Multi-Task Learning for iRAP Attribute Classification and Road Safety Assessment

Kačan, Marin; Oršić, Marin; Šegvić, Siniša; Ševrović, Marko

doi:10.1109/itsc45102.2020.9294305

Cited by 6 publications

(5 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The ratings indicate that the risk of death or serious injury is highest on a 1-star road and lowest on a 5-star road. SRS are also particularly useful in enabling evidence-based decisions on road improvements (43). SRS are based on road engineering design features.…”

Section: Ranking Methodsmentioning

confidence: 99%

Identifying Optimal Locations for Speed Enforcement Cameras

Mohammad

Eleimat²

2022

Transportation Research Record: Journal of the Transportation R

View full text Add to dashboard Cite

Speed enforcement cameras are used to improve traffic safety. These cameras are usually placed in areas with significant crashes related to speeding. However, in many countries, detailed crash data are not available or are of poor quality. This study aims to propose and explore three simple ranking methods for identifying optimal locations for speed enforcement cameras that do not depend on crash records. The first method is based on the Star Rating Scores and the traffic volume data. The second method relies on developing a data-driven model for predicting speeding violation rate (VR) based on road geometry aspects and comparing it with available road segment crash risk. The third method uses a combination of the first two methods. The proposed methods and prediction models were applied to a case study and successfully identified the optimal location for speed enforcement cameras. The developed prediction models suggest that speed limit, junction spacing, and the number of access points per kilometer are the most common variables to estimate speed VRs for both urban and rural roads. Availability of service roads and the number of lanes were variables that affect the speed VR on urban roads only. The proposed methods are considered valuable simple tools for road safety practitioners to effectively prepare speed enforcement plans for different road networks.

show abstract

Section: Ranking Methodsmentioning

confidence: 99%

Identifying Optimal Locations for Speed Enforcement Cameras

Mohammad

Eleimat²

2022

Transportation Research Record: Journal of the Transportation R

View full text Add to dashboard Cite

show abstract

“…Their experiments showed that using MTL with the CNN-based VGG architecture helps enhance the precision of road rating and improve the estimation of some road risk attributes, including roadside hazards and lane width, by utilizing attention layers. Additionally, MTL based on CNN could improve the performance of iRAP, a charity aiming to rate road safety and enhance road assessment, by concurrently training iRAP attributes derived from segments of sequential video input [65]. The attributes of the iRAP methodology are divided into seven categories: road details, observed flow, speed limit, road middle-side objects, roadside objects, intersections, and road user vulnerability factors such as pedestrians.…”

Section: ) Road Assessmentmentioning

confidence: 99%

“…Using MTL is effective in concurrently predicting the semantic segmentation for labeling each object in the image, depth estimation for objects distance measurement, light detection and ranging (LiDAR) Using usRAP with auxiliary tasks such as intersection, lane numbers, and road conditions based on MTL increases the road safety assessment. 2020 Kačan et al [65] Training concurrently iRAP attributes based on MTL enhances the rating of the road.…”

Section: ) Surrounding Cars Detectionmentioning

confidence: 99%

Survey on Multi-Task Learning in Smart Transportation

Alzahrani,

Wang,

Liao

et al. 2024

IEEE Access

View full text Add to dashboard Cite

Artificial Intelligence (AI) has been widely adopted in numerous fields and enabled various smart systems because of its strong ability to perform tasks, including prediction, event detection, and status estimation, among others. As one of the typical smart systems empowered by AI and Internet of Things (IoT) technologies, the smart transportation system has made dramatic progress for traditional transportation in numerous aspects, including autonomous driving, smart traffic lights, navigation, and traffic forecasting, among others. Deep learning is an essential component to enable such smart systems. Typically, specific deep learning models, e.g., Recurrent Neural Networks (RNN) and Convolutional Neural Networks (CNN), can be trained on collected transportation data for a particular task. However, traditional deep learning techniques rely on data sufficiency to build an effective model. Additionally, each trained model can only work on one single task. This has limited the efficacy of deep learning techniques in numerous application scenarios. To this end, multi-task learning (MTL) has been studied to train a single model that can work for multiple tasks. This technique effectively allows the learning model to expand to specific tasks with only a limited amount of data affiliated. In the meantime, MTL significantly reduces the training time of each task. The success of MTL requires that there are potential relationships among different tasks. Many tasks in the smart transportation system are related. For instance, traffic speed and vehicle volume estimations for each road are highly correlated. Based on this, research on applying MTL in smart transportation systems has been studied recently. This paper reviews the recent efforts to use MTL in smart transportation systems and conducts an extensive survey to provide insights. In particular, we categorize the MTL applications in smart transportation systems into traffic forecasting, traffic sign recognition, vehicle recognition, travel time estimation, road safety estimation, taxi demand prediction, and autonomous driving. Ultimately, we discuss challenges and future research directions in applying MTL in smart transportation systems.INDEX TERMS Multi-task learning (MTL), smart transportation, deep learning.

show abstract

“…With regard to the first category, visual inspections are defined in [33] as manual measurements operated by personnel walking or slowly driving along roadways. While in [46][47][48], it was argued that visual methods are slow, cumbersome, labor-intensive and time-consuming, in [33,49] it was highlighted that they are quite easy to learn, require simple equipment, and provide data of sufficient quality for most decisions and applications. Automatic surveys evolved in parallel to the development of solutions, such as remote sensing technologies [50] as well as artificial intelligence [51] and deep learning techniques [52] that allow an easier and quicker data collection, with growing efficiency in terms of effectiveness [53]; today, they are preferred to manual surveys [54].…”

Section: Introductionmentioning

confidence: 99%

Indexing the Maintenance Priority of Road Safety Barriers in Urban and Peri-Urban Contexts: Application of a Ranking Methodology in Bologna, Italy

Nalin,

Simone,

Lantieri

et al. 2023

Infrastructures

View full text Add to dashboard Cite

The need for clear and updated information is pivotal when authorities plan and perform routinary, periodic and emergency maintenance of both road network and their roadside assets, e.g., curbs, signals, and barriers. With particular regard to road barriers, the development of remote sensing technologies, such as Laser Imaging Detection and Ranging (LiDAR), has played a disruptive role in acquiring information, so the surveys today are predominantly automatic, faster, and less biased than the traditional (i.e., visual and manual) inventorying methodologies. However, even though they are accurate, these emerging procedures usually focus only on the surveyed elements and do not provide any other information about the surrounding environment or about the qualitative degradation of the elements. The primary objective of this research effort was to present a ranking methodology for enhancing road safety in urban contexts. Due to an innovative synthetic index which takes into account both the deterioration and the location of the surveyed elements, maintenance priority of road barriers was outlined in Bologna, Italy. All the collected information was georeferenced in a Geographic Information System (GIS) environment and hence plotted in thematic maps for an easier analysis. In addition, compliance to the norm was verified. The research was tested to provide public authorities with an effective tool in the evaluation of maintenance activities and road safety policies.

show abstract

Multi-Task Learning for iRAP Attribute Classification and Road Safety Assessment

Cited by 6 publications

References 16 publications

Identifying Optimal Locations for Speed Enforcement Cameras

Identifying Optimal Locations for Speed Enforcement Cameras

Survey on Multi-Task Learning in Smart Transportation

Indexing the Maintenance Priority of Road Safety Barriers in Urban and Peri-Urban Contexts: Application of a Ranking Methodology in Bologna, Italy

Contact Info

Product

Resources

About