Fangzhou Sun

Real-Time and Predictive Analytics for Smart Public Transportation Decision Support System

Public bus transit plays an important role in city transportation infrastructure. However, public bus transit is often difficult to use because of lack of real- time information about bus locations and delay time, which in the presence of operational delays and service alerts makes it difficult for riders to predict when buses will arrive and plan trips. Precisely tracking vehicle and informing riders of estimated times of arrival is challenging due to a number of factors, such as traffic congestion, operational delays, varying times taken to load passengers at each stop. In this paper, we introduce a public transportation decision support system for both short-term as well as long-term prediction of arrival bus times. The system uses streaming real-time bus position data, which is updated once every minute, and historical arrival and departure data - available for select stops to predict bus arrival times. Our approach combines clustering analysis and Kalman filters with a shared route segment model in order to produce more accurate arrival time predictions. Experiments show that compared to the basic arrival time prediction model that is currently being used by the city, our system reduces arrival time prediction errors by 25% on average when predicting the arrival delay an hour ahead and 47% when predicting within a 15 minute future time window.

Transit-hub: a smart public transportation decision support system with multi-timescale analytical services

Public transit is a critical component of a smart and connected community. As such, citizens expect and require accurate information about real-time arrival/departures of transportation assets. As transit agencies enable large-scale integration of real-time sensors and support back-end data-driven decision support systems, the dynamic data-driven applications systems (DDDAS) paradigm becomes a promising approach to make the system smarter by providing online model learning and multi-time scale analytics as part of the decision support system that is used in the DDDAS feedback loop. In this paper, we describe a system in use in Nashville and illustrate the analytic methods developed by our team. These methods use both historical as well as real-time streaming data for online bus arrival prediction. The historical data is used to build classifiers that enable us to create expected performance models as well as identify anomalies. These classifiers can be used to provide schedule adjustment feedback to the metro transit authority. We also show how these analytics services can be packaged into modular, distributed and resilient micro-services that can be deployed on both cloud back ends as well as edge computing resources.

DelayRadar: A multivariate predictive model for transit systems

Effective public transit operations are one of the fundamental requirements for a modern community. Recently, a number of transit agencies have started integrating automated vehicle locators in their fleet, which provides a real-time estimate of the time of arrival. In this paper, we use the data collected over several months from one such transit system and show how this data can be potentially used to learn long term patterns of travel time. More specifically, we study the effect of weather and other factors such as traffic on the transit system delay. These models can later be used to understand the seasonal variations and to design adaptive and transient transit schedules. Towards this goal, we also propose an online architecture called DelayRadar. The novelty of DelayRadar lies in three aspects: (1) a data store that collects and integrates real-time and static data from multiple data sources, (2) a predictive statistical model that analyzes the data to make predictions on transit travel time, and (3) a decision making framework to develop an optimal transit schedule based on variable forecasts related to traffic, weather, and other impactful factors. This paper focuses on identifying the model with the best predictive accuracy to be used in DelayRadar. According to the preliminary study results, we are able to explain more than 70% of the variance in the bus travel time and we can make future travel predictions with an out-of-sample error of 4.8 minutes with information on the bus schedule, traffic, and weather.

Unsupervised Mechanisms for Optimizing On-Time Performance of Fixed Schedule Transit Vehicles

The on-time arrival performance of vehicles at stops is a critical metric for both riders and city planners to evaluate the reliability of a transit system. However, it is a non-trivial task for transit agencies to adjust the existing bus schedule to optimize the on-time performance for the future. For example, severe weather conditions and special events in the city could slow down traffic and cause bus delay. Furthermore, the delay of previous trips may affect the initial departure time of consecutive trips and generate accumulated delay. In this paper, we formulate the problem as a single-objective optimization task with constraints and propose a greedy algorithm and a genetic algorithm to generate bus schedules at timepoints that improve the bus on-time performance at timepoints which is indicated by whether the arrival delay is within the desired range. We use the Nashville bus system as a case study and simulate the optimization performance using historical data. The comparative analysis of the results identifies that delay patterns change over time and reveals the efficiency of the greedy and genetic algorithms.

SpeedPro: A Predictive Multi-Model Approach for Urban Traffic Speed Estimation

Data generated by GPS-equipped probe vehicles, especially public transit vehicles can be a reliable source for traffic speed estimation. Traditionally, this estimation is done by learning the parameters of a model that describes the relationship between the speed of the probe vehicle and the actual traffic speed. However, such approaches typically suffer from data sparsity issues. Furthermore, most state of the art approaches does not consider the effect of weather and the driver of the probe vehicle on the parameters of the learned model. In this paper, we describe a multivariate predictive multi-model approach called SpeedPro that (a) first identifies similar clusters of operation from the historic data that includes the real-time position of the probe vehicle, the weather data, and anonymized driver identifier, and then (b) uses these different models to estimate the traffic speed in real-time as a function of current weather, driver and probe vehicle speed. When the real-time information is not available our approach uses a different model that uses the historical weather and traffic information for estimation. Our results show that the purely historical data is less accurate than the model that uses the real-time information.

A Feasibility Study of Autonomically Detecting In-Process Cyber-Attacks

A cyber-attack detection system issues alerts when an attacker attempts to coerce a trusted software application to perform unsafe actions on the attackers behalf. One way of issuing such alerts is to create an application-agnostic cyber- attack detection system that responds to prevalent software vulnerabilities. The creation of such an autonomic alert system, however, is impeded by the disparity between implementation language, function, quality-of-service (QoS) requirements, and architectural patterns present in applications, all of which contribute to the rapidly changing threat landscape presented by modern heterogeneous software systems. This paper evaluates the feasibility of creating an autonomic cyber-attack detection system and applying it to several exemplar web-based applications using program transformation and machine learning techniques. Specifically, we examine whether it is possible to detect cyber-attacks (1) online, i.e., as they occur using lightweight structures derived from a call graph and (2) offline, i.e., using machine learning techniques trained with features extracted from a trace of application execution. In both cases, we first characterize normal application behavior using supervised training with the test suites created for an application as part of the software development process. We then intentionally perturb our test applications so they are vulnerable to common attack vectors and then evaluate the effectiveness of various feature extraction and learning strategies on the perturbed applications. Our results show that both lightweight on-line models based on control flow of execution path and application specific off-line models can successfully and efficiently detect in-process cyber-attacks against web applications.

Poster Abstract: A Distributed and Resilient Platform for City-Scale Smart Systems

The advent of the Internet of Things (IoT) is driving several technological trends. The first trend is an increased level of integration between edge devices and commodity computers. This trend, in conjunction with low power-devices, energy harvesting, and improved battery technology, is enabling the next generation of information technology (IT) innovation: city-scale smart systems. These types of IoT systems can operate at multiple time-scales, ranging from closed-loop control requiring strict real-time decision and actuation to near real-time operation with humans-in-the-loop, as well as to long-term analysis, planning, and decision-making.