Sai Chand

Autonomous Vehicles: Disengagements, Accidents and Reaction Times

Autonomous vehicles are being viewed with scepticism in their ability to improve safety and the driving experience. A critical issue with automated driving at this stage of its development is that it is not yet reliable and safe. When automated driving fails, or is limited, the autonomous mode disengages and the drivers are expected to resume manual driving. For this transition to occur safely, it is imperative that drivers react in an appropriate and timely manner. Recent data released from the California trials provide compelling insights into the current factors influencing disengagements of autonomous mode. Here we show that the number of accidents observed has a significantly high correlation with the autonomous miles travelled. The reaction times to take control of the vehicle in the event of a disengagement was found to have a stable distribution across different companies at 0.83 seconds on average. However, there were differences observed in reaction times based on the type of disengagements, type of roadway and autonomous miles travelled. Lack of trust caused by the exposure to automated disengagements was found to increase the likelihood to take control of the vehicle manually. Further, with increased vehicle miles travelled the reaction times were found to increase, which suggests an increased level of trust with more vehicle miles travelled. We believe that this research would provide insurers, planners, traffic management officials and engineers fundamental insights into trust and reaction times that would help them design and engineer their systems.

Long-Range Dependence of Traffic Flow and Speed of a Motorway: Dynamics and Correlation with Historical Incidents

Speed and flow of vehicles tend to have several effects on the dynamics of a transport system. Fluctuations of these variables can implicate congestion, can lower predictability, and may even catalyze crashes. A concept of fractal theory called the Hurst exponent—a measure of the long-range dependence (LRD) of a time series—was used to understand the fluctuations in flow and speed of a motorway in Sydney, Australia. The spatial and temporal variation of the LRD for flow (Hflow) and speed (Hspeed) at several monitor sites is discussed. Furthermore, the effects of number of lanes on flow and speed predictability are explored. It was observed that the flow predictability of two-lane sections was significantly lower when compared with three-lane and four-lane sections. Conversely, the speed predictability of four-lane sections was considerably higher than that of two-lane and three-lane sections. Finally, traffic congestion was defined with regard to the LRD of speed, and its correlation with historical incident rates was measured. It was ascertained that monitor sites with a historically high proportion of large Hspeed were correlated with unsafe locations. This study could lead to many applications of fractal analysis on highways and urban traffic.

Deployment and Calibration Considerations for Large-Scale Regional Dynamic Traffic Assignment: Case Study for Sydney, Australia

Dynamic traffic assignment (DTA) has received increasing attention in recent years, and there are numerous examples of practical implementations. This work adds to the literature by describing the ongoing experience of building the first large-scale simulation-based DTA model in Australia. The input data for the model are summarized, and an in-depth discussion and an analysis of model output and the calibration process are presented. Current results put 80% of the 322 calibration points spread across the network within an acceptable bound of error, but the project found that alternative metrics of network performance also must be considered so that other aspects of model realism are not neglected. The described DTA model could be used for evaluating important policy decisions and infrastructural development in the context of the macro- and mesoscale network operation. Additionally, this project is a proof of concept for the Australian region and may provide insight to practitioners interested in emerging areas of transport planning and traffic modeling.

Evaluation of Fluctuating Speed and Lateral Movement of Vehicles: Comparison Between Mixed Traffic and Homogeneous Traffic

Urban roads in developing countries become congested more often because of the substantial lateral movement of vehicles and fluctuations in speed. Popular microscopic models such as car-following and lane-changing models are not suitable for the analysis of such traffic conditions unless some modifications accounting for heterogeneity are included. However, there seems to be an underlying mechanism behind the fluctuations that needs to be investigated thoroughly. The Hurst exponent concept of chaos theory was used to identify the hidden trends in vehicular movement. Real-life trajectory data from an urban arterial in Chennai, India, were analyzed and then compared with a homogeneous data set in the United States. The Hurst exponent for mixed traffic was found to be significantly less than that for the homogeneous data; this finding indicated the strong trends of lateral movement and speed in homogeneous traffic. The variation of Hurst exponents with vehicle type and average lateral positions was explored in mixed traffic. Results from this study will help modelers propose better microscopic simulation models accounting for the fluctuations in speed and lateral movement.

Framework for Estimating the Impact of Camera-Based Intelligent Transportation Systems (ITS) Technology on Incident Duration:

Safety is a major motivator of intelligent transportation systems (ITS) projects, and most efforts have addressed the potential to avoid incidents. Managing and reducing the duration of incidents is another key application for ITS despite challenges in distinguishing the true versus the reported duration of an incident. This paper presents a framework for modeling the impact of camera-based (closed-circuit television or CCTV) ITS technology on incident duration including an increase in the reported duration and a reduction in the true duration. The framework is validated against a data set of 121,793 accidents in New South Wales, Australia, covering 4.5 years. The results demonstrate that the use of CCTVs for incident duration contributes a 4.5 min reduction in average duration (as earlier detection can lead to more efficient clearance) and a 9% reduction in variance in the duration (as a uniform detection method supports standardized response procedures). These impacts are only visible when the 8.5 min median detection delay (the difference between the recorded duration and the true duration) is modeled and accounted for. These results offer a quantitative support tool for decision makers wishing to assess the value of incident-detection ITS projects.

Application of Fractal theory for crash rate prediction: Insights from random parameters and latent class tobit models

The repercussions from congestion and accidents on major highways can have significant negative impacts on the economy and environment. It is a primary objective of transport authorities to minimize the likelihood of these phenomena taking place, to improve safety and overall network performance. In this study, we use the Hurst Exponent metric from Fractal Theory, as a congestion indicator for crash-rate modeling. We analyze one month of traffic speed data at several monitor sites along the M4 motorway in Sydney, Australia and assess congestion patterns with the Hurst Exponent of speed (Hspeed). Random Parameters and Latent Class Tobit models were estimated, to examine the effect of congestion on historical crash rates, while accounting for unobserved heterogeneity. Using a latent class modeling approach, the motorway sections were probabilistically classified into two segments, based on the presence of entry and exit ramps. This will allow transportation agencies to implement appropriate safety/traffic countermeasures when addressing accident hotspots or inadequately managed sections of motorway.

Large-Scale Dynamic Traffic Assignment: Practical Lessons from an Application in Sydney, Australia

Traditional static traffic assignment models no longer meet the strategic planning needs of most major metropolitan areas, especially in regard to evaluating major infrastructure projects. One promising possibility is dynamic traffic assignment (DTA), which has been receiving greater attention in the research community for the last ten years. This work describes the ongoing experience of building the first large-scale DTA model in Australia. We divide our experiences into categories regarding data, implementation, and visualization, and we discuss the challenges faced as well as our methods for overcoming those challenges. Finally, we discuss initial model results and the calibration process. In the future, the DTA model described here could aid in evaluating important policy decisions and infrastructural development in the context of the macro/meso-scale network operation. This project serves as a proof of concept for the Australia region and may provide valuable insight to other practitioners interested in emerging areas of transport planning and traffic modeling.