AV deployments will be limited to operational domains (OD) where vehicles can readily demonstrate safe operation. Limitations in ODs may stem from factors that challenge an automated vehicle’s ability to accurately perceive the surrounding environment and effectively make decisions, such as adverse weather and degraded lane markings. City and state agencies may wish to extend the ODs to improve benefits, such as to improve economic opportunities and accessibility, connect strategic locations, and simply be an attractive place for AV testing and deployment. While agencies may not be able to control some of these OD limitations, they can take steps to modify infrastructure to improve the OD.
Infrastructure modifications may include infrastructure-to-vehicle (I2V) communication systems, signage, and civil infrastructure such as curbs and barriers to provide different levels of segregation between the CAVs and other road users. Uniform and well-maintained traffic control devices, such as lane markings and traffic signs, may improve the extent of AV OD. AV functionality depends on perception algorithms to accurately detect and respond to infrastructure based on sensor information. Just as humans learn to drive through experience, many perception algorithms use machine learning that is trained to detect and classify objects and events based on past experience. Atypical conditions are more challenging for perception systems. Segregation can create a less complex environment by eliminating mixed road users that can be unpredictable, or can even take advantage of an AV’s conservative behavior. Infrastructure owners and operators want to understand how the OD of near term deployments may benefit from infrastructure modifications.
This research will review and identify potential infrastructure modifications that could improve the OD of AVs. The analysis will:
· Investigate aspects of technology and operation that influence OD, such as vehicle connectivity, dedicated lanes, AV sensors, perception algorithms, operating speeds, and pickup/drop-off locations;
· Review lessons learned from AV testing and deployment activities;
· Identify and characterize aspects of physical and digital infrastructure elements that may limit OD, such as V2I, curbs, barriers, reflectivity, geometry, and quality of data; and
· Develop implementation guidance for infrastructure modifications, including potential improvements to OD, costs, and impacts to other road users.
This research will provide state and local transportation agencies with guidance on how to modify infrastructure to improve the OD of AVs. The assessment will provide insights based on AV technology and operations. It will provide a catalogue of potential infrastructure modifications, and describe how these modifications will impact ODs. Key considerations for prioritizing potential modifications will be provided to infrastructure owners and operators to enable decision making and investments.
In addition to highly automated vehicles, the study should include automated driving system (ADS) technologies that are available now. The scope should include the digital infrastructure, data management, and work zones. The scope of this task should be coordinated with that of NCHRP 20-102(24) to minimize overlap.