As loyal followers of our blog may have noticed, I have a keen interest in developments relating to automated vehicle (“AV”) or “self-driving” technology. For a couple of my recent posts on the topic, including some of the legal implications in this area, please see here and here. In my last post, I noted that the National Highway Traffic Safety Administration (“NHTSA”) scheduled two public meetings with respect to “operational guidance” for automated vehicles. Specifically, NHTSA sought “input on planned operational guidelines for the safe deployment of automated vehicles” including “information on the roadway scenarios and operational environments that highly automated vehicles will need to address, and the associated design and evaluation processes and methods needed to ensure that AV systems are able to detect and appropriately react to these scenarios.”
The first meeting was held on April 8 at U.S. Department of Transportation headquarters in Washington, DC. A thorough account of the meeting is available here. One of the interesting points from the meeting relates to the speed at which the agency’s guidance is being developed: while the agency is moving quickly and has a goal of issuing guidance by July, some attending the meeting suggested that a slower and more deliberate approach is needed–and that the traditional (and much slower) process of issuing regulations and standards should be utilized. Such regulations are also enforceable whereas “guidance” is typically more general and subject to interpretation.
The problem, as NHTSA acknowledges, is that issuing new regulations takes an average of eight years and the agency cannot wait that long because self-driving technologies are already on the road. NHTSA Administrator Mark Rosekind noted at the meeting that there are vehicles on the road today that automatically steer, change lanes and adjust speed in response to traffic. According to Dr. Rosekind, without federal instructions, “people are just going to keep putting stuff out on the road with no guidance on how do we do this the right way.” Dr. Rosekind further noted that he sees AV as “game-changing technology” that could save the lives of many of the more than 30,000 people killed each year on the nation’s roads.
Still others at the meeting expressed concern that self-driving technology cannot yet safely operate under a number of conditions, including:
- Poorly marked pavement, including parking lots and driveways, without clear lane markings.
- Bad weather that interferes with vehicle sensors.
- An inability to take directions from a policeman.
- An inability to follow inconsistent traffic-control devices such as horizontal versus lateral traffic lights.
In addition, meeting attendees discussed issues of cybersecurity, including that self-driving cars, like other systems that rely on wireless technology, may be vulnerable to hacking.
NHTSA is hosting a second public meeting on AV operational guidance on April 27 at Stanford University. In advance of the meeting, NHTSA has announced several topics for discussion:
- Evaluation and testing of scenarios the AV system should detect and correctly operate in: Within the AV system’s operating envelope, consider how to identify the scenarios that could be encountered by the AV system (e.g., behavioral competencies/normal driving, pre-crash scenarios, etc.) and what design and evaluation (testing) processes and methods are needed to ensure that the vehicle can detect and appropriately react to these scenarios. Consider whether third-party testing is appropriate for validating test results.
- Detection and communication of operational boundaries: If there are limitations on where AV technology will operate – what methods should the AV technology use to sense when it is reaching the operational domain limit and how should that be communicated to the driver?
- Environmental operation and sensing: Consider what environmental conditions AV systems will likely operate in. For environmental conditions in which AV systems are not designed to operate, discuss methods used to detect these conditions.
- Driver transitioning to/from AV operating mode: For AV systems that rely on transferring vehicle operation back to the driver, discuss approaches to (a) ensuring safe transitioning back to a fully capable non-impaired driver (e.g., geo-fencing, adverse weather) and (b) how non-optimal driver behavior (e.g., decision errors, erratic behavior, driver impairment) will be addressed by the AV system.
- AV for persons with disabilities: Consider the unique needs of people with different types of disabilities in the design, development, and policy setting for self-driving cars and related automation.
- Data: Consider data recording capabilities of system(s) necessary to monitor the correct operation of the AV system, and what are appropriate triggers (crash, near crash, etc.) to determine system operational status or possible malfunction of the system. Also consider how recorded data could be accessed and by whom. During the testing phase, consider what data should be made public for further analysis and understanding.
- Crash avoidance capability: Consider the capabilities of AV systems with respect to detecting roadway hazards (other vehicles, pedestrians, animals, etc.) such that common crash scenarios involving these hazards (control loss, crossing paths head-on, etc.) can be detected and either avoided or mitigated.
- Electronics systems safety: Consider methods and potential documentation that could be produced with respect to functional safety and cybersecurity.
- Non-passenger AVs: Consider differences between AVs designed for delivery of goods and products that are not intended to have a human operator or potentially even human passengers.
- Aspects of AV technology that may not be suitable or ready for guidelines: For these areas, information would be useful on alternative approaches to assure safety.
- Identification of industry voluntary standards, best practices, etc., related to automated vehicle operation.
- Information AVs may need to communicate to pedestrians and other vehicles (manual or automated) just as a driver would. Consider situations such as pedestrians crossing a travel lane in a parking lot and how this communication should be accomplished.
- Conditions in which AV s may need to be able to identify and communicate to a central location or authority that a problem has occurred. Consider situations where passengers may be delivered to their destination but a medical problem or potential incapacitation en route may potentially suggest considerations for vehicle capabilities that could handle such cases.
- Operation of an AV with open safety recall: Consider if automated vehicles should be allowed to operate in automated mode in cases when there is an open safety recall on that vehicle or if automated functions should be restrained until recall repairs are completed (perhaps reversion to manual driving when possible). Consider if AV s with open recalls should be allowed to operate on public roads at all, and if so, under what conditions.
- Other topics needed for operational guidance: Other topics that would be beneficial to address in an operational guidance document to facilitate innovation and safe deployment of these systems on public roadways.