Browsing by Author "Lu, Pan"

Now showing 1 - 11 of 11

Detecting Sources of Ride Roughness by Ensemble Connected Vehicle Signals
(2022) Bridgelall, Raj; Bhardwaj, Bhavana; Lu, Pan; Tolliver, Denver D.; Dhingra, Neeraj; Upper Great Plains Transportation Institute
It is expensive and impractical to scale existing methods of road condition monitoring for more frequent and network-wide coverage. Consequently, defects that increase ride roughness or can cause accidents will go undetected. This paper presents a method to enable network-wide, continuous monitoring by using low-cost GPS receivers and accelerometers on board regular vehicles. The technique leverages the large volume of sensor signals from multiple traversals of a road segment to enhance the signal quality by ensemble averaging. However, ensemble averaging requires position-repeatable signals which is not possible because of the low resolution and low accuracy of GPS receivers and the non-uniform sampling of accelerometers. This research overcame those challenges by integrating methods of interpolation, signal resampling, and correlation alignment. The experiments showed that the approach doubled the peak of the composite signal by decreasing signal misalignment by a factor of 67. The signal-to-noise ratio increased by 10 dBs after combining the signals from only 6 traversals. A probabilistic model developed to estimate a dynamic signal-detection threshold demonstrated that both the false-positive and false-negative rates approached zero after combining the signals from 15 traversals. The method will augment the efficiency of follow-up inspections by focusing resources to locations that consistently produce rough rides.
Enhancement of Signals from Connected Vehicles to Detect Roadway and Railway Anomalies
(2019) Bridgelall, Raj; Chia, Leonard; Bhardwaj, Bhavana; Lu, Pan; Tolliver, Denver D.; Dhingra, Neeraj; Upper Great Plains Transportation Institute
Frequent network-wide monitoring of the condition of roadways and railways prevent fatalities, injuries, and financial losses. Even so, agencies cannot afford to inspect vast transportation networks using present methods. Therefore, the idea of using low-cost sensors aboard connected vehicles became appealing. However, low-cost sensors introduce new challenges to improve poor signal quality which causes detection errors. Common approaches apply computationally complex filters to individual signal streams, which limits further improvements. This paper presents a method that combines signals from each traversal in a manner that leads to ever-increasing signal quality. The proposed method addresses the challenges of poor accuracy and precision of position estimates from global positioning system (GPS) receivers, and errors from the non-uniform sampling of low-cost accelerometers. The result is improved signal quality from a 20% improvement in signal alignment over GPS and a 90-fold enhancement in distance precision.
Mining Connected Vehicle Data for Beneficial Patterns in Dubai Taxi Operations
(2018) Bridgelall, Raj; Lu, Pan; Tolliver, Denver D.; Xu, Tie; Upper Great Plains Transportation Institute
On-demand shared mobility services such as Uber and micro-transit are steadily penetrating the worldwide market for traditional dispatched taxi services. Hence, taxi companies are seeking ways to compete. This study mined large-scale mobility data from connected taxis to discover beneficial patterns that may inform strategies to improve dispatch taxi business. It is not practical to manually clean and filter large-scale mobility data that contains GPS information. Therefore, this research contributes and demonstrates an automated method of data cleaning and filtering that is suitable for such types of datasets. The cleaning method defines three filter variables and applies a layered statistical filtering technique to eliminate outlier records that do not contribute to distributions that match expected theoretical distributions of the variables. Chi-squared statistical tests evaluate the quality of the cleaned data by comparing the distribution of the three variables with their expected distributions. The overall cleaning method removed approximately 5% of the data, which consisted of errors that were obvious and others that were poor quality outliers. Subsequently, mining the cleaned data revealed that trip production in Dubai peaks for the case when only the same two drivers operate the same taxi. This finding would not have been possible without access to proprietary data that contains unique identifiers for both drivers and taxis. Datasets that identify individual drivers are not publicly available.
Modeling Pavement Performance and Preservation
(North Dakota State University, 2011) Lu, Pan
The number of highway lane-miles in the United States increased by 7% from 1980 to 2007, while vehicle-miles of travel almost doubled. During the same period, the Federal Highway Trust Fund (the major source of funding for highways) grew by only 40% in constant 1980 dollars. With growth in trade and commerce, truck traffic levels are expected to increase significantly in the future. Highway agencies throughout the United States are facing complex decisions about maintaining, repairing, and renewing existing pavements in the most cost-effective ways. Decision makers need to learn: to what degrees different pavement preservation treatments will improve a pavement condition; how pavement conditions will change over time; when to apply which treatment to what section; and what budget level will be needed to maintain and improve pavement conditions. The objectives of this dissertation are to 1) estimate the effectiveness of appropriate different levels of pavement preservation treatments, 2) evaluate pre-treatment and posttreatment pavement performances, and 3) use the uniformed results (of the first two objectives) to develop a decision making tool for integrated pavement management systems. The dissertation will utilize data from the Long Term Pavement Performance (LTPP) program. LTPP data will be used to estimate statistical models of the benefit effectiveness of preservation-related treatments and pavement performance, including models of performance jump--i.e., the instantaneous improvement in the performance or condition of a pavement due to a maintenance treatment. The forecast values from the statistical models will be used as inputs to optimization models that will allow for the simultaneous solution of several objectives or constraints. The results will benefit pavement management systems and improve pavement preservation planning in the United States.
Railroad Track Condition Monitoring Using Inertial Sensors and Digital Signal Processing: A Review
(2018) Chia, Leonard; Bhardwaj, Bhavana; Lu, Pan; Bridgelall, Raj; Upper Great Plains Transportation Institute
Inertial sensors such as accelerometers and gyroscopes have been widely used since the early 1990s to monitor the condition of transportation assets. Recent improvements in their performance, a reduction in cost, and sensor miniaturization has resulted in a growing interest expanding their use. This research is an extensive and systematic review of their application considerations, challenges, and opportunities for improvements in railroad track condition monitoring. Research questions were developed to guide the selection of relevant articles from databases. The authors report key findings in the areas of sensor specification, sensor location, and sensor signal processing.
Ranking Risk Factors in Financial Losses From Railroad Incidents: A Machine Learning Approach
(2023) Dhingra, Neeraj; Bridgelall, Raj; Lu, Pan; Szmerekovsky, Joseph; Bhardwaj, Bhavana; Upper Great Plains Transportation Institute
The reported financial losses from railroad accidents since 2009 have been more than US$4.11 billion dollars. This considerable loss is a major concern for the industry, society, and the government. Therefore, identifying and ranking the factors that contribute to financial losses from railroad accidents would inform strategies to minimize them. To achieve that goal, this paper evaluates and compares the results of applying different non-parametric statistical and regression methods to 15 years of railroad Class I freight train accident data. The models compared are random forest, k-nearest neighbors, support vector machines, stochastic gradient boosting, extreme gradient boosting, and stepwise linear regression. The results indicate that these methods are all suitable for analyzing non-linear and heterogeneous railroad incident data. However, the extreme gradient boosting method provided the best performance. Therefore, the analysis used that model to identify and rank factors that contribute to financial losses, based on the gain percentage of the prediction accuracy. The number of derailed freight cars and the absence of territory signalization dominated as contributing factors in more than 57% and 20% of the accidents, respectively. Partial-dependence plots further explore the complex non-linear dependencies of each factor to better visualize and interpret the results.
Rolling-Stock Automatic In-Situ Line Deterioration and Operating Condition Sensing
(2013) Bridgelall, Raj; Lu, Pan; Tolliver, Denver D.; Upper Great Plains Transportation Institute
Track and equipment failures dominate railroad accident causes. Railroads must visually inspect most tracks in service as often as twice weekly to comply with the Federal Track Safety Standards. They augment visual inspections with automated non-destructive-evaluation (NDE) equipment to locate developing and mature defects. However, the defect formation rate is escalating with increasing traffic load density and continuously declining railroad employment per track-mile. This indicates a widening gap between the rate of defect formation and the resources available to find them before they result in accidents, delays, and lost revenue. With resources thinly stretched and the rate of defect formation escalating with traffic load-density, railroads are seeking to enhance the efficiency of inspections and maintenance of way. This paper describes the development of a Rolling-stock Automatic In-situ Line Deterioration & Operating Condition Sensing (RAILDOCS) system to automatically locate and classify track and rail vehicle defects. The approach incorporates a new low-cost wireless sensor technology and Cloud computing method to guide and focus inspection activities to locations of equipment and track defect symptoms, leading to efficient diagnosis and remediation. RAILDOCS has on-board sensors which will continuously monitor track and vehicle condition and transmit a 3D inertial signature for a remote processor to analyze and produce a complete and updated picture of aggregate track and equipment quality. RAILDOCS complement more expensive visual and NDE methods by reallocating time spent on defect discovery to detailed inspections of prioritized defect symptom locations. Symptom sensors integrate micro-electro-mechanical (MEMS), global positioning system (GPS) satellite receivers, wireless communications, and microprocessors technology. Cloud computing and signal processing algorithms produce a track quality index, and forecast optimum maintenance triggers.
Signal Feature Extraction and Combination to Enhance the Detection and Localization of Railroad Track Irregularities
(2020) Bhardwaj, Bhavana; Bridgelall, Raj; Lu, Pan; Dhingra, Neeraj; Upper Great Plains Transportation Institute
Tracks are critical and expensive railroad asset, requiring frequent maintenance. The stress from heavy car axle loads increases the risk of deviations from uniform track geometry. Irregularities in track geometry, such as track warping, can cause an excessive harmonic rocking condition that can lead to derailments, traffic delays, and associated financial losses. This paper presents an approach to enhance the location identification accuracy of track geometry irregularities by combining measurements from sensors aboard Hi-Rail vehicles. However, speed variations, position recording errors, low GPS update rates, and the non-uniform sampling rates of inertial sensors pose significant challenges for signal processing, feature extraction, and signal combination. This study introduces a method of extracting features from the fused data of inertial sensors and GPS receivers with multiple traversals to locate and characterize irregularities of track geometry. The proposed method provides robust detection and enhanced accuracy in the localization of irregularities within spatial windows along the track segment. Tradeoff analysis found that the optimal spatial window size is 5-meter.
Signal Filter Cut-off Frequency Determination to Enhance the Accuracy of Rail Track Irregularity Detection and Localization
(2019) Bhardwaj, Bhavana; Bridgelall, Raj; Chia, Leonard; Lu, Pan; Dhingra, Neeraj; Upper Great Plains Transportation Institute
A continuous condition monitoring system to detect and localize railroad track irregularities is achievable with inertial sensors onboard revenue service trains. However, the inaccurate geospatial position estimates of GPS receivers and the non-uniform sampling of inertial sensors adds noise and reduces signal strength. Consequently, the signal-to-noise ratio decreases, which leads to higher rates of false positives and false negatives. Appropriate signal filtering, alignment, and combination from multiple traversals can enhance the signal-to-noise ratio. However, it is not straightforward to determine the best cut-off frequency for the filter. This paper introduces a method that is suitable for any signal filtering approach. The frequency window of the resultant energy and variance of ensemble averaged FFTs informs the best cut-off frequency. The results affirm that a lowpass finite impulse response filter with the selected cutoff frequency progressively increases the signal-to-noise ratio with increasing filter order, thus demonstrating the effectiveness and practicality of the method.
Technology Developments and Impacts of Connected and Autonomous Vehicles: An Overview
(2022) Ahmed, Hafiz Usman; Huang, Ying; Lu, Pan; Bridgelall, Raj; Upper Great Plains Transportation Institute
The scientific advancements in the vehicle and infrastructure automation industry are progressively improving nowadays to provide benefits for the end-users in terms of traffic congestion reduction, safety enhancements, stress-free travels, fuel cost savings, and smart parking, etc. The advances in connected, autonomous, and connected autonomous vehicles (CV, AV, and CAV) depend on the continuous technology developments in the advanced driving assistance systems (ADAS). A clear view of the technology developments related to the AVs will give the users insights on the evolution of the technology and predict future research needs. In this paper, firstly, a review is performed on the available ADAS technologies, their functions, and the expected benefits in the context of CVs, AVs, and CAVs such as the sensors deployed on the partial or fully automated vehicles (Radar, LiDAR, etc.), the communication systems for vehicle-to-vehicle and vehicle-to-infrastructure networking, and the adaptive and cooperative adaptive cruise control technology (ACC/CACC). Secondly, for any technologies to be applied in practical AVs related applications, this study also includes a detailed review in the state/federal guidance, legislation, and regulations toward AVs related applications. Last but not least, the impacts of CVs, AVs, and CAVs on traffic are also reviewed to evaluate the potential benefits as the AV related technologies penetrating in the market. Based on the extensive reviews in this paper, the future related research gaps in technology development and impact analysis are also discussed.
Weigh-In-Motion System in Flexible Pavements Using Fiber Bragg Grating Sensors Part A: Concept
(2019) Al-Tarawneh, Mu'ath; Huang, Ying; Lu, Pan; Bridgelall, Raj; Upper Great Plains Transportation Institute
Weight data of vehicles play an important role in traffic planning, weight enforcement, and pavement condition assessment. In this paper, a weigh-in-motion (WIM) system that functions at both low-speeds and high-speeds in flexible pavements is developed based on in-pavement, three-dimensional glass-fiber-reinforced, polymer-packaged fiber Bragg grating sensors (3D GFRP-FBG). Vehicles passing over the pavement produce strains that the system monitors by measuring the center wavelength changes of the embedded 3D GFRP-FBG sensors. The FBG sensor can estimate the weight of vehicles because of the direct relationship between the loading on the pavement and the strain inside the pavement. A sensitivity study shows that the developed sensor is very sensitive to sensor installation depth, pavement property, and load location. Testing in the field validated that the longitudinal component of the sensor if not corrected by location has a measurement accuracy of 86.3% and 89.5% at 5 mph and 45 mph vehicle speed, respectively. However, the system also has the capability to estimate the location of the loading position, which can enhance the system accuracy to more than 94.5%.