| Abstrakt: |
To fully leverage “smart” transportation infrastructure data-stream investments, the creation of applications that provide real-time meaningful and actionable corridor-performance metrics is needed. However, the presence of gaps in data streams can lead to significant application implementation challenges. To demonstrate and help address these challenges, a digital twin smart-corridor application case study is presented with two primary research objectives: (1) explore the characteristics of volume data gaps on the case study corridor, and (2) investigate the feasibility of prioritizing data streams for data imputation to drive the real-time application. For the first objective, a K-means clustering analysis is used to identify similarities and differences among data gap patterns. The clustering analysis successfully identifies eight different data loss patterns. Patterns vary in both continuity and density of data gap occurrences, as well as time-dependent losses in several clusters. For the second objective, a temporal-neighboring interpolation approach for volume data imputation is explored. When investigating the use of temporal-neighboring interpolation imputations on the digital twin application, performance is, in part, dependent on the combination of intersection approaches experiencing data loss, demand relative to capacity at individual locations, and the location of the loss along the corridor. The results indicate that these insights could be used to prioritize intersection approaches suitable for data imputation and to identify locations that require a more sensitive imputation methodology or improved maintenance and monitoring. |
