Source author record

N. A. E

N. A. E appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

Catalog footprint

What is connected

1works
1topics
3close collaborators

Actions

Connect this record

Log in to claim

Research graph

See the researcher in context

Open full explorer

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

Published work

1 published item(s)

preprint2022arXiv

A General Measure of Collision Hazard in Traffic

A collision hazard measure that has the essential characteristics to provide a measurement of safety that will be useful to AV developers, traffic infrastructure developers and managers, regulators and the public is introduced here. The Streetscope Collision Hazard Measure (SHM) overcomes the limitations of existing measures, and provides an independent leading indication of safety. * Trailing indicators, such as collision statistics, incur pain and loss on society, and are not an ethically acceptable approach. * Near-misses have been shown to be effective predictors of incidents. * Time-to-Collision (TTC) provides ambiguous indication of collision hazards, and requires assumptions about vehicle behavior. * Responsibility-Sensitive Safety (RSS), because of its reliance on rules for individual circumstances, will not scale up to handle the complexities of traffic. * Instantaneous Safety Metric (ISM) relies on probabilistic predictions of behaviors to categorize events (possible, imminent, critical), and does not provide a quantitative measure of the severity of the hazard. * Inertial Measurement Unit (IMU) acceleration data is not correlated with hazard or risk. * A new measure, based on the concept of near-misses, that incorporates both proximity (separation distance) and motion (relative speed) is introduced. * Near-miss data has been shown to be predictive of the likelihood and severity of incidents. The new measure presented here gathers movement data about vehicles continuously and a quantitative score reflecting the hazard encountered or created (from which the riskiness or safeness of the behavior of vehicles can be estimated) is computed nearly continuously.