An AI and equity driven framework for mobile photo enforcement deployment

移动自动速度执法(ASE)计划已经在世界各地的许多司法管辖区运行了十多年，包括在加拿大。尽管已发表的文献中有很多关于评估这类方案的交通安全结果的文献，但在绘制方案投入以实现这些结果方面却很少。换句话说，关于项目操作(即选址、操作员调度)以及它们如何与由此产生的交通安全结果相联系的决策还没有得到广泛的探索。鉴于最近加拿大各地的司法管辖区对ASE技术重新产生了兴趣，将注意力重新吸引到此类项目是如何设计和运营以及如何做出决策变得至关重要。然而，最近对警察中的偏见的关注表明，调查ASE等基于社区的交通执法项目的结果也很重要，特别是场地选择和场地执法结果是否在城市地区的性别、种族和其他社会人口特征方面表现出偏见。为了应对这一挑战，该研究项目将开发一种系统但可转移的方法来部署ASE资源，调查并明确考虑公平和效率，使用大数据和机器学习技术。这项研究的预期结果是通过促进支持GIS的插补和可视化工具的开发来提高ASE在密集城市环境中的部署效率，这些工具将i)优先处理股权问题，ii)在有限资源下优化ASE部署的效率，以及iii)提高对速度限制的遵从性，这将最终提高安全性。利用从这项研究中获得的知识和开发的方法，ASE运营商将能够通过可视化和部署工具支持他们的部署，这将帮助他们在考虑城市社会公平影响的同时，就使用现有ASE改善交通安全做出更明智的决定。这项研究将产生大量新知识，以促进警务公平和ASE决策支持工具，使所有加拿大人受益。

Mobile automated speed enforcement (ASE) programs have been in operation in many jurisdictions around the world, including in Canada, for over a decade. Although there is much in the published literature in assessing the traffic safety results of such programs, there has been very little in terms of mapping program inputs to the achievement of these outcomes. In other words, the decisions that are made with respect to program operations (i.e., site selection, operator scheduling) and how they connect to resulting traffic safety outcomes has not been explored extensively. Given the recently renewed interest in ASE technology by jurisdictions throughout Canada, it has become critical to pulling attention back to how such programs are designed and operated, and how decisions are made. However, the recent attention to biases in policing reveals that it is also important to investigate the outcomes of such community-based traffic law enforcement programs as ASE, particularly whether site selection and site enforcement outcomes exhibit biases with respect to gender, race, and other socio-demographic characteristics within urban areas. To address this challenge, this research project will develop a systematic yet transferrable approach to deploy ASE resources, investigating and explicitly considering equity and efficiency, using Big Data and machine learning techniques. The expected outcomes of this research are to improve the efficiency of ASE deployment in a dense urban environment, by facilitating the development of GIS-enabled interpolation and visualization tools that would i) prioritize equity issues, ii) optimize the efficiency of ASE deployment under limited resources, and iii) increase compliance to speed limits, which will ultimately improve safety. Using knowledge gained and methods developed from this research, ASE operators will be able to support their deployment with a visualization and deployment tool, which will assist them in making more informed decisions on the use of existing ASE to improve traffic safety while also considering urban social equity implications. This research will generate a significant body of new knowledge to advance policing equity and ASE decision support tools that will benefit all Canadians.