Projects

 

• Cyber-Physical Systems for Connected and Autonomous Vehicles

 

Project title: Development of Cyber-Physical Systems for Connected and Autonomous Vehicle Research

 

Funding agency: South eastern Transportation Center

 

Publications: Peer-reviewed Journals and conference papers

​

Description: The Center for Connected Multimodal Mobility (C2M2) is developing a cyber-physical system (CPS) for connected and autonomous vehicle research by utilizing and enhancing the functionalities of the Clemson University Connected and Autonomous Vehicle Testbed (CU-CAVT), located in Clemson, South Carolina, to support the future research at partner institutions. Our CPS at CU-CVT includes heterogeneous wireless communication technologies and data infrastructure for real-time connected vehicle (CV) data exchange, streaming, fusion and archiving. One of the key features of CU-CAVT testbed is the implementation of layered architecture, which enables the reduction of data delivery delay for the services provided by our CPS and supports reduced bandwidth requirements and data loss. These are necessary to accommodate a large number of CAVs, and to simultaneously support multiple and diverse CAV applications.

​

​

• Data Analytics for Smart Transportation Systems

​

Project title: Big Data Analytics for Connected Vehicle Data Infrastructure Resiliency

 

Funding agency: South eastern Transportation Center

 

Publications: Peer-reviewed Journals and conference papers

 

Description: Emerging connected vehicle technologies (CVT) provide vehicles with a 360 degree of awareness, which will warn the motorist of any crash imminent conditions and thus reduce such accidents. CVT applications also support efficient mobility and environmentally sustainable travel. A complex and massive amount of data will be collected from onboard CV sensors, transportation infrastructure and mobility data sources, social media, and news and weather sources in connected transportation systems. However, the primary challenge in enabling CV applications involves aggregating and processing collected data for redistribution, to satisfy specific CV application requirements based on time and spatial contexts. The proposed research will develop alternate Big Data infrastructure architectures for the safe and secure operations of diverse CV applications related to safety, mobility and environment, and evaluate alternate big data infrastructure for diverse CV applications. For this purpose, centralized and distributed data infrastructure will be developed and evaluated considering diverse CV application requirements via simulation and real-world field tests. We will also develop information service priority rules based on CV safety applications priority needs. The research will provide a much needed guidance to establish a sustainable and resilient data infrastructure for future connected vehicle technology deployments in the real-world.

​

• Intelligent Connected and Autonomous Vehicle Applications using Machine Learning

​

Project title: US Ignite: Track 1: Enabling Connected Vehicle Applications through Advanced Network Technology

 

Funding agency: National Science Foundation (NSF)

​

Publications: Peer-reviewed Journals and conference papers

 

Description: By the end of the decade, the US Department of Transportation (DOT) will likely require all new vehicles to be Connected Vehicles (CV), capable of communicating with other vehicles and roadside infrastructure through wireless communications in order to reduce the number of crashes and save lives. The crash avoidance applications supported by vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) connectivity exchange safety critical information such as speed, location and direction of movement to assess the crash risk based on the proximity of vehicles. While standards such as Dedicated Short Range Communications (DSRC) will play a key role, other technologies such as WiFi, LTE (cellular), or other emerging technologies, can lower overall systems cost as well as supplement the availability, coverage, and peak data rate requirements of DSRC-based systems. The South Carolina Connected Vehicle Testbed (SC-CVT) is located along a 10-mile segment of Interstate I-85 near Clemson's International Center for Automotive Research (ICAR) campus in Greenville South Carolina. Two specific example CV applications that will be developed are traffic incident detection and queue warning. These two applications provide a convenient starting point for illustrating how CV applications can benefit from advanced network technology that integrates multiple wireless technologies in a CV system. Heterogeneous networks (HetNets) are networks that integrate and exploit multiple concurrently available networking technologies. For critical applications requiring resource allocation optimization in order to meet safety-driven performance requirements. This project is using a combination of Software Defined Networking, local computing provided by GENI racks, and control based on statistical learning theory to demonstrate optimized HetNet operation on the SC-CVT.

​

​

• Incident Management

​

Project title: Integration of the Incident Command System (ICS) Protocol for Effective Coordination of Multi - Agency Response to Traffic Incidents

 

Funding agency: South Carolina Department of Transportation (SCDOT)

 

Publications: Peer-reviewed Journals and conference papers

 

In recent years, there has been an increased focus on Traffic Incident Management (TIM) and incorporation of the Incident Command System (ICS) to reduce traffic congestion on the nation's Interstates. In fact, studies show that for every minute a freeway lane is blocked due to an incident, there is a corresponding time of four minutes of travel delay, and the likelihood of a secondary crash increases by 2.8% for every minute that the primary incident remains a hazard. Efficiencies resulting from adoption of coordinated multi-agency response through ICS can reduce the impact of non-recurrent traffic congestion caused by traffic incidents on major roadways. This project investigated the effectiveness of multi-stakeholder agency coordinated ICS strategies for managing traffic incidents by considering their potential impacts in reducing incident duration. Between 2012 and 2013, there were 129 fatal incidents recorded on SC Interstates. Roughly half of the fatal incident recovery times (dispatch of response to incident cleared) were in excess of six hours, with maximum recovery times well over 12 hours. These clearance and recovery times are significantly higher than those in other states. States which have implemented enhanced ICS and TIM procedures are consistently achieving major incident clearance times of one and a half hours or less. After analyzing the current state-of-practice in South Carolina, and national best practices, the potential areas for improvement in incident response were identified. The TIM areas most in need of improvement were response and clearance strategies, with major gaps noted in towing, coroner, HAZMAT, and crash investigation procedures.

​

• Traffic flow theory

 

Publications: Peer-reviewed Journals and conference papers

​

Microscopic roadway traffic simulators, which attempt to mimic real-world driver behaviors, are based on carfollowing models and have been widely used as a cost-effective tool for intelligent transportation system (ITS) evaluation. In addition to evaluation, ITSs can benefit from accurate car-following models that can provide current estimations and future predictions of various traffic situations to support real-time traffic management. The accuracy and reliability of these applications are greatly dependent on the appropriate calibration of car-following models. In this paper, the authors developed a process to apply a stochastic calibration method with appropriate regularization to estimate the distribution of parameters for car-following models. The calibration method is based on the Markov chain Monte Carlo simulation that uses the Bayesian estimation theory. An intelligent driver model and Helly’s car-following model were utilized to evaluate this method. The Bayesian approach provided better results in terms of the cost function than the deterministic optimization algorithm. With the Bayesian approach, the mean square error per vehicle is decreased with the increased number of vehicles. Analysis also revealed that the Bayesian approach predicted drivers’ speed and acceleration/deceleration profiles more closely to the real-world data compared with the deterministic approach evaluated in this paper. Positive validation outcomes suggest potential efficacy of the calibration approach presented in this paper for future applications.