safety is alluring. To decrease the number of motorcycle crashes and resulting injuries and fatalities, it is paramount that automated vehicle technology, including highly automated vehicles, be capable of recognizing and properly reacting to motorcycles in all traffic situations and settings, including in parking lots, amid urban congestion, at intersections and on highways, expressways and rural roads.
Automated vehicles can bring a greater measure of safety to motorcyclists and cannot be overlooked. Distracted driving is one of the major causes of motorcycle crashes that are frequently the fault of the motorist. A properly designed, complete automated system of control, highly refined in its ability to recognize motorcycles, can truly save lives. The AMA welcomes the potential of this type of vehicle, once thoroughly vetted. Unfortunately, the industry is still many miles away from the development of a system that is able to interact safely with motorcycles in many common real world situations.
The rush to market of driver-assist systems, semi-autonomous vehicles and highly automated vehicles—referred to collectively as AVs—poses a significant threat to motorcyclists when the developers of this technology and the vehicle manufacturers are not held to the highest safety standards throughout the entire development and implementation process. If AV systems are not conceived and developed with motorcycles and motorcyclists in mind, the eventual result could be that motorcycles would be excluded from certain roadways, or, worse, banned from roads altogether.
Motorcyclists have been an integral component of daily transportation and recreational activity on our public roads and highways for over a century. The AMA has an indelible history of protecting access for motorcyclists to our transportation infrastructure.
Motorcycles meeting federal design, safety, sound and emission requirements operate legally on our public roadways. The issue of safe access is also directly related to the high cost of licensing, ownership and fuel taxes, including the substantial expense to the motorcycle consumer of meeting state and federal laws pertaining to safety and emissions in the manufacturing process. Additionally, motorcyclists are and have been direct contributors to both the building and policing of our highway systems and hence contribute directly to the safety of all motorists.
Motorcycling provides many benefits to riders and to the public at large. On average, motorcycles and scooters consume fewer resources and emit less carbon dioxide per mile, take up less space in parking areas and impose very little wear and tear on our nation’s roads and infrastructure, especially compared to automobiles, trucks and SUVs. Furthermore, motorcycling tourism and events provide substantial economic benefit in the form of revenue and tax receipts to towns, municipalities and counties that cater to motorcycling enthusiasts.
The AMA has been monitoring the development of safety-enhancing technology for decades—and AV systems since the early 2000s. The AMA is committed to ensuring that AV technology benefits all road users, and, specifically, motorcyclists.
Working with the U.S. Department of Transportation and its agencies, including the National Highway Traffic Safety Administration and the Federal Highway Administration, as well as elected officials, automakers, technology companies and software developers—the AMA has strongly advocated that motorcycle safety must be an integral component of every AV technology program. The goals of this effort include ensuring that:
- Motorcycles are included when AV, vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communication systems are designed, tested and implemented;
- The implemented communications systems, whether using Wi-Fi or another technology, are secure from hacking; and
- These technologies protect the privacy of all road users.
If manufacturers can self-certify that their AVs comply with existing applicable safety standards, no federal law or regulation prevents AVs from being built or from being tested on roadways shared with motorcycles. Neither the 2016 Federal Automated Vehicles Policy nor the 2017 Automated Driving Systems guidance change addresses this situation. They instead sacrifice safety for technological flexibility. Administrative agencies actions often reflect pressure from manufacturers to bring products to market without adequate testing.
The federal policy lists examples of behavioral competencies that AV manufacturers should assess, test and validate. However, as expressed above, what are the sanctions for failure to address motorcycles, which are a legal class of motor vehicle with distinct operating characteristics? Currently, the answer is none.
As a result, the AMA continues to demand that motorcyclists be included in the ongoing discussion, planning and implementation of the policies and regulations governing the rollout of AVs.
Failure to specifically address motorcycles in statutory and regulatory language amounts to the abandonment of motorcycle safety by legislators and regulators. Essentially the issue of distracted and inattentive driving will mushroom into a monumental hazard for motorcyclists when flawed AV technology enters the transportation mainstream.
Moreover, the AMA opposes any provisions that limit or eliminate motorcycle access to public roadways. The AMA maintains that acceptance of a policy that attempts to perpetuate a view that motorcyclists present a "social burden" on America's highways based on flawed AV technology is both discriminatory and contrary to the long-term interests of motorcycling.
From the beginning, the AMA has offered its expertise repeatedly during AV development. The AMA now calls upon the president and Congress to immediately direct the appropriate federal agencies to implement automated-vehicle policy and guidelines to improve and ensure the safety of motorcyclists.
In summary, the American Motorcyclist Association position regarding AVs, V2V and V2I technology is such that these technologies must be thoroughly reviewed and tested to maximize crash avoidance involving motorcyclists. However, the rush to market for maximization of sales of AVs without complete and competent analysis of the relationship of motorcyclist safety to the AV environment, is an invitation to injury and death.
Simultaneously, the president and Congress, along with pertinent federal agencies, must hold vehicle manufacturers and those developing this technology accountable by enacting regulations and/or guidelines that include consideration of motorcycles and motorcyclists in the development and deployment processes.
These regulations must come with sanctions to be genuinely effective at achieving the level of safety motorcyclists are entitled to whether it is in fairness for their monetary contributions to road use or more importantly enjoy the right to be safe operating on the public streets, roads and highways.
Finally, we encourage all stakeholders involved in the development, testing and rollout of AV technology to include the AMA in their efforts.