"Investigating the Factors Affecting Injury Severity of Single-Vehicle Rollover Crashes in the United States"

This study developed a generalized ordered logit model to investigate the effects of various factors on injury severity of occupants in single-vehicle rollover crashes. The study used five-year crash data in the United States from 2012 to 2016. The effects of explanatory variables considered in the model development included roadway attributes, crash and environmental information, driver characteristics, and vehicle features. Results showed that the likelihood of serious and fatal injuries increases in rollover crashes with occupant's partial or complete ejection, speeding, higher posted speed limits, roadside and median rollovers, undulating terrain, blacktop road surface, rural roads, evening time, weekdays, older driver age, no occupant protection, careless or inattentive driving, driver-passenger engagement, aggressive driving, and passenger car. The deployment of air bag was associated with lower serious and fatal injuries. Moreover, the regional variations were more likely to affect injury severity in rollover crashes. The study findings can help safety stakeholders in prioritizing countermeasures through a better understanding of factors in rollover crash injury severity.-Ihsan Khan

On-board sensors allow for the combining of features extracted from inertial sensor signals, across multiple train traversals, to significantly enhance their signal-to-noise ratio, which minimizes false positives and false negatives. However, position registration errors, relatively slow update rates of the GPS receivers on low-cost devices, and non-uniform sample interval variabilities of their inertial sensors result in feature alignment errors that can actually degrade the signal-to-noise ratio. This project introduces four heuristic methods to align the inertial signals from multiple traversals. The authors selected the best-performing method by ranking the position variability of a known ground truth anomaly and the corresponding statistical distribution that maximally agrees with a Gaussian.-Leonard Chia

"A Deterministic Mathematical Model to Support Future Investment Decisions for Developing Inland Container Terminals"

Agriculture is a leading sector in the Midwest economy. Within it, grain production is especially important to natural resource-based economies in the upper Midwest. Exporters are put at a competitive disadvantage when they are unable to obtain containers at a reasonable cost for their exports. This situation has increased the desire with inland shippers for more and closer availability of empty containers to facilitate grain transportation.

To mitigate the shortage of containers and excessive empty vehicle miles, it is proposed to strategically establish inland depots in regions with sufficiently high trade volumes. Inland depots aim to minimize total system costs for empty container repositioning, while providing customers with their desired level of service. Mathematical models are formulated to evaluate the proposed system, while also testing the realistic scenarios. The implementation of the model is performed as case study for soybean container shipments in the study region of Minnesota. The proposed system has been demonstrated to significantly reduce empty vehicle miles traveled and total system costs in the study region, yielding benefits to the regional exporters and individual stakeholders. Findings of interest show that even if the soybean bean trade volumes in the region remain static or even decrease, inland depots will still result in noteworthy system cost and empty vehicle mile savings. Finally, the model can be applied similarly to other commodities and/or be used to analyze the potential for new intermodal points in other geographic regions.-Satpal Wadhwa

"Application of Decision Tree Method on Bridge Deterioration Forecasting"

Bridges are essential transportation infrastructures. Nowadays, bridge maintenance is conducted after a bridge evaluation. However, considering the cost of time and labor of visual evaluation, an analyzing predictive model can provide guidance to identify bridge deterioration rate. Traditionally, Markov Chain models and Statistical models have been adapted to forecast bridge deterioration, but both of them have disadvantages that can't be neglected. 

Decision tree method, a novel computational modeling technique that uses flowchart-like tree structure, has been widely used for classification and prediction in many scientific and medical fields. The research team is using Decision Tree method to forecast bridge deteriorate rates.  The research conducted DT method on 1,869 records of Deck data and 2,378 records of Superstructure and Substructure data, exact forecast rates for Deck, Superstructure and Substructure are 54%, 60% and 53%, respectively. Forecasting results within 1 and 2 rating differences are all 89% and above. The Research Team recommends Decision Tree as an alternative for making reliable predictions which are close to actual observations.-Mingwei Guo