On May 7, 2016, news spread around the world about Joshua Brown, who was killed instantly when his vehicle collided into a tractor-trailer crossing an uncontrolled intersection in central Florida. Mr. Brown's accident was particularly momentous as it marked the first fatality in a vehicle using semi-autonomous functions engaged.
Mr. Brown's Tesla Model S was equipped with Autopilot, which detects objects in its path and surroundings using a combination of cameras and ultrasonic sensors. This technology allows the vehicle to independently maintain and adjust speed, change lanes and brake, all without express driver input.
This first of its kind collision sparked concern in the technology industry. Debates ensued as to whether the safety feature created to eliminate (or at the very least, reduce) motor vehicle accidents was the cause of the accident, whether human error was to blame, or some combination of the two.
What Went Wrong?
The U.S Department of Transportation's National Highway Transport Safety Administration ("NHTSA") opened an inquiry into Mr. Brown's death to determine the cause of the accident.
The Autopilot technology implemented in the Tesla Model S, and other Tesla models, is intended to be used as an aid to the driver in navigating the road. Tesla expressly advises drivers to focus their full attention on the road and keep both hands on the wheel during operation.
On January 19, 2017, the NHTSA concluded its inquiry, stating that:
An attentive driver has superior situational awareness in most of these types of events, particularly when coupled with the ability of an experienced driver to anticipate the actions of other drivers.1
The NHTSA assembled a Crash Investigations team to investigate the crash site and to study and evaluate the vehicle. The results of the accident reconstruction suggested that the tractor-trailer ahead of Mr. Brown should have been visible to him for at least seven seconds prior to impact. Unfortunately, Mr. Brown failed to see the oncoming vehicle and did not activate the breaks.
Also, according to the report, the Autopilot system in the subject vehicle performed properly and without defect. To which, the NHTSA declared the Automatic Emergency Breaking (AEB) system to be in-line with industry state of the art for AEB performance, but it also noted that braking for crossing path collisions are not in the performance capabilities of the system.2
The report highlighted that the vehicle still required human input in certain circumstances. In other words, the Autopilot system is not a fully autonomous system, but an advanced Driver Assistance System (ADAS). ADAS requires a driver to be fully engaged with the road traffic in order to take immediate action, should it be required.
The report ultimately concluded that the cause of the accident was two-fold: The Automatic Emergency Braking (AEB) system did not provide any warning or automated braking for the collision event, and the driver did not take any actions to avoid the collision. It states:
A safety-related defect trend has not been identified at this time and further examination of this issue does not appear to be warranted.3
Autonomous v Semi-Autonomous
The elementary question remains as to what is the difference between an autonomous vehicle and a semi-autonomous vehicle?
The Ontario Ministry of Transportation defines autonomous vehicles as "driverless or self-driving vehicles capable of detecting and navigating the surrounding environment, and have the potential to help improve road safety and fuel efficiency, as well as reduce traffic congestion and greenhouse gas emissions".4
This definition does not specifically differentiate between fully and semi-autonomous vehicles. Whereas in Ontario, fully autonomous vehicles have not been approved for everyday road use, the number of semi-autonomous vehicles available for purchase is steadily growing.
The specific capabilities and features of each vehicle so-equipped with semi-autonomous functions vary by manufacturer and their models; however, all such examples require the driver to be attentive and aware of the surrounding road conditions and traffic.
Clearly, the ultimate responsibility, at law, in North America over semi-autonomous vehicles thus far has been found to lay with the driver. In Ontario specifically, fully autonomous vehicles have not been approved for everyday road use yet; however, it remains unclear whether the responsibility at law will shift away from the driver once fully autonomous vehicles are approved for operation.
This undoubtedly will be an issue that drivers, insurers and lawyers will have to address in the very near future, as the distinction between a vehicle's drive mode, autonomous versus semi-autonomous, will have a significant impact of whom is at fault for a motor vehicle accident, the driver or the vehicle's designer/manufacturer.
Automation ona Spectrum
Unfortunately, the distinction between an autonomous vehicle and a semi-autonomous vehicle is not black and white.
The Society of Automotive Engineers (SAE) International has attempted to reduce the confusion of autonomous vehicles by creating a harmonized classification system that has categorized vehicles based on their level of autonomy. According to the SAE, this system was created with the goal of educating and clarifying the role that drivers play in each level of automation.5 The classification may also assist in sorting responsibility for accidents based on the specific vehicle(s) involved.
According to the SAE, the classification has six levels of driving automation that span from "no automation" ("0") to "full automation" ("5"). Vehicles classified as Level 0, 1 or 2 require the human driver to monitor the driving environment, whereas that responsibility in Levels 3 through 5 fall on the automated driving system. Vehicles that fall under Level 5 automation can achieve all aspects of operational and tactical driving tasks under all roadway and environmental conditions that can be managed by a human driver. The specifics of each category can be found in the SAE's published International Standard J3016.6
The elements in each level, however, indicate minimum, rather than maximum system capabilities, leaving the possibility that any particular vehicle could fall into different levels depending on which automated features are in use at a specific time. This is clearly contemplated by the SAE's description of system capability as "some driving modes".7 Only vehicles falling in Levels 0 and 5 are unambiguous in this regard.
The Ontario Ministry of Transportation, in recognizing the utility in this technology, has launched the first automated vehicle pilot program in Canada. The program, led by the Erwin Hymer Group and Blackberry's QNX at the University of Waterloo, intends to test automated vehicles that operate at Levels 3, 4 and 5 based on the SAE's classification.8 No results have been published at this time.
Despite Mr. Brown's unfortunate accident, automated vehicles have statistically made driving safer. According to the NHTSA, more than 30,000 people die in motor vehicle related collisions each year in the United States and 94% of those collisions have been attributed to human error.9 With the gradual introduction and advancement of automated vehicles, the NHTSA is hoping that the contribution of human error in collisions will be greatly reduced. This is all said to start at the vehicle manufacturer level.
Since Mr. Brown's accident, Tesla has made further changes to its technology to promote driver attention during Autopilot's operation. For example, if a driver chooses to ignore the safety message on the dashboard, including the reminder for the driver to keep his or her hands on the wheel, Autopilot will disengage after several warnings and will not be reactivated until the car is shifted into park. Prior to the update, the previous system would simply slow the vehicle to a stop if the driver ignored these warnings.
Despite the continuing advancements in driving technology, operators of motor vehicles must remain cognizant about the capabilities and limitations of their automated driving system. Driver education, including awareness of the vehicle's capabilities and the requirement for continued human input, will be paramount to determining respective fault in the case of motor vehicle collisions.
1 NHTSA report, Automatic Vehicle Control Systems (19 January 2017), online.
4 Ministry of Transportation, Automated Vehicles Coming to Ontario Roads. November 28, 2016.
8 Supra, note 4.
9 NHTSA, Automated Vehicles.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.