Dr. Haobing Liu is a Research Engineer II and Instructor in School of Civil and Environmental Engineering at the Georgia Institute of Technology. His career and research interests center on energy, emissions, and air quality modeling of transportation sectors, with special interests in exploring innovative technologies and policies that shape sustainable transportation systems. Haobing's Ph.D. dissertation research, “Modeling the Impact of Road Grade on Driving Behavior, Vehicle Energy Consumption, and Emissions” examines vehicle operations in response to road grade changes, and the incorporation of such behaviors into energy, emissions and air quality modeling system. Haobing developed a streamlined machine learning method using Python to generate high-resolution road grade based on United States Geological Survey Digital Elevation Model (DEM), an open source LiDAR data. The straightforward method and publicly available DEM enable researchers to easily implement in any other regions across the United States. With the grade information available, Haobing also built a Hierarchical Bayesian Model to examine the impact of road grade on vehicle operations, including the operation heterogeneity across drivers, vehicle types, and road types. The model indicated a strong correlation between driving behavior and road grade, while such relationship varies by drivers and vehicle types. The results also highlighted the importance of integrating road grade and grade-behavior correlation into vehicle energy consumption and emissions modeling, especially for heavy-duty vehicles. The findings can lead to enhanced agency guidance on improved emissions modeling and have the potential to improve vehicle activity model with grade impact integrated. In addition to dissertation research, Haobing has also leveraged various emerging data science and computer science techniques to address critical inquiries in sustainable transportation system: One of Haobing's research assessed life-cycle energy consumption and emissions for intercity passenger travel for more than 200 city-city pairs by aviation, intercity bus, and automobile. The modeled processes included vehicle manufacturing, infrastructure construction and maintenance, upstream fuel production, and vehicle operation and maintenance. The research to date indicates that lifecycle energy consumption and emissions per passenger-mile of travel are lowest for fully-loaded intercity buses. For medium to long-distance travel (500+ miles), aviation is energy more efficient than automobiles, due to the high fuel efficiency of the air cruise mode, while automobiles are more efficient than aviation for shorter distances, sue to takeoff and landing energy use. This research provides a basis for future policies designed to encourage mode shifts by a range of service. In 2016-2017, Haobing developed a real-time near-road air quality modeling system by linking hourly traffic data and meteorology data with vehicle emissions model and dispersion model in parallel supercomputing cluster. The new tool will facilitate server-side automation of transportation and air quality analysis for transportation planning and management. In 2014-2015, Haobing participated in the project of emissions modeling and eco-driving for transit. An eco-driving strategy has been designed to avoid high-speed and hard acceleration driving, and results indicated eco-driving can reduce fuel consumption of diesel and CNG transit fleet by about 4%. These results are of great significance, considering eco-driving training does not require significant capital investment. In the Summer of 2015, Haobing worked as an ORISE Fellow in the National Vehicle Fuel and Emission Lab at the U.S. Environmental Protection Agency (US EPA). I conducted a successful project on vehicle activity and emission modeling at highway ramps that were an improvement for MOVES model, a regulatory vehicle emission model published by US EPA for conformity analysis in the United States. EPA plans to integrate my research result into MOVES 2019 version since it is an improvement in modeling accuracy on highway ramps. Another research Haobing led was vehicle classification method for MOVES modeling. A classification method was proposed to generate vehicle type distribution for improved MOVES input at project-level analysis. The paper has been presented in 2015 TRB meeting, The Transportation and Air Quality Committee (ADC20) of the Transportation Research Board selected our paper as its "Spotlight Presentation" for that year, since the paper was ranked 1st among 220 papers reviewed. In this field, Haobing has authored or co-authored more than 15 peered-review papers in well-known transportation and energy journals, including Transportation Research Part-C, Transportation Research Part-D, Applied Energy, Journal of Transportation Engineering (ASCE), and Journal of Transportation Research Record.
Sustainable Transportation, Energy, Emissions and Air Quality Modeling in Transportation Sectors, Data Mining and Statistical Modeling for Transportation Applications, Geoprocessing and Big Data Analysis in Transportation Based on Parallel Computing Clusters