The challenges, which were selected by a panel of experts, were categorized into three major groups.
Nationwide, United States
Identifying High-Priority Needs for Law Enforcement Interactions with Autonomous Vehicles Within the Next Five Years
Summary Information
A recent report published by the RAND Corporation and the University of Denver sought to explore the intersection of law enforcement (LE) and automated vehicles (AVs). As AVs begin to be deployed, there is a growing need to understand the problems and opportunities that will emerge from LE interactions such as traffic stops and investigations.
On behalf of the National Institute of Justice (NIJ), the researchers convened a workshop in July 2019 with selected participants to discuss public safety scenarios related to traffic stops, collisions, emergencies, and tangential interactions (such as AVs involved as evidence in an investigation). The participants then ranked which needs they felt were most important. Following the workshop, the researchers categorized the 17 needs that were identified as being high priority into three groups: cybersecurity and communication protocols, stakeholder communication and collaboration, and standard procedures and training needs for LE interacting with AVs. The researchers used these needs to create potential paths for resolution, which may be employed by potential deployers.
The identified high-priority challenges related to cybersecurity and vehicular communication, as well as the authors’ proposed methods of meeting them, are below:
|
Challenge |
Proposed Method of Meeting Challenge |
|---|---|
| LE often is unaware of whether a vehicle is capable of or authorized to run in automated mode. | Identify the costs and benefits of various options to identify capabilities and authorization to run in automated mode (e.g., annotation on the electronic registration records, indicator on license plate). |
| There is a lack of understanding about cyber threats against privately owned or commercial vehicles. | Conduct a threat analysis and risk assessment (TARA) of AVs and design tools to detect cyberattacks and facilitate investigation for LE. |
| There is a lack of understanding about cyber threats to LE activities (e.g., traffic stops, identification, remote kill-switch, data exfiltration). | Analyze attack models from the perspective of LE agencies. |
| LE often is unaware of whether a vehicle is actually running in automated mode. | Examine the costs and benefits of various options of communicating with AVs running in automated mode (e.g., vehicle-to-vehicle, communication between the light bar and AV). |
| Using only lights or sound to initiate a traffic stop can leave an AV vulnerable to hacking from unauthorized entities. | Utilize a system that allows authenticated agencies to communicate their intentions. |
| LE always needs to have the ability to communicate with something other than a computer (e.g., a responsible human that is the owner or operator). | Develop an equivalent to "license and documentation" that allows LE to check the authorization to operate an AV (and potentially talk with a responsible human). |
| Verbal and nonverbal (and voice amplified) communication for traffic stops can be quite complex (e.g., asking vehicles to move into the next traffic lane, asking vehicles to move forward 100 yards, asking vehicles to move into a parking lot, indicating that a traffic stop is complete, or initiating special procedures for felony stops). | Identify the most-promising technological solutions that could be used in situations in which verbal communications are used (including assessments of potential adoption rates). |
Identified challenges and proposed solutions related to stakeholder communications were:
| Challenge | Proposed Method of Meeting Challenge |
|---|---|
| There is a lack of real-world knowledge among LE and other agencies about AV system capabilities and limitations. | Conduct workshops and ride-alongs for LE staff and other agency staff to raise knowledge levels. |
| AV developers have the potential to report relevant information to dispatchers and first responders but lack clarity on what kinds of information would be most useful (e.g., location of incident, number of passengers, airbag deployment, vital signs). | Conduct information-gathering exercises (e.g., workshops and surveys) to develop ideal approaches for conveying information to first responders. |
| At present, LE does not have a thorough understanding of the kinds of information that is being collected by AVs and how long it is maintained so that they can request the most appropriate information (for the purposes of crash reconstruction). | Survey LE and crash reconstruction experts to identify the type and quality of information that would be most useful, and pass this along to standards committees. |
| It is important for LE and other first responders to easily be able to stay current on procedures for accessing an AV (cutting power, towing, etc.). | Develop model web portals that could inform original equipment manufacturers about the kinds of information that LE would benefit from (e.g., LE points of contact, instructional videos on how to cut power). |
| There is a need for increased knowledge among AV system developers about how their product affects the community from a LE perspective (e.g., traffic stop procedures, unique aspects of a jurisdiction). | Conduct workshops and ride-alongs for AV system developers to raise knowledge levels. |
| It might be difficult for LE agencies to deal with AVs in areas of planned and unplanned closures (especially AVs that are operating without a capable operator as a backup). | Identify best practices for cities and other entities that have information about upcoming closures to share information on events, closures, etc. (This is useful only for planned closures.) |
Finally, the challenges and proposed solutions relating to standards and training needs were as follows:
| Challenge | Proposed Method of Meeting Challenge |
|---|---|
| LE often is unaware of whether a vehicle is actually running in automated mode. | Develop model training and guides for LE for identifying and interacting with AVs running in automated mode. |
| LE might have to respond to scenarios in which vehicles have been hacked and are not operable (e.g., from ransomware). | Develop guides and tools for potential LE responses to AV hacking. |
| LE might have to adapt tactics and procedures when AV technologies are employed incorrectly, improperly, or illegally. | Develop a guide containing likely scenarios in which AVs are used illegally and potential solutions. |
| There is not a consistent set of "traffic stop" maneuvers that AVs will be expected to perform. | Develop a general description of the kinds of behaviors that LE will expect AVs to be able to perform that is representative across the United States. |
