Conflict Detection and Resolution among Smart Services in Smart Cities
An increasing number of smart services are being developed and deployed in cities around the world. IoT platforms have emerged to integrate smart city services and city resources, and thus improve city performance in the domains of transportation, emergency, environment, public safety, etc. Despite the increasing intelligence of smart services and the sophistication of platforms, the safety issues in smart cities are not addressed adequately, especially the conflicts arising from the integration of smart services. In our work, we propose to build a system to detect and resolve conflicts among smart services. In particular, we propose a novel city-based stochastic spatial-temporal logic (CityTL) with novel operators, quantitative semantics and optimized monitoring algorithms to specify and learn smart city requirements. In addition, we propose to develop deep spatial-temporal predictive models to forecast city future states with consideration of requested actions and uncertainties. By verifying predicted states with CityTL specified requirements, we will be able to detect conflicts. To resolve the conflicts, we will generate resolution options by considering the complex dependencies between actions. When not all the requirements are met, we will provide different ways to calculate the trade-offs of satisfaction degrees of different resolution options. We will evaluate individual models and the whole system comprehensively using large amounts of real city data, and then deploy and test the system in the operating center of a real smart city.
Alfred Weaver (Chair), John Stankovic (Advisor), Lu Feng, Haiying Shen, and John Lach