Abstract Fast state-of-the-art multi-object-tracking (MOT) schemes, such as reported in challenges MOT16 and Mot20, perform tracking on a single sensor, often couple tracking and detection, support only one kind of object representation or don’t take varying latencies and update rates into account. We propose a fast generic MOT system for use in real world applications which is capable of tracking objects from different sensor / detector types with their respective latencies and update rates. An SORT inspired online tracking scheme is extended by time awareness using timelines as unifying concept. The system supports different object, sensor and filter and tracking types by modularizing and generalizing the online tracking scheme, while ensuring high performance using an efficient data-oriented C++-template-based implementation. Using the proposed system we achieve, with comparable evaluation metrics, framerates up to ten times higher than the fastest MOT schemes publicly listed for the axis-aligned bounding-box tracking challenges MOT17 and MOT20.
{"title":"Fast Timeline Based Multi Object Online Tracking","authors":"Martin Hünermund, Maik Groneberg, Nils Brauckmann","doi":"10.2478/ttj-2023-0007","DOIUrl":"https://doi.org/10.2478/ttj-2023-0007","url":null,"abstract":"Abstract Fast state-of-the-art multi-object-tracking (MOT) schemes, such as reported in challenges MOT16 and Mot20, perform tracking on a single sensor, often couple tracking and detection, support only one kind of object representation or don’t take varying latencies and update rates into account. We propose a fast generic MOT system for use in real world applications which is capable of tracking objects from different sensor / detector types with their respective latencies and update rates. An SORT inspired online tracking scheme is extended by time awareness using timelines as unifying concept. The system supports different object, sensor and filter and tracking types by modularizing and generalizing the online tracking scheme, while ensuring high performance using an efficient data-oriented C++-template-based implementation. Using the proposed system we achieve, with comparable evaluation metrics, framerates up to ten times higher than the fastest MOT schemes publicly listed for the axis-aligned bounding-box tracking challenges MOT17 and MOT20.","PeriodicalId":44110,"journal":{"name":"Transport and Telecommunication Journal","volume":"199 1","pages":"65 - 72"},"PeriodicalIF":1.4,"publicationDate":"2023-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"76572169","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
O. Barzilai, Havana Rika, Nadav Voloch, Maor Meir Hajaj, O. L. Steiner, N. Ahituv
Abstract Traffic lights monitoring that considers only traffic volumes is not necessarily the optimal way to time the green/red allocation in a junction. A “smart” allocation should also consider the necessities of the vehicle’s passengers and the needs of the people those passengers ought to serve. This paper deals with a “smart” junction, where several cars approach the intersection from different directions and a traffic light is set to comply to a sequence of time intervals of red and green lights in each direction. The novel approach presented here is based not only on traffic congestion parameters, but also on the social and economic characteristics of the passengers (e.g. a handicapped person, a medical doctor, an employee who is extremely required in a certain organization due to an emergency situation). This paper proposes to enhance the smart junction with a fast lane, which has a flexible entry permit based on social and economic criteria. Machine learning (specifically, Reinforcement Learning (RL)) is added to the junction’s algorithm with the aim of optimizing the social utility of the junction. For the purposes of this study, the utility of the junction is defined by the total social and economic potential benefits given a certain red/green time allocation is set. This is defined as the measure of the reward function which contains positive factors for vehicles which crossed the junction or advanced their position and a negative factor for vehicles which remains in their positions. In addition, a weight value for the vehicles with high priority is also part of the equation. A simplified version of the smart junction has been used, serving as a model for incorporating RL into the “smart’ junction with Fast Lane (FL). Specifically, the Q-Learning algorithm is used to maximize the reward function. Simulation results show that prioritizing high priority vehicles via FL is influenced by the weights and factors given to the reward components. Farther research should enhance the “Smart” junction with FL to a more complex and realistic one using a varying amount of vehicles crossing the junction.
{"title":"Using Machine Learning Techniques to Incorporate Social Priorities in Traffic Monitoring in a Junction with a Fast Lane","authors":"O. Barzilai, Havana Rika, Nadav Voloch, Maor Meir Hajaj, O. L. Steiner, N. Ahituv","doi":"10.2478/ttj-2023-0001","DOIUrl":"https://doi.org/10.2478/ttj-2023-0001","url":null,"abstract":"Abstract Traffic lights monitoring that considers only traffic volumes is not necessarily the optimal way to time the green/red allocation in a junction. A “smart” allocation should also consider the necessities of the vehicle’s passengers and the needs of the people those passengers ought to serve. This paper deals with a “smart” junction, where several cars approach the intersection from different directions and a traffic light is set to comply to a sequence of time intervals of red and green lights in each direction. The novel approach presented here is based not only on traffic congestion parameters, but also on the social and economic characteristics of the passengers (e.g. a handicapped person, a medical doctor, an employee who is extremely required in a certain organization due to an emergency situation). This paper proposes to enhance the smart junction with a fast lane, which has a flexible entry permit based on social and economic criteria. Machine learning (specifically, Reinforcement Learning (RL)) is added to the junction’s algorithm with the aim of optimizing the social utility of the junction. For the purposes of this study, the utility of the junction is defined by the total social and economic potential benefits given a certain red/green time allocation is set. This is defined as the measure of the reward function which contains positive factors for vehicles which crossed the junction or advanced their position and a negative factor for vehicles which remains in their positions. In addition, a weight value for the vehicles with high priority is also part of the equation. A simplified version of the smart junction has been used, serving as a model for incorporating RL into the “smart’ junction with Fast Lane (FL). Specifically, the Q-Learning algorithm is used to maximize the reward function. Simulation results show that prioritizing high priority vehicles via FL is influenced by the weights and factors given to the reward components. Farther research should enhance the “Smart” junction with FL to a more complex and realistic one using a varying amount of vehicles crossing the junction.","PeriodicalId":44110,"journal":{"name":"Transport and Telecommunication Journal","volume":"28 1","pages":"1 - 12"},"PeriodicalIF":1.4,"publicationDate":"2023-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87208604","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Mosleh M. Abualhaj, A. Abu-Shareha, Sumaya N. Al-Khatib
Abstract Voice over IP (VoIP) is widely utilized by organizations, schools, colleges, and so on. Nevertheless, VoIP numerous challenges that hinder its spread. One of the significant challenges is the poor exploit of the VoIP technology network bandwidth (BW), caused by the huge preamble of the VoIP packet. This paper suggests a novel methodology to manage this huge preamble overhead challenge. The proposed methodology is named runt payload VoIP packet (RPV). The core principle of the RPV methodology is to reemploy and exploit the VoIP packet preamble’s data (fields) that are superfluous by VoIP technology, especially for unicast IP voice calls. Generally, those fields will be used to convey the VoIP packet payload. Consequently, diminish or zero the length of the payload and, therefore, spare the BW. The results of the investigation into the suggested RPV methodology indicated significant enhancement in the BW exploitation of VoIP technology. For instance, the saved BW in the examined environment with the LPC codec came to up to 25.9%.
IP语音(Voice over IP, VoIP)被广泛应用于组织、学校、学院等。然而，VoIP面临着许多阻碍其传播的挑战。其中一个重要的挑战是VoIP技术的网络带宽(BW)的利用不佳，这是由于VoIP数据包的巨大序数造成的。本文提出了一种新颖的方法来管理这一巨大的序言开销挑战。所提出的方法被命名为小载荷VoIP分组(RPV)。RPV方法的核心原理是重新利用和利用VoIP技术中多余的VoIP数据包序言数据(字段)，特别是对于单播IP语音呼叫。通常，这些字段将用于传递VoIP数据包的有效载荷。因此，减少或零有效载荷的长度，从而节省BW。对建议的RPV方法的调查结果表明，VoIP技术的BW利用显著增强。例如，在使用LPC编解码器的测试环境中，保存的BW高达25.9%。
{"title":"Utilizing Voip Packet Header’s Fields to Save the Bandwidth","authors":"Mosleh M. Abualhaj, A. Abu-Shareha, Sumaya N. Al-Khatib","doi":"10.2478/ttj-2023-0004","DOIUrl":"https://doi.org/10.2478/ttj-2023-0004","url":null,"abstract":"Abstract Voice over IP (VoIP) is widely utilized by organizations, schools, colleges, and so on. Nevertheless, VoIP numerous challenges that hinder its spread. One of the significant challenges is the poor exploit of the VoIP technology network bandwidth (BW), caused by the huge preamble of the VoIP packet. This paper suggests a novel methodology to manage this huge preamble overhead challenge. The proposed methodology is named runt payload VoIP packet (RPV). The core principle of the RPV methodology is to reemploy and exploit the VoIP packet preamble’s data (fields) that are superfluous by VoIP technology, especially for unicast IP voice calls. Generally, those fields will be used to convey the VoIP packet payload. Consequently, diminish or zero the length of the payload and, therefore, spare the BW. The results of the investigation into the suggested RPV methodology indicated significant enhancement in the BW exploitation of VoIP technology. For instance, the saved BW in the examined environment with the LPC codec came to up to 25.9%.","PeriodicalId":44110,"journal":{"name":"Transport and Telecommunication Journal","volume":"38 1","pages":"33 - 42"},"PeriodicalIF":1.4,"publicationDate":"2023-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89651780","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Ken Koshy Varghese, Sajjad Mahdaviabbasabad, Guido Gentile, Mohamed Eldafrawi
Abstract Advances in machine learning technology and the availability of big data from GPS systems have led to the development of effective methods for modelling transportation demand and forecasting the future. Most previous research concentrated on demand prediction using a variety of machine learning and deep learning models that took into account spatial and temporal relationships. This paper investigates the impact of spaces and time granularity for a Spatio-temporal demand modelling framework. Using taxi demand data from New York City, our study compares the prediction performance of deep learning models such as Long Short-Term Memory (LSTM), Convolution Neural Networks (CNN) and Temporal-Guided Networks (TGNet), modelled with a grid-based tessellation strategy. The findings of this study could assist researchers in better understanding how the granularity of space and time helps deep learning models perform better for demand forecasting problems.
{"title":"Effect of Spatio-Temporal Granularity on Demand Prediction for Deep Learning Models","authors":"Ken Koshy Varghese, Sajjad Mahdaviabbasabad, Guido Gentile, Mohamed Eldafrawi","doi":"10.2478/ttj-2023-0003","DOIUrl":"https://doi.org/10.2478/ttj-2023-0003","url":null,"abstract":"Abstract Advances in machine learning technology and the availability of big data from GPS systems have led to the development of effective methods for modelling transportation demand and forecasting the future. Most previous research concentrated on demand prediction using a variety of machine learning and deep learning models that took into account spatial and temporal relationships. This paper investigates the impact of spaces and time granularity for a Spatio-temporal demand modelling framework. Using taxi demand data from New York City, our study compares the prediction performance of deep learning models such as Long Short-Term Memory (LSTM), Convolution Neural Networks (CNN) and Temporal-Guided Networks (TGNet), modelled with a grid-based tessellation strategy. The findings of this study could assist researchers in better understanding how the granularity of space and time helps deep learning models perform better for demand forecasting problems.","PeriodicalId":44110,"journal":{"name":"Transport and Telecommunication Journal","volume":"50 1","pages":"22 - 32"},"PeriodicalIF":1.4,"publicationDate":"2023-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"91326842","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Abstract Clay soils can cause significant distress in road construction due to their low strength. Stabilizing such soil improve with binder agents prior to the geotechnical works can significantly its performance and ensure safety and stability of roads while exploitation. This research envisaged the use of five different binders (lime, energy fly ash, bio fly ash, slag, cement) as an additive stabilizing agents to improve the strength parameters of soil as required in engineering industry standards. The variations of strength was assessed using measurements of P-wave velocity of the elastic waves propagating through soil specimens stabilized by different combination of binders. Measurements were performed on 28th day of soil treatment. The best effects of added binders were noted in the following combinations: cement / energy fly ash / bio fly ash (P-waves >3100 m/s), followed by combination lime / energy fly ash / GGBFS (P-waves >2800 m/s) and cement / lime / energy fly ash (P-waves >2700 m/s). Adding lime is effective due to its fixation and chemical bond with particles. The study contributes to the industrial tests on soil strength for constructing roadbed.
{"title":"Laboratory Experiments on Soil Stabilization to Enhance Strength Parameters for Road Pavement","authors":"P. Lindh, Polina Lemenkova","doi":"10.2478/ttj-2023-0008","DOIUrl":"https://doi.org/10.2478/ttj-2023-0008","url":null,"abstract":"Abstract Clay soils can cause significant distress in road construction due to their low strength. Stabilizing such soil improve with binder agents prior to the geotechnical works can significantly its performance and ensure safety and stability of roads while exploitation. This research envisaged the use of five different binders (lime, energy fly ash, bio fly ash, slag, cement) as an additive stabilizing agents to improve the strength parameters of soil as required in engineering industry standards. The variations of strength was assessed using measurements of P-wave velocity of the elastic waves propagating through soil specimens stabilized by different combination of binders. Measurements were performed on 28th day of soil treatment. The best effects of added binders were noted in the following combinations: cement / energy fly ash / bio fly ash (P-waves >3100 m/s), followed by combination lime / energy fly ash / GGBFS (P-waves >2800 m/s) and cement / lime / energy fly ash (P-waves >2700 m/s). Adding lime is effective due to its fixation and chemical bond with particles. The study contributes to the industrial tests on soil strength for constructing roadbed.","PeriodicalId":44110,"journal":{"name":"Transport and Telecommunication Journal","volume":"46 1","pages":"73 - 82"},"PeriodicalIF":1.4,"publicationDate":"2023-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"91129655","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Stefan Sass, Markus Höfer, Michael Schmidt, S. Schmidt
Abstract Automated cargo bikes are intended to complement public transportation in a sharing concept and provide an alternative transportation option for people and goods. In highly automated driving without a seated user, real-time trajectory prediction of other road users is crucial for collision avoidance with other motor vehicles or vulnerable road users (VRU). For this purpose, moving obstacles are detected by environmental sensors and classified and tracked using object detection and tracking algorithms. The current and past position data as well as environmental information are used to predict future positions. In this paper, we present several AI-based trajectory prediction models that are specifically suited for this use case. Our focus is not only on the accuracy of trajectory prediction, but additionally on a robust, real-time and practical application. We consider models that can predict the trajectories with position estimation or distributions for position estimation for each time step in the future. For this aim, we present generative network structures based on Conditional Variational Autoencoder (CVAE) in different variants. After training, the models are integrated into our production system and their computation time is determined on the hardware we use.
{"title":"Autonomous Cargo Bike Fleets – Approaches for AI-Based Trajectory Forecasts of Road Users","authors":"Stefan Sass, Markus Höfer, Michael Schmidt, S. Schmidt","doi":"10.2478/ttj-2023-0006","DOIUrl":"https://doi.org/10.2478/ttj-2023-0006","url":null,"abstract":"Abstract Automated cargo bikes are intended to complement public transportation in a sharing concept and provide an alternative transportation option for people and goods. In highly automated driving without a seated user, real-time trajectory prediction of other road users is crucial for collision avoidance with other motor vehicles or vulnerable road users (VRU). For this purpose, moving obstacles are detected by environmental sensors and classified and tracked using object detection and tracking algorithms. The current and past position data as well as environmental information are used to predict future positions. In this paper, we present several AI-based trajectory prediction models that are specifically suited for this use case. Our focus is not only on the accuracy of trajectory prediction, but additionally on a robust, real-time and practical application. We consider models that can predict the trajectories with position estimation or distributions for position estimation for each time step in the future. For this aim, we present generative network structures based on Conditional Variational Autoencoder (CVAE) in different variants. After training, the models are integrated into our production system and their computation time is determined on the hardware we use.","PeriodicalId":44110,"journal":{"name":"Transport and Telecommunication Journal","volume":"17 1","pages":"55 - 64"},"PeriodicalIF":1.4,"publicationDate":"2023-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"83230880","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Darius Bazaras, Kristina Čižiūnienė, Kristina Vaičiūtė
Abstract The continuously changing market of companies offering logistics services has challenged logistics organizations to adapt to the needs of service users and providers. Faster information processing offers new ways of communication with suppliers and optimization of distribution systems. Increasing information flows have an increasing potential to affect the management, structure, functioning and development of enterprises. The development of technological infrastructure and information technology systems by organizations allows service participants to exchange information, adjust information flows and restore relevant information through the use of technology, which poses new challenges for resource management. The level of innovation of production processes, productivity and the quality of products directly depends on the transmission of information and technology. To successfully work in the existing market, organizations must not only purchase IT systems or technologies but also constantly upgrade system applications, improve technologies or acquire brand-new IT systems. The article analyses the peculiarities of the impact of the development of technological infrastructure and its use as an instrument for the development of logistics organizations and quality assurance of logistics services. The quality research carried out allowed identifying the problems relating to technological infrastructure in logistics organizations.
{"title":"Assessment of the Interaction of the Logistics Company’s Information Technologies with the Technological Infrastructure","authors":"Darius Bazaras, Kristina Čižiūnienė, Kristina Vaičiūtė","doi":"10.2478/ttj-2023-0005","DOIUrl":"https://doi.org/10.2478/ttj-2023-0005","url":null,"abstract":"Abstract The continuously changing market of companies offering logistics services has challenged logistics organizations to adapt to the needs of service users and providers. Faster information processing offers new ways of communication with suppliers and optimization of distribution systems. Increasing information flows have an increasing potential to affect the management, structure, functioning and development of enterprises. The development of technological infrastructure and information technology systems by organizations allows service participants to exchange information, adjust information flows and restore relevant information through the use of technology, which poses new challenges for resource management. The level of innovation of production processes, productivity and the quality of products directly depends on the transmission of information and technology. To successfully work in the existing market, organizations must not only purchase IT systems or technologies but also constantly upgrade system applications, improve technologies or acquire brand-new IT systems. The article analyses the peculiarities of the impact of the development of technological infrastructure and its use as an instrument for the development of logistics organizations and quality assurance of logistics services. The quality research carried out allowed identifying the problems relating to technological infrastructure in logistics organizations.","PeriodicalId":44110,"journal":{"name":"Transport and Telecommunication Journal","volume":"106 12S1 1","pages":"43 - 54"},"PeriodicalIF":1.4,"publicationDate":"2023-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81826897","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Maik Groneberg, Olaf Poenicke, Chirag Mandal, Nils Treuheit
Abstract The paper describes a system approach to use LiDAR sensors for capturing dynamic point cloud data in industrial process environments and to interpret the captured scenes with AI based object detection. The object detection is used to distinguish between humans and other mobile objects in safety relevant workspaces. Several AI methods relevant for such application are analysed. One method is applied with annotated test data and evaluated concerning its accuracy.
{"title":"Lidar and AI Based Surveillance of Industrial Process Environments","authors":"Maik Groneberg, Olaf Poenicke, Chirag Mandal, Nils Treuheit","doi":"10.2478/ttj-2023-0002","DOIUrl":"https://doi.org/10.2478/ttj-2023-0002","url":null,"abstract":"Abstract The paper describes a system approach to use LiDAR sensors for capturing dynamic point cloud data in industrial process environments and to interpret the captured scenes with AI based object detection. The object detection is used to distinguish between humans and other mobile objects in safety relevant workspaces. Several AI methods relevant for such application are analysed. One method is applied with annotated test data and evaluated concerning its accuracy.","PeriodicalId":44110,"journal":{"name":"Transport and Telecommunication Journal","volume":"38 1","pages":"13 - 21"},"PeriodicalIF":1.4,"publicationDate":"2023-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"91090729","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Abstract Predicting the most favorable traveling routes for Vehicles plays an influential role in Intelligent Transportation Systems (ITS). Shortest Traveling Routes with high congestion grievously affect the driving comfort level of VANET users in populated cities. As a result, increase in journey time and traveling cost. Predicting the most favorable traveling routes with less congestion is imperative to minimize the driving inconveniences. A major downside of existing traveling route prediction models is to continuously learn the real-time road congestion data with static benchmarking datasets. However, learning the new information with already learned data is a cumbersome task. The main idea of this paper is to utilize incremental learning on the Hybrid Learning-based traffic Congestion and Timing Prediction (HL-CTP) to select realistic, congestion-free, and shortest traveling routes for the vehicles. The proposed HL-CTP model is decomposed into three steps: dataset construction, incremental and hybrid prediction model, and route selection. Firstly, the HL-CTP constructs a novel Traffic and Timing Dataset (TTD) using historical traffic congestion information. The incremental learning method updates the novel real-time data continuously with the TDD during prediction to optimize the performance efficiency of the hybrid prediction model closer to real-time. Secondly, the hybrid prediction model with various deep learning models performs better by taking the route prediction decision based on the best sub-predictor results. Finally, the HL-CTP selects the most favorable vehicle routes selected using traffic congestion, timing, and uncertain environmental information and enhances the comfort level of VANET users. In the simulation, the proposed HL-CTP demonstrates superior performance in terms of Mean Square Error (MSE), Mean Absolute Error (MAE), and Root Mean Square Error (RMSE).
{"title":"Dynamic Traveling Route Planning Method for Intelligent Transportation Using Incremental Learning-Based Hybrid Deep Learning Prediction Model with Fine-Tuning","authors":"Shridevi Jeevan Kamble, Manjunath R. Kounte","doi":"10.2478/ttj-2022-0024","DOIUrl":"https://doi.org/10.2478/ttj-2022-0024","url":null,"abstract":"Abstract Predicting the most favorable traveling routes for Vehicles plays an influential role in Intelligent Transportation Systems (ITS). Shortest Traveling Routes with high congestion grievously affect the driving comfort level of VANET users in populated cities. As a result, increase in journey time and traveling cost. Predicting the most favorable traveling routes with less congestion is imperative to minimize the driving inconveniences. A major downside of existing traveling route prediction models is to continuously learn the real-time road congestion data with static benchmarking datasets. However, learning the new information with already learned data is a cumbersome task. The main idea of this paper is to utilize incremental learning on the Hybrid Learning-based traffic Congestion and Timing Prediction (HL-CTP) to select realistic, congestion-free, and shortest traveling routes for the vehicles. The proposed HL-CTP model is decomposed into three steps: dataset construction, incremental and hybrid prediction model, and route selection. Firstly, the HL-CTP constructs a novel Traffic and Timing Dataset (TTD) using historical traffic congestion information. The incremental learning method updates the novel real-time data continuously with the TDD during prediction to optimize the performance efficiency of the hybrid prediction model closer to real-time. Secondly, the hybrid prediction model with various deep learning models performs better by taking the route prediction decision based on the best sub-predictor results. Finally, the HL-CTP selects the most favorable vehicle routes selected using traffic congestion, timing, and uncertain environmental information and enhances the comfort level of VANET users. In the simulation, the proposed HL-CTP demonstrates superior performance in terms of Mean Square Error (MSE), Mean Absolute Error (MAE), and Root Mean Square Error (RMSE).","PeriodicalId":44110,"journal":{"name":"Transport and Telecommunication Journal","volume":"40 1","pages":"293 - 310"},"PeriodicalIF":1.4,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"90468156","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Abstract The usage of mobile phones is nowadays reaching full penetration rate in most countries. Smartphones are a valuable source for urban planners to understand and investigate passengers’ behavior and recognize travel patterns more precisely. Different investigations tried to automatically extract transit mode from sensors embedded in the phones such as GPS, accelerometer, and gyroscope. This allows to reduce the resources used in travel diary surveys, which are time-consuming and costly. However, figuring out which mode of transportation individuals use is still challenging. The main limitations include GPS, and mobile sensor data collection, and data labeling errors. First, this paper aims at solving a transport mode classification problem including (still, walking, car, bus, and metro) and then as a first investigation, presents a new algorithm to compute waiting time and access time to public transport stops based on a random forest model. Several public transport trips with different users were saved in Rome to test our access trip phase recognition algorithm. We also used Convolutional Neural Network as a deep learning algorithm to automatically extract features from one sensor (linear accelerometer), obtaining a model that performs well in predicting five modes of transport with the highest accuracy of 0.81%.
{"title":"Smartphone-Based Recognition of Access Trip Phase to Public Transport Stops Via Machine Learning Models","authors":"Seyed Hassan Hosseini, Guido Gentile","doi":"10.2478/ttj-2022-0022","DOIUrl":"https://doi.org/10.2478/ttj-2022-0022","url":null,"abstract":"Abstract The usage of mobile phones is nowadays reaching full penetration rate in most countries. Smartphones are a valuable source for urban planners to understand and investigate passengers’ behavior and recognize travel patterns more precisely. Different investigations tried to automatically extract transit mode from sensors embedded in the phones such as GPS, accelerometer, and gyroscope. This allows to reduce the resources used in travel diary surveys, which are time-consuming and costly. However, figuring out which mode of transportation individuals use is still challenging. The main limitations include GPS, and mobile sensor data collection, and data labeling errors. First, this paper aims at solving a transport mode classification problem including (still, walking, car, bus, and metro) and then as a first investigation, presents a new algorithm to compute waiting time and access time to public transport stops based on a random forest model. Several public transport trips with different users were saved in Rome to test our access trip phase recognition algorithm. We also used Convolutional Neural Network as a deep learning algorithm to automatically extract features from one sensor (linear accelerometer), obtaining a model that performs well in predicting five modes of transport with the highest accuracy of 0.81%.","PeriodicalId":44110,"journal":{"name":"Transport and Telecommunication Journal","volume":"4 1","pages":"273 - 283"},"PeriodicalIF":1.4,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"73123726","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：