Highway maintenance management is a hot area for innovation when it comes to using technology, machine learning/AI, and data to optimize operations and manage system condition. This poster session highlights some of the latest advancements.
Use of Mobile Photogrammetry Method for Highway Asset Management
Mohammad Farhadmanesh, University of UtahShow Abstract
Chandler Cross, University of Utah
Ali Hassandokht Mashhadi, University of Utah
Abbas Rashidi, University of Utah
Jessica Wempen, University of Utah
Highway asset condition is of the utmost importance when it comes to transportation maintenance and pedestrian safety. Transportation facility managers must have up-to-date information on the status of all transportation assets to keep the transportation facilities operating at their highest level. Due to the sheer volume of transportation assets, an efficient and affordable data collection procedure is necessary to gather important data and create an asset inventory. Some pioneer departments of transportation in the United States decided to use mobile LiDAR (Light Detection and Ranging) to monitor highway assets and pavement condition data. Not only is the laser scanning equipment expensive, but the operator in charge of using the equipment and processing the point clouds must have special technical knowledge that may not be accessible to every individual. More recently, image-based reconstruction, known as photogrammetry, has emerged as a cheaper and simpler technology than LiDAR. Image-based reconstruction can be done using a camera such as a DSLR camera or even a smartphone. This paper presents a full review of various conducted research studies on highway asset management and pavement condition assessment using spatial data modelings such as LiDAR and photogrammetry. The two case studies presented in this work are in the same categories previously mentioned, and data are evaluated and compared between both technologies, LiDAR and photogrammetry. Point clouds were evaluated in terms of density, generated sign size accuracy, and generated sign density. Limitations, challenges, benefits, and recommendations are discussed.
Trust in Automated Truck-Mounted Attenuators: A Survey on Worker Perceptions
Shiva Pourfalatoun, Colorado State University, Fort CollinsShow Abstract
Erika Miller (email@example.com), Colorado State University
Automated Truck-Mounted Attenuators (ATMAs) have the potential to improve work zone safety by removing a human driver from a vehicle intended for impact. However, this technology is expensive and can be detrimental to safety and project success if used incorrectly. Therefore it is important to understand user's perceptions of ATMAs and how training can improve appropriate adoption of this technology. The objective of this study was to evaluate how work zone worker's perceive the usefulness and understand the capabilities of automation in Truck-Mounted Attenuators. A survey was administered to 13 Department of Transportation workers that had used and been trained on ATMAs. Workers reported an overall positive acceptance of this technology, such as expected reductions in crash severity, reasonable workload associated with operating procedures for the automation, and trust in the automation's reliability. However, they noted concerns regarding their trust in the automation under various contexts, such as poor visibility and denser traffic volumes. Trust in the technology was greatest among workers with higher levels of ATMA training and longer experience working with the ATMA. This research presents a novel perspective on user acceptance of ATMA technology. These findings can help jurisdictions achieve the safety improvements that deployment of automation in work zones offers.
Traffic Sign Extraction using Deep Hierarchical Feature Learning and LiDAR Data on Rural Highways
Maged Gouda (firstname.lastname@example.org), University of AlbertaShow Abstract
Alexander Epp, University of Alberta
Ryden Graham, University of Alberta
Karim El-Basyouny, University of Alberta
The application of deep learning techniques on point cloud data holds significant promise for efficient data segmentation and classification of traffic signs. This study proposes adjustments to the PointNet++ neural network and leverages the use of local geometric features in the training process. Further, models with different combinations of intensity, roughness, z-gradient, and different point densities were trained using labeled data from seven highway segments in Alberta, Canada. The results indicate that the model combining intensity and z-gradient with the lower point density significantly outperformed all other models in terms of precision, recall, and F1-score. More so, there was improved performance in accuracy and processing times compared to previous studies on sign detection using point cloud data. The overall per sign detection performance shows a 99.2% recall (98% per point) and a 98% F1-score (97% per point). The intensity with roughness combination yielded lower per sign performance compared to intensity-only and intensity in combination with z-gradient models. Overall, the inclusion of z-gradient significantly increased sign detection in terms of precision, recall, and F1-score, by 9%, 4.9%, and 7.1%, respectively, allowing the model to yield notable performance improvements for outdoor scene recognition. Comparison was also made with existing sign detection methods on the Paris-Lille-3D benchmark, finding higher recall rates than existing studies. The proposed approach suggests that with adjustments, the PointNet++ neural network architecture can achieve remarkable results on large metric scale scenes for sign extraction using point cloud data.
Multi Asset Hotspot Analysis for Predictive Roadway Maintenance
Arash Karimzadeh (email@example.com), University of North Carolina, CharlotteShow Abstract
Sepehr Sabeti, University of North Carolina, Charlotte
Hamed Tabkhi, University of North Carolina, Charlotte
Omidreza Shoghli, University of North Carolina, Charlotte
Given multiple budget and revenue constraints that transportation sector encounters, predictive analytics alongside adequate data collection enable maintenance agencies to make effective decisions, prioritize maintenance tasks and provide efficient life-cycle planning. To this end, risk-based predictive models provide promising capability in representing the susceptibility of assets to future defects. Therefore, the main objective of this study is to provide an integrated framework for predicting the occurrence probability of multiple defects on different highway asset types. We identified several gaps in previous models, including limitations in predictive frameworks given the inadequate scope of available inspection data, expert-based selection of contributing factors, and ignoring the interrelationships between neighboring assets. We, therefore, propose a risk-based method that combines a risk score generator and a machine learning algorithm to predict the hotspots of multiple defects in a given roadway. We then chose a case study to measure the efficiency of the proposed model. Our framework produced a significant accuracy in predicting future risk scores of erosion, obstruction, and cracking on paved ditches given historical weather, traffic, maintenance, and inspection data of three selected neighboring assets (unpaved ditches, slopes, small pipes and box culverts). Besides, we investigated the contribution of the considered factors to further study the importance of individual contributors. The framework offers decision makers a holistic view of degradation risks of multiple assets, which could enable them to prepare an integrated asset management program. Additionally, a similar framework can be applied to other linear infrastructure systems such as water networks and railroads.
COVID-19: Effect of the Pandemic on Work Zone Schedule Optimization
Celina Semaan, New Jersey Institute of TechnologyShow Abstract
Steven Chien, New Jersey Institute of Technology
The outbreak of the novel coronavirus COVID-19 has caused enormous impacts on various social, economic, and health spectrums. Rapid decline in the mobility trends have been detected with the declaration of the stay-at-home order on March 21 in New Jersey, upon which non-essential workers started operating remotely. The traffic pattern started shifting with the lift of the stay-at-home order on June 9. While maintenance projects are deemed essential in NJ (1), transportation agencies are investigating ways to work in a safe environment, yet finish their maintenance projects on time for the reopening stage. This paper investigates the effect of traffic demand variations during COVID-19 lockdown on work zone optimization using the Artificial Bee Colony algorithm. A case study was conducted to evaluate the optimized schedules in three different phases. The first phase depicts the normal traffic conditions, the second phase reflects the lockdown conditions due to the declaration of the stay-at-home order, and the third outlines the traffic conditions after lifting the order. INRIX data is used to estimate the decrease in traffic volumes during those phases. The results indicate that work zone optimization during COVID-19 lockdown period may significantly reduce the project duration while saving on the project total cost. In addition, the crew assigned during lockdown period may consist of few workers, which allows social distancing measures. The findings of this study can assist transportation agencies with managing the work zone optimization scheduling problem during current and any future lockdown conditions.
Joint optimization of multi-scale decisions in budget allocation, inspection frequency, and maintenance policies for transportation infrastructure systems
Heeseung Shon, Korea Advanced Institute of Science and Technology (KAIST)Show Abstract
Jinwoo Lee (firstname.lastname@example.org), Korea Advanced Institute of Science and Technology (KAIST)
We present a transportation infrastructure management framework that involves system-level budget allocation, group-level inspection, and segment-level maintenance, rehabilitation, and reconstruction (MR&R) strategies. The innovation in inspection technologies has enabled us to investigate the condition of a group of multiple sections of infrastructure almost simultaneously by an instrumented vehicle that can travel at the same speed as traffic. Previous related studies have mainly focused on at most a bi-level structure, with all the managerial activities such as inspection and MR&R at the segment level and allocation of a pooled budget at the system level. However, this cannot reflect the current group-level inspection practice, positioned between the segment and system levels. The proposed methodology adopts a bottom-up approach to account for heterogeneous segment- and group-specific properties and optimally determines multi-scale decisions to minimize the discounted societal costs. We formulate the multi-scale problem as a combination of: (i) a Markov Decision Process and (ii) approximate convex programming. The finding confirms that the optimal inspection period will increase with inspection costs and that frequent inspections are necessary if the expected lifetime costs of the object are huge. As a numerical study, we analyzed a roadway network including three expressway groups and one local road group near Daejeon City in Korea. The results reveal that a road group under heavier traffic and faster deterioration needs higher inspection frequency, and segments included in a group need different thresholds of condition-based MR&R policies depending on their current condition states and other segment-specific influential factors.
Framework for Leveraging Data from Autonomous Trucks to Support State Maintenance Operations
Morgan Avera (email@example.com), University of Texas, AustinShow Abstract
C. Michael Walton, University of Texas, Austin
An urgent need to move goods safely and efficiently has propelled the development of autonomous trucks (ATs). Using a combination of sensors, computing can replace a human driver for long, dull stretches of highway driving. As ATs are increasingly deployed, they are collecting massive amounts of information which could be valuable to those managing and maintaining roadways. Simultaneously, state agencies are struggling to maintain the roadways which provide vital connectivity. Routine maintenance addresses day to day concerns which are often the hardest to track considering the lack of predictability. While it is easy to replace a fallen sign, it is often hard to know when a sign has fallen. The data being collected by newly deployed ATs can be leveraged to assist in identifying routine maintenance concerns, enabling state agencies to keep their roads in a better condition. This study lays out how state agencies can implement a data-sharing framework to leverage the operation of ATs on their roadways. Working together, a platform can be built into existing systems that allows AT companies to report maintenance concerns identified during operation. The reports generated would have higher veracity than typical reporting mechanisms as ATs are equipped with high-quality cameras. Using data provided by Kodiak Robotics, a prototype mapping module was created to showcase how this system would work. Input was provided by private sector and public sector representatives, with both groups agreeing that there is valuable data available which should be leveraged by the state.
DISCLAIMER: All information shared in the TRB Annual Meeting Online Program is subject to change without notice. Changes, if necessary, will be updated in the Online Program and this page is the final authority on schedule information.