univerge site banner
Original Article | Open Access | https://doi.org/10.34104/ajeit.024.01040121 | doi: Aust. J. Eng. Innov. Technol., 2024; 6(5), 104-121

Real-time Vehicle Monitoring and Route Prediction using Geolocation through LSTM

Md. Al-Amin Hossain Bijoy* Mail Img ,
Pallab Sarker Mail Img ,
Nazmus Sakib Siam Mail Img ,
Md. Nazmul Ahsan Mail Img ,
Munna Hossain Mail Img ,
Md. Shahiduzzaman Mail Img

Abstract

Vehicle travel, such as by bus, automobile, etc., is a common way to use public transportation since its convenient, economical, easily accessible, and eco-friendly. However, there are moments when its challenging to keep an eye on everyone while traveling, avoid traffic bottlenecks, and provide enough service, which makes the passengers bored. In this paper, we offer a real-time geolocation Android application that gathers real-time geolocation data (longitude and latitude), date, time, and speed, and integrates it with Google Firebase to assist predictive analytics for the operational efficiency of a vehicle firm. After being exported from the database as a JSON file, the data must be transformed into a CSV file. The company will be able to anticipate a vehicles future location on a predetermined route based on date, time, and speed by using this database to train an LSTM (Long short-term memory) model, a sort of sequential neural network. Determining the vehicles expected arrival time, allocating passengers, and predicting traffic conditions will therefore be useful in preventing congestion. Utilizing this technology allows a car company to improve scheduling, boost monitoring, and provide better overall service.

INTRODUCTION

Public transportation by vehicles such as buses and cars plays a pivotal role in modern urban life, offering affordability, accessibility, convenience, and environmental sustainability. However, as these modes of transportation continue to gain popularity, managing passengers effectively, mitigating traffic congestion, and ensuring a seamless travel experience has become increasingly challenging. This project seeks to address these challenges through the development and implementation of a real-time geolocation Android application, tightly integrated with Google Firebase, and using the location data in a LSTM model to enable predictive analytics for the operational efficiency of a vehicle firm.

The increasing use of location-aware sensors such as GPS, WiFi, and Bluetooth devices has produced large amounts of urban trajectory data, which capture the spatio-temporal footprint of the movement of individual people and vehicles travelling around cities. Trajectory data contain rich information on travel behaviours and network-wide traffic dynamics and offer new opportunities to discover and predict individual and collective mobility patterns of users in urban road (Sun, Jie and Jiwon kim, 2021). The use of this massive GPS data produced by location acquisition technologies is an important part of smart city and intelligent transportation system, and produced many opportunities for researchers in different application domains, such as route prediction by time, traffic congestion estimation (Hung qiu et al., 2020), travel route planning, behavior analysis, transportation monitoring, point of interest recommendation, inference of taxi status, identifying travel trips and activities or destination prediction, etc. (Besse P.C et al., 2024). 

In Dhaka, the air pollution that has been a serious concern for the past 20 years is mostly caused by rapid urbanization, inadequate infrastructure, vehicle emissions, and traffic congestion (Ahmed et al., 2024). Transportation companies, particularly those operating a fleet of buses, often grapple with the task of ensuring a smooth and efficient passenger experience. Passengers desire accurate information regarding bus arrival times, minimized wait times, and a reliable estimation of their journeys duration. In addition, companies need to optimize their scheduling, adapt to dynamic traffic conditions, and allocate passengers efficiently to ensure the best use of resources and maintain a high level of service quality. To meet these multifaceted demands, our project introduces a comprehensive solution that relies on real-time geolocation data. This data, comprising longitude, latitude, speed, and time information, is collected through a purpose-built Android application. The information is then securely stored in Google Firebase, a powerful cloud-based database service, allowing for the retrieval and analysis of real-time location and traffic data (Sami et al., 2021).

One of the critical objectives of this project is to export the collected data from the database into a structured CSV file. This CSV file will serve as the foundation for training a LSTM model. This model can forecast a cars future location along a predetermined path, estimate the expected arrival time at specific stops, and provide insights into traffic conditions to help the company avoid congestion and delays. The implementation of this technology has the potential to revolutionize the way vehicle companies manage their operations. By harnessing predictive analytics and real-time data, these companies can enhance their monitoring capabilities, offer passengers a more seamless travel experience, and optimize scheduling to maximize operational efficiency. The result is a win-win situation for both the company and the passengers, fostering a more sustainable and passenger-centric public transportation system for the future. Travel by public vehicles, such as buses and cars, represents a cost-effective, accessible, and eco-friendly mode of transportation that is essential for urban and suburban communities. However, despite its popularity, several challenges persist in managing passenger experiences, mitigating traffic-related issues, and ensuring high-quality service delivery. Passengers often face uncertainties related to vehicle arrival times, endure frustrating traffic jams, and may experience a lack of engaging services during their journey. These problems are detrimental to both the passengers satisfaction and the operational efficiency of vehicle companies.

The primary issue that this project seeks to address is the lack of a comprehensive solution for managing and improving the public transportation experience. Current transportation systems often rely on fixed schedules that are subject to disruptions caused by unpredictable traffic conditions, which can result in long waiting times and inefficiencies in resource allocation. Additionally, the absence of real-time passenger engagement mechanisms can lead to a disengaged and less satisfied rider community. Moreover, the absence of tools to predict future vehicle locations, estimated arrival times, and traffic conditions further exacerbates these challenges. This lack of predictive capabilities hinders a vehicle companys ability to optimize its scheduling, allocate passengers efficiently, and provide accurate information to travelers.

The root challenges are:

  1. Predicting future vehicle locations
  2. Managing passengers
  3. Avoiding traffic jams

The problem addressed by this project is the need for a data-driven, predictive solution to enhance the operational efficiency and passenger experience of public vehicle transportation. The absence of such a solution results in operational challenges, passenger dissatisfaction, and suboptimal resource allocation, all of which can be alleviated through the proposed system. Public transportation is a cornerstone of urban mobility, offering a cost-effective, accessible, and environmentally friendly mode of commuting through vehicles like buses and cars. However, despite the many advantages of public transportation, numerous challenges persist in providing a seamless and efficient experience for passengers. These challenges include difficulties in managing passenger allocation, navigating traffic congestion, maintaining service quality, and addressing passenger dissatisfaction resulting from unpredictable delays. In todays fast-paced world, where time is of the essence, the need to address these challenges has never been more critical.

Managing Passenger Allocation 

Vehicle companies often struggle to efficiently allocate passengers to vehicles and routes, leading to overcrowding, uneven passenger distribution, and extended waiting times. These issues can result in decreased passenger satisfaction and operational inefficiencies.

Avoiding Traffic Jams 

Traffic congestion remains a significant issue in urban areas, causing delays and unpredictability in transportation services. Vehicle companies require strategies to mitigate traffic-related disruptions and offer reliable journey times.

Providing Quality Service 

Ensuring that passengers receive timely, quality service is a paramount concern. Passengers expect accurate information about vehicle arrival times and transparent communication regarding their journey.

Passenger Engagement 

Boredom during the journey is a common issue, leading to decreased passenger satisfaction. Engaging passengers during their travel experience can help alleviate this concern and enhance overall customer experience. The problem of managing passengers, avoiding traffic jams, and providing proper service is exacerbated by the fact that vehicle companies often lack real-time data on the location of their vehicles and the traffic conditions on their routes. This makes it difficult to make informed decisions about how to allocate resources and optimize operations. In addition, vehicle companies often rely on manual processes to manage their operations. This can lead to errors and inefficiencies. For example, a bus driver may not be able to accurately estimate their arrival time at a particular stop due to traffic conditions. This can lead to passengers waiting longer than expected and missing appointments or work.

There is a need for a solution that can help vehicle companies overcome these challenges and improve their operational efficiency. A real-time geolocation Android application, integrated with Google Firebase, can provide vehicle companies with the data and tools they need to manage their operations more effectively, a LSTM model to predict future  obstacles and estimate time, can solve the above issues and this type of system have been proposed by us. Such a system could be used to:

  1. Track the location of vehicles in real time.
  2. Forecast traffic conditions.
  3. Optimize vehicle routes.
  4. Provide passengers with accurate arrival time information.
  5. Manage passenger allocation.
  6. Improve overall service delivery.

By harnessing the power of real-time geolocation data and LSTM model, vehicle companies can improve their operational efficiency, provide better service to passengers, and reduce costs. The goal of this project is to create an Android application for real-time geolocation that is coupled with Google Firebase to enable predictive analytics for the operational efficiency of a car company. Vehicle speed, time, and geolocation information (longitude and latitude) will be gathered in real time by the proposed system. In order to train an LSTM model, this data will be exported as a CSV file and kept in a Google Firebase database. The LSTM model will be trained to forecast traffic conditions to prevent congestion, manage passenger distribution, and predict a vehicles future location on a predetermined route. We will also be able to obtain the vehicles predicted arrival time. A car firm can improve total service delivery, streamline scheduling, and strengthen monitoring capabilities using this technology. The particular goals of the research are as follows:

Develop a Real-Time Geolocation Android Application (RTMS) 

Create a robust Android application (RTMS) that can collect and transmit real-time geolocation data, including longitude, latitude, speed, and time, to effectively track vehicle movements. This task is mainly used for collecting the location data.

Integrate Google Firebase for Data Storage and Collection

Establish a secure and efficient integration with Google Firebase to store and manage the collected geolocation data, ensuring reliability, accessibility, and data integrity.

Data Export to CSV File 

Implement a data export mechanism to transform the information stored in Google Firebase into a structured CSV file format, ensuring that data remains organized and usable for further analysis.

LSTM Model Development 

Design and develop a model capable of accurately predicting a vehicles future location on a predefined route, estimating the expected arrival time, and forecasting traffic conditions to address operational challenges. The additional research objectives include:

  1. Passenger Allocation and Service Management
  2. Traffic Condition Forecasting
  3. Enhanced Passenger Engagement
  4. Operational Efficiency Enhancement
  5. Transport Companies Real-time Monitoring Enhancement

The successful achievement of these research objectives will empower vehicle companies to transform their operations, offer a more efficient and enjoyable experience to passengers, and contribute to the creation of a sustainable and passenger-centric public transportation system. A comprehensive solution that relies on real-time geolocation data and LSTM model to enhance the operational efficiency of vehicle companies. The potential benefits of the proposed application for both vehicle companies and passengers, including improved monitoring capabilities, optimized scheduling, enhanced service delivery, reduced traffic congestion, and increased environmental sustainability.

Review of Literature

Public transportation, including vehicle options such as buses and cars, remains a widely adopted and accessible mode of travel due to its cost-effectiveness, convenience, availability, and environmental friendliness. Nonetheless, challenges persist in ensuring a seamless passenger experience, including efficient passenger management, congestion avoidance, service quality, and passenger engagement. Traditional solutions have often fallen short in addressing these multifaceted issues. There are several existing solutions that vehicle companies can use to address these challenges. For example, some vehicle companies use GPS tracking systems to track the location of their vehicles. This information can be used to improve passenger information and to identify and avoid traffic jams. Additionally, some vehicle companies use mobile apps to allow passengers to book tickets, track the location of their vehicle, and receive estimated arrival times. However, existing solutions have several limitations. For example, it cannot predict a vehicles future location, estimate arrival times, optimize passenger allocation, and forecast traffic conditions. In response, this project introduces an innovative approach, leveraging a real-time geolocation Android application integrated with Google Firebase. This technology surpasses existing solutions by enabling the collection of real-time geolocation data, including longitude, latitude, speed, and time, and exporting it to a CSV file for LSTM model training. This novel solution empowers vehicle companies to predict a vehicles future location, estimate arrival times, optimize passenger allocation, and forecast traffic conditions, ultimately enhancing monitoring capabilities, scheduling optimization, and service delivery, thus addressing the limitations of prior methods.

Though vehicle monitoring and collection of real-time location data are common nowadays for researchers. There is a significant amount of research out there being done on different aspects of vehicle monitoring and collection of real-time location data and shows their outcomes. Traffic jam is a common problem in lot of country. Nowadays there is a significant amount of research out there done. They (Skhosana et al., 2020) introduces an intelligent real-time transport information system addressing the issue of unreliable public transport in developing countries. The system utilizes Firebase for Backend-as-a-Service, with three subsystems for commuters, bus drivers, and managers. A neural network predicts ridership numbers per route, integrated into a web app for managers. An Android app collects ridership data for network input. Evaluation with real-world data since 2001 demonstrates the systems effectiveness, reducing mileage, fuel costs, and increasing operator profit and rider satisfaction. In this paper (D. Dhinakaran et al., 2022) presents SALT, a Secure Android Location Tracker system designed to overcome limitations in existing tracking solutions. SALT operates without individual consent, gathering location data every 10 seconds and transmitting it to the secure Firebase cloud. This ensures tracking continuity even in scenarios where GPS is off or the devices SIM card is swapped. Using JSON format for data transit enhances security. Future work aims to integrate the system into the Android OS build to prevent erasure during factory resets. Authors (Skhosna et al., 2021) addresses challenges in public transportation, focusing on buses. It introduces a cloud-based system with mobile and web application interfaces. The mobile app provides real-time bus information to commuters and enables tracking for drivers, collecting ridership data. The web app acts as a dashboard for operators to gain insights. Developed on the Firebase BaaS platform, the system integrates a machine learning model predicting daily ridership. The solution is holistic, scalable, cost-efficient, and leveraging contemporary technologies for improved public transport management.

The authors of these papers (Skhosana, Menzi, 1970) explore leveraging Big Data and the Internet of Things in the public transportation sector of developing countries. Focusing on buses, the paper introduces a cloud-based system with mobile and web applications. The mobile app offers real-time bus information for commuters and enables tracking for drivers, collecting ridership data. The web app serves as a dashboard for operators to extract insights. Developed on the Firebase BaaS platform, the system integrates a machine learning model predicting daily ridership, providing a scalable, cost-efficient, and technologically advanced solution for public transport management. Shridevi Jeevan Kamble et al. (Kamble et al., 2020) addresses urban traffic congestion by employing a machine learning (ML) approach that utilizes intelligent internet of vehicles. It focuses on predicting congestion based on parameters like hard delay constraints and GPS vehicle trajectory speed. The Gaussian process in ML is used for traffic speed prediction, leveraging three datasets: training set, prediction set, and road sector data frame. The study identifies and evaluates three distinct time slots for monitoring vehicle traffic congestion, assessing average speeds during these periods using the dataset.

Authors in this paper (Krause et al., 2019) addresses the issue of traffic congestion and wasted time in U.S. commuting. It introduces short-term destination prediction using a database of GPS driving traces and integrates point of interest (POI)/land use data. The dataset includes over 20,000 trips, and machine learning methods are employed to calculate trip purposes. A novel prediction model structure is developed, incorporating trip purpose information, resulting in improved accuracy and speed compared to traditional models, offering a deeper understanding of travel motivations for effective short-term destination prediction. Authors of this paper (Julio et al., 2016) explores the use of machine learning algorithms (ANN, SVM, Bayes Networks) to predict bus travel speeds using real-time traffic data. The proposed methods outperform naive prediction models and demonstrate the effectiveness of incorporating real-time data for accurate bus travel speed prediction.

This paper (Bansal et al., 2020) proposes a machine learning-based pothole detection system called DeepBus, which uses IoT sensors to detect potholes in real-time and provide a centralized map for users and authorities. In this paper (Souza et al., 2018) introduces Asfault, a cost-effective system utilizing smartphone sensors and machine learning to assess road pavement conditions in real-time. Achieving over 90% classification accuracy, it offers practical applications for road maintenance and route planning. They (Jisha et al., 2017) introduce an IoT-based school bus monitoring system using RFID/GPS/GSM/GPRS technologies. It includes a prediction algorithm for bus arrival times, enhancing student safety, and enabling parents to monitor routes via an Android application. The authors (Meng et al., 2020) proposes a novel Long Short-Term Memory with Dynamic Time Warping (D-LSTM) model for short-term traffic speed prediction using GPS positioning data. DTW is introduced to fine-tune the time feature, aligning the current data distribution with historical data. The attention mechanism focuses on relevant input data, improving prediction accuracy. Experiments demonstrate that D-LSTM outperforms existing models, especially for weekend traffic prediction.

This paper (Zhao et al., 2019) presents a novel method for predicting traffic speed on expressways using BeiDou satellite navigation system (BDS) data. The method addresses the challenges of abnormal data, missing data, and sparse data by employing data screening, filling, and a deep learning model based on Long Short-Term Memory (LSTM). Additionally, the proposed method incorporates a three-regime algorithm to handle non-recurrent traffic congestion. Experimental results demonstrate that the proposed method outperforms Support Vector Regression (SVR) in terms of prediction accuracy and robustness. The authors (Gui, Zhipeng et al., 2021) describes real-time destination prediction for individual drivers is crucial for various location-based services. However, existing methods often overlook the impact of spatial factors such as urban functionalities and trajectory points. To address this gap, the authors propose LSI-LSTM, an attention-aware LSTM model that incorporates location semantics and location importance to improve prediction accuracy. Experiments demonstrate the effectiveness of LSI-LSTM compared to baseline methods. In this paper (Besse et al., 2018) proposes a novel method for predicting the final destination of vehicle trips based on their initial partial trajectories. The method involves clustering trajectories to capture user behavior and modeling traffic flow patterns using a mixture of 2-D Gaussian distributions. This results in a density-based clustering of locations, which generates a data-driven grid of similar points within each pattern. The model predicts the final destination of a new trajectory based on its first locations using a two-step procedure: assigning the new trajectory to the clusters it most likely belongs to and using characteristics from trajectories within those clusters to predict the final destination. Experimental results demonstrate the effectiveness of the method on two different datasets, indicating its ability to adapt to different subsets.

This paper (Seitbekova et al., 2020) present a novel method for predicting bus arrival times using historical bus GPS data, real-time traffic information, and bus dwell time at bus stops. Their approach divides bus arrival time into bus dwelling time and bus travel time and predicts each separately. For bus travel time, they use a clustering approach followed by an LSTM neural network. For bus dwelling time, they use historical dwelling time and location analysis. The method is evaluated on GPS data from 1200 buses and shows small mean absolute error for buses that are not far from the departure station. The authors conclude that their method can be used to provide additional information for bus passengers and to estimate possible travel time in bus journeys. This paper (Mikhailov et al., 2021) discusses the impact of COVID-19 on the tourism industry and proposes an approach for car tourist trajectory prediction using a bidirectional LSTM neural network model. The proposed system utilizes sensor data from the tourists smartphone and information processing capabilities of surrounding devices to analyze tourist behavior during car-based attraction-visiting trips. The obtained results can be used for tourist trip behavior analysis. In this paper (Choudhury et al., 2019) explores the challenging problem of predicting waiting times for fleet trucks on long-haul journeys in the logistics industry, emphasizing the critical connection between unexpected delays and detention costs. The paper introduces a novel approach termed "vehicle speed trend analysis and hours of service forecasting" using LSTM networks. By converting real-time location updates into speed sequences and employing historical data, the proposed method demonstrates effective forecasting of speed trends for the majority of a vehicles journey and successfully learns the trend in hours of service for fleet trucks traveling long distances with a mean absolute error of less than an hour. They (Liu et al., 2020) explores the challenges in bus arrival prediction due to the complex and dynamic nature of urban traffic conditions. The authors address the limitations of traditional methods using historical data and real-time speed for short and long-term predictions, respectively. To overcome these challenges, they propose a comprehensive approach using a combination of long short-term memory (LSTM) and Artificial Neural Networks (ANN) based on spatial-temporal feature vectors, demonstrating high accuracy in bus arrival predictions through experiments with an entity dataset.

This paper (Zhao et al., 2018) explores the challenges in bus arrival prediction due to the complex and dynamic nature of urban traffic conditions. The authors address the limitations of traditional methods using historical data and real-time speed for short and long-term predictions, respectively. To overcome these challenges, they propose a comprehensive approach using a combination of long short-term memory (LSTM) and Artificial Neural Networks (ANN) based on spatial-temporal feature vectors, demonstrating high accuracy in bus arrival predictions through experiments with an entity dataset.

Problem Analysis

Although researchers are trying their best to propose a model that are quite similar to us relating to public transport by vehicle but there are several differences. The fourth, sixth and seventh paper introduces an intelligent real-time transport information system for public transportation in developing countries, utilizing Firebase, neural networks, and mobile apps. A cloud-based system with mobile and web applications that focuses on real-time bus information, tracking, ridership data, and machine learning predictions. Leveraging Big Data and the Internet of Things in the public transportation sector, focusing on cloud-based systems, mobile apps, web apps, real-time bus information, tracking, ridership data, and machine learning predictions. Their projects are different in their approaches and specific applications. These are not about monitoring a vehicle. Paper thirteen specifically addresses school bus monitoring, emphasizing student safety and parent monitoring, which is different from the broader public transportation context of the proposed project. And the paper- fifth, eighth, ninth, eleventh, twelfth, thirteenth primarily focus on addressing traffic congestion, road conditions, or pothole detection. They may employ machine learning for predictive purposes but differ in their core objectives compared to the proposed project. So we can see that there are solutions but these are not fulfill as we mentioned in our project rather there are some extra features that passengers dont need to think. Similar to Xianwei Mengs paper, our project utilizes LSTM to capture temporal patterns in vehicle movements. However, our focus on real-time prediction differentiates it from Mengs work, which focuses on short-term traffic speed prediction using historical data. Zhaos paper addresses the challenges of abnormal, missing, and sparse data in traffic speed prediction using BeiDou satellite navigation system (BDS) data. Our project shares this focus on handling real-world data imperfections, but it extends beyond traffic speed prediction to encompass real-time vehicle monitoring and route prediction. Guis paper highlights the importance of spatial factors in real-time destination prediction for individual drivers. Our project aligns with this emphasis on spatial context by utilizing geolocation data to predict vehicle routes. However, our project expands upon this concept by providing real-time updates and predictions. Philippe Cs paper proposes a method for predicting final destinations of vehicle trips based on initial partial trajectories. Our project shares this goal of destination prediction, but it distinguishes itself by utilizing LSTM to capture temporal patterns in vehicle movements, enabling more accurate real-time predictions. Yerkezhan Seitbe-kovas paper addresses the problem of bus arrival time prediction using historical bus GPS data, real-time traffic information, and bus dwell time at bus stops. While our project focuses on vehicle monitoring and route prediction rather than arrival time estimation, it shares a common goal of providing accurate real-time information to users. Sergei Mikhailov proposes a method for predicting car tourist trajectories using a bidirectional LSTM neural network model. Our project shares this focus on car-based route prediction but differentiates itself by providing real-time monitoring and prediction capabilities. Arunabha Choudhury explores the challenging problem of predicting waiting times for fleet trucks on long-haul journeys in the logistics industry. While our project focuses on real-time vehicle monitoring and route prediction for general vehicles, it shares a common interest in improving logistics efficiency. Hongjie Liu addresses the challenges in bus arrival prediction due to complex and dynamic urban traffic conditions. Our project shares this focus on handling real-world traffic complexities, but it extends beyond bus arrival prediction to encompass real-time vehicle monitoring and route prediction for general vehicles. Jiandong Zhaos paper also explores the challenges in bus arrival prediction due to complex and dynamic urban traffic conditions. While our project focuses on real-time vehicle monitoring and route prediction for general vehicles, it shares a common interest in improving traffic management and user experience. 

Our project stands out by combining real-time vehicle monitoring, route prediction, and LSTM-based modeling to provide valuable insights into vehicle movements and potential route changes. It addresses the limitations of existing methods by focusing on real-time prediction and handling real-world data imperfections. So the existing systems cannot solve this kind of problem properly and completely because of their limitations and thus a solution is highly essential.

METHODOLOGY

In response to the pressing challenges faced by public transportation, particularly in the realm of passenger management, traffic navigation, and service quality, this project introduces a pioneering solution. We propose the development of a real-time geolocation Android application (RTMS), seamlessly integrated with Google Firebase, to harness the power of predictive analytics for enhancing a vehicle companys operational efficiency. By continuously collecting real-time geolocation data, including longitude, latitude, speed, and time, this innovative application lays the foundation for improved decision-making within the company. The data is skillfully exported to a structured CSV file, serving as the basis for training a LSTM model. This model, the cornerstone of our proposed system, will enable the company to predict a vehicles future location on a predefined route, estimate arrival times with precision, allocate passengers efficiently and forecast traffic conditions to avert congestion. By adopting this technology, a vehicle company can bolster its monitoring capabilities, fine-tune scheduling, and elevate the quality of service delivery, ushering in a new era of operational excellence and passenger satisfaction.

Feasibility Analysis

The proposed real-time geolocation Android application (RTMS), coupled with Google Firebase, to facilitate predictive analytics for a car companys operational efficiency is feasible from a technical, economic, and operational standpoint.

Technical Feasibility

The proposed system is technically feasible because all of the required technologies are available and mature. The system will use Android for the mobile app, Google Firebase for the backend database and cloud computing platform, and a LSTM model to predict vehicles future route according to the date, arrival times and traffic conditions.

Economic Feasibility

The proposed system is economically feasible because it is relatively inexpensive to develop and deploy. The cost of developing the Android app and the Google Firebase backend will be relatively low, and there are several open-source models that can be used to predict vehicles future route according to the date, arrival times and traffic conditions.

Operational Feasibility

The proposed system is operationally feasible because it is easy to use and maintain. The Android app will be easy for passengers and vehicle operators to use, and the Google Firebase backend will be easy for vehicle companies to manage. Overall, the proposed system is feasible from a technical, economic, and operational standpoint. The system has the potential to offer several benefits to vehicle companies and passengers alike. In addition to the technical, economic, and operational feasibility considerations, there are a few other factors that need to be considered when evaluating the feasibility of the proposed system.

Data privacy and security

The proposed system will collect a significant amount of data, including the location of vehicles and passengers. It is important to ensure that this data is collected and stored in a secure manner.

Acceptance by vehicle companies and passengers 

To be successful, the proposed system needs to be adopted by vehicle companies and passengers alike. It is important to conduct market research and user testing to ensure that the system meets the needs of these users.

Regulatory compliance 

The proposed system needs to comply with all applicable laws and regulations. It is important to conduct legal research to ensure that the system is compliant. The proposed real-time geolocation Android application that integrates with Google Firebase, to provide predictive analytics for a vehicle companys operational efficiency is feasible from a technical, economic, and operational standpoint. However, it is important to consider the factors of data privacy and security, acceptance by vehicle companies and passengers, and regulatory compliance when evaluating the feasibility of the system. The feasibility analysis must comprehensively address these dimensions to ascertain the viability and potential success of the project. A favorable feasibility analysis suggests that the project has a high likelihood of delivering its intended benefits in terms of enhanced operational efficiency and improved service delivery for the vehicle company.

Requirement Analysis

This project necessitates a comprehensive analysis of various requirements to ensure the successful implementation of the proposed system, aiming to enhance a vehicle companys operational efficiency.

Hardware and Software Requirements

  1. Android mobile devices equipped with GPS capabilities for data collection.
  2. Adequate computing resources for the develop-ment and training of the LSTM model.
  3. Software tools for Android application development, database integration, and model training (e.g., Android Studio, Google Firebase, Google Colab, Python libraries).

Data Collection and Integration Requirements

  1. Reliable internet connectivity for real-time data transmission.
  2. Integration with Google Firebase for secure storage, retrieval, and management of geolocation data.
  3. Exporting JSON file after collecting longitude, latitude, speed, date and time data.

Data Processing and LSTM Model Requirements

  1. Robust data processing capabilities to export data from the database to a CSV file.
  2. Access to different libraries and frameworks for model development, training, and deployment. Using LSTM layers for the model.

Sufficient data for model training, including historical geolocation data and related parameters.

Model Development and Prediction Requirements

  1. A well-structured and scalable LSTM model capable of predicting vehicle locations and estimating arrival times.
  2. Real-time data input for model predictions, with a focus on accuracy and loss.
  3. Regular model updates to adapt to changing conditions 

Data Security and Privacy Requirements

  1. Implementation of robust data security mea-sures to protect passenger and operational data.
  2. Compliance with data privacy regulations and ethical considerations, including user consent and data anonymization.

User Acceptance Testing and Feedback Requirements

  1. Gathering feedback from passengers and vehicle company staff to validate system usability, accuracy, and user satisfaction.
  2. Iterative testing and improvements based on user feedback to enhance the systems performance and user experience.

Operational Integration and Training Requirements

  1. Training of staff and operators in efficient system utilization.
  2. Seamless integration into the existing vehicle company operations, with minimal disruption.

Legal and Regulatory Requirements

  1. Compliance with transportation and data privacy regulations, including data retention and sharing policies.
  2. Ethical considerations for data collection, storage, and usage.

By meticulously addressing these requirements, the capstone project aims to develop and implement a comprehensive solution that enhances the operational efficiency of vehicle companies, thereby improving service delivery and passenger experiences. The following research methodology will be used to develop and evaluate the proposed real-time geo-location Android application (RTMS), integrated with Google Firebase, to facilitate predictive analytics for a vehicle companys operational efficiency:

Fig. 1: Android Application Interface (Login, Signup and Map).

Android App Development

First, we must develop a real-time geolocation Android application capable of continuously collecting data using Mapbox (Map), including longitude, latitude, speed, and time. Then integrating the application with Google Firebase to securely store and manage the collected data.

Data Collection

The second step will be to collect data on the location, speed, and time of vehicles. This data can be collected using a variety of methods, but we are using smartphone apps and collecting the data- longitude, latitude, speed, date and time and stored this data into google firebase.

Data Preprocessing

Once the data has been collected, it will need to be preprocessed. First, we have to export the JSON file and convert it into CSV file so that we can get the data in a suitable way for using in our LSTM model.

LSTM Model Development and Training

The next step will be to train a LSTM model to predict a vehicles future location in a specific route.

Fig. 2: LSTM Model
The core of this real-time vehicle monitoring and route prediction system is recurrent neural network (RNN) architecture, specifically the Long Short-Term Memory (LSTM) network. LSTMs are well-suited for modeling sequential data, such as geolocation data, as they can effectively capture long-term dependencies within the data. The proposed LSTM model comprises four layers:

Input Layer 
This layer receives the input sequence of geolocation data. The input shape of the layer is defined as (xtrain.shape, 1), indicating that the sequence consists of xtrain.shape [1] features (longitude, latitude, etc.) and each feature is represented as a single value.

First LSTM Layer (128 Units) 
This layer processes the input sequence and extracts temporal patterns using 128 LSTM units. The return_sequences=True parameter ensures that the layer outputs a sequence of vectors, allowing the model to capture long-term dependencies.

Second LSTM Layer (64 Units)
This layer further refines the temporal patterns extracted by the first LSTM layer using 64 LSTM units. The return_sequences=False parameter indicates that the layer outputs a single vector, summarizing the temporal information.

Dense Layer (25 Units)
This fully connected layer transforms the output of the second LSTM layer into a 25 dimensional vector.

Output Layer (2 Units) 
This final fully connected layer produces the predicted longitude and latitude values. The model is compiled using the Adam optimizer and the mean squared error (MSE) loss function. MSE is a suitable choice for this task as it measures the average squared difference between the predicted and actual geolocation values.

System Testing and Validation
Conduct rigorous testing to validate the accuracy, reliability, and speed of the systems predictions, passenger allocation, and traffic forecasts. Gather feedback from users and conduct iterative testing to refine the systems performance and usability.

System Evaluation
Once all the previous tasks done, it will be evaluated to assess its effectiveness in improving a vehicle companys operational efficiency. The evaluation will be conducted in two phases:

Local-Area Evaluation
The first phase of the evaluation will be conducted in a local area using several types of vehicles. This phase will involve testing the Android app with a simulated dataset of vehicle data.

Real-World Evaluation 
The second phase of the evaluation will be conducted in a real-world setting. This phase will involve testing the Android app with a fleet of real vehicles. The results of the evaluation will be used to improve the Android app and the LSTM model.
Fig. 3: Methodology.
A real-time geolocation Android application integrated with Google Firebase will be developed and evaluated using the suggested research approach in order to enable predictive analytics for the operational efficiency of a car firm. The research methodology is designed to ensure that the system is developed and evaluated in a rigorous and scientific manner.
Design and Implementation
The project has been meticulously designed to enhance the operational efficiency and passenger experience within the public transportation sector. The projects design was composed of several key components, as outlined below:

Developing the Android Application
For IDE we have used the Android Studio. It is the official integrated development environment for Android app development. It provides a rich set of tools for designing, coding, and debugging Android applications. Android Studio simplifies the app development process and offers features like code completion, visual layout editors, and a robust emulator for testing apps. For programming language we have used Kotlin. It is a modern, statically-typed pro-gramming language developed by JetBrains. It is fully interoperable with Java and has gained popularity as an alternative to Java for Android app development. Kotlin offers concise syntax, improved safety, and enhanced readability, making it a preferred choice for many Android developers. Our application integrated Mapbox for real-time mapping and navigation capabilities while providing user authentication through Firebase, ensuring secure access and data collection. Mapbox is a platform that offers mapping and location-based services. It provides tools and APIs for integrating maps and location services into mobile and web applications. Mapbox allows developers to create custom maps, add geolocation features, and build navigation experiences within their apps. Its often used for real-time mapping and navigation solutions. Google Firebase is a comprehensive mobile and web application development platform. It offers a wide range of services, including real-time databases, authentication, cloud storage, and more. Firebase simplifies the development process by providing a unified backend infrastructure, enabling developers to focus on building features rather than managing server infrastructure. It is commonly used for data storage, authentication, and real-time data synchronization in mobile apps.
1. We have used two pages in our application- login and signup and one map navigation option using Mapbox (including traffic view). We have also used different types of icons and backgrounds for the signup and login option and several functions in coding section.
2. We also use speed, time and date and calculate the default speed so that users can concern on the speed and also we can predict what should be the speed on a specific road while travelling.
3. For tracking the location information several types of methods can be used like- by GPS data, by wifi or by cellular data. In our case we used the mobile GPS data to extract the location information.

Data Collection and Storage
The Mapbox component of the application collected crucial location information, including Longitude, Latitude, Speed, and Time, ensuring a continuous feed of real-time data. The data was securely stored in Google Firebase, which also managed user authentication, safeguarding data integrity and user privacy.
1. The users signup details are directly connected to google firebase and also used for authentication that if the user mail already exist in the database or not. When a user will try to login, it will be directly authenticated by the database.
2. In the database, the user details have been listed accordingly to the “Vehicle Number”. Every signup details have been shown under “Vehicle Number” and each users every location information has been shown under each listed users.

Data Export and Conversion
To prepare the data for model training, we extracted it from the Firebase server in JSON format. This data was then meticulously converted into a structured CSV file, ready for use in model training.

LSTM Model Development
With a dataset derived from the CSV file, we embarked on the development and training of a LSTM model. This model was designed to predict various factors, including the estimated arrival time of the vehicle and the prevailing traffic situation. It also played a pivotal role in managing passenger allocation based on real-time data.

RESULTS

The implementation of this project represents a significant advancement in the realm of public transportation. The system setup will encompass the development of a sophisticated Android application integrated with Mapbox for real-time geolocation data collection and Google Firebase for data storage and user authentication. The culmination of this imple-mentation stage leads to the projects evaluation, where the effectiveness of the LSTM model in predicting vehicle locations, estimated arrival times, and traffic conditions is rigorously assessed. This evaluation is complemented by a comprehensive discussion of the results, including the accuracy of predictions and the systems impact on passenger experiences. Through this capstone project, we aim to demonstrate the feasibility and potential transformative impact of our solution in enhancing operational efficiency and service quality within the public transportation sector.

System Setup

A real-time geolocation Android application is at the core of this setup, seamlessly integrated with Google Firebase for data collection and storage. This application continuously captures vital location data, including longitude, latitude, speed, and time. The data is securely stored in Firebase and, upon export as a structured JSON file and converts into CSV file, becomes the foundation for LSTM model training. This model is designed to predict a vehicles future location, estimate arrival times, optimize passenger allocation, and forecast traffic conditions. This system setup provides the necessary infrastructure for improving operational efficiency, scheduling optimization, and   overall service delivery for vehicle companies, thus addressing the challenges faced in public transportation.

Evaluation

The implementation of this project has yielded significant results in advancing public transportation efficiency and passenger satisfaction. The real-time geolocation Android application, integrated with Google Firebase and LSTM, successfully collected and processed vital data, enabling accurate predictions of vehicle location, estimated arrival times, passenger allocation, and traffic conditions. Through rigorous testing and user feedback, the systems performance and usability were verified, resulting in a refined and user-friendly solution. The projects implementation underscores its potential to revolutionize the public transportation sector by optimizing scheduling and enhancing overall service delivery, paving the way for a more efficient and enjoyable travel experience for passengers. The core of the project, the LSTM model, emerged as a powerful tool for predicting vehicle location, estimating arrival times, and managing passenger allocation. The models training with the CSV dataset yielded promising results, offering the potential for precise predictions.

DISCUSSION

Through this implementation, the study has demonstrated promising results. The accuracy of collecting and storing the location information with the speed and time information is very good. The system effectively predicts a vehicles future location on predefined routes according to a specific time. The loss of the model is quite good according to the dataset.

Fig. 4: LSTM Model Performance and loss .

Fig. 5: LSTM Model Performance and loss Graph.

The discussion surrounding these outcomes under-scores the potential for revolutionizing public trans-portation by predicting specific location according to time and as a result improving scheduling, minimizing congestion, and delivering a more satisfying travel experience for passengers. This project lays the foundation for more efficient and sustainable urban mobility, emphasizing the importance of data-driven solutions in modern transportation systems.

Summary

The project, successfully executed a comprehensive system setup. It involved the development of an Android application using Google Firebase and Mapbox for real-time geolocation data collection. The project then meticulously transformed this data into a structured CSV file to train a LSTM model. The results were impressive, with the model effectively predicting a vehicles future location on predefined routes, estimating arrival times, and providing valuable insights for passenger allocation and traffic forecasting. The projects results and their implications were thoughtfully discussed, underscoring the promising future of this innovative solution. This segment highlights the thesis works Standards, Sustainability, Impacts, and Ethics. The commitment to ethical considerations, adherence to industry standards, and a well- defined project plan are essential for the successful development and implementation of the proposed system. Finally, the planned works Schedules, Tasks, and Milestones are shown.

Standards

The project will adhere to industry standards and best practices in several key areas:

Data Privacy and Security Standards 

Ensuring compliance with regulations and ethical principles to safeguard user data and privacy.

Android App Development Standards

Following guidelines and best practices for Android application development to ensure a user-friendly and efficient application.

Firebase Integration Standards

Adhering to Firebase integration best practices for seamless and secure real-time data management.

LSTM Model Implementation Standards 

Employing best practices in implementing Long Short-Term Memory (LSTM) models, a type of sequential neural network, for accurate predictions.

Impact

The proposed project holds significant promise in revolutionizing the landscape of public transportation management. By addressing the inherent challenges faced in the management of vehicles and passengers, the project aims to bring about a transformative impact on operational efficiency, service quality, and overall user experience. In the realm of public transportation, where reliability and efficiency are paramount, the real-time geolocation Android application stands as a beacon of innovation. The integration of Google Firebase, coupled with the utilization of LSTM, a sophisticated sequential neural network, elevates the projects capabilities to predict not only a vehicles future location but also its estimated arrival time, enabling precise scheduling and passenger allocation. The impact of this project extends to various facets of the transportation industry. Efficient route prediction ensures optimal utilization of resources, minimizing fuel consumption and reducing operational costs. The ability to manage passenger allocation in real-time addresses a critical aspect of public transportation, enhancing the overall experience for commuters. Furthermore, the projects forecasting of traffic conditions contributes to proactive decision-making, allowing for the avoidance of congestion and providing a smoother travel experience. In the broader context of environmental sustainability, the project aligns with the growing emphasis on eco-friendly transportation options. By optimizing scheduling and route planning, the project contributes to the reduction of carbon footprints associated with unnecessary fuel consumption and traffic congestion. This aligns with the global push toward sustainable and green transportation solutions. Moreover, the projects reliance on ethical considerations, data privacy standards, and transparent communication with users sets a benchmark for responsible technology deployment. This ensures that the positive impact is not only technological but also extends to user trust and societal well-being. The project has the potential to revolutionize the public transportation sector, making it more efficient, user-friendly, and environmentally conscious. Its impact spans operational enhancements, economic savings, user satisfaction, and contributes to the broader goals of sustainable and responsible technology deployment.

Summary

Schedules, Tasks, and Milestones based on the timeline, we have divided the whole thesis work into 3 parts. In this chapter, we discussed how many times we invested in this project, especially our Schedules, Tasks, and Milestones.

CONCLUSION

In the pursuit of enhancing the efficiency and passenger experience in public transportation, the project titled "Real-time Vehicle Monitoring and Route Prediction using Geolocation through LSTM" has charted a transformative path. The project recognized the inherent advantages of vehicles as an affordable, accessible, and eco-friendly mode of travel, while also acknowledging the common challenges of passenger management, traffic congestion, and service quality. Through a carefully designed approach, the integration of a real-time geolocation Android application, Google Firebase, Mapbox, and LSTM, the project aimed to revolutionize the operational dynamics of vehicle companies. By seamlessly collecting and processing real-time geolocation data, including longitude, latitude, speed, and time, the project laid the foundation for precise predictive analytics. The exported data, structured into a CSV file, became the lifeblood for training LSTM model, enabling the company to forecast vehicle locations, estimate arrival times, manage passenger allocation, and preempt traffic congestions. This forward-thinking approach was not merely a technological innovation; it represented a paradigm shift in public transportation management. The projects implementation, grounded in the synchronization of LSTM model and the Android application, brought these predictions to life, offering passengers a more predictable and enjoyable journey. The passenger allocation strategy and real-time traffic forecasting mechanisms contributed to optimize scheduling and reduced congestion. User feedback and iterative testing validated the systems usability and performance, ensuring its alignment with user needs. The project signifies a substantial leap toward the future of public transportation. It demonstrates that harnessing technology and data-driven insights can reshape how vehicle companies operate, ultimately elevating service delivery, improving passenger experiences, and paving the way for a more efficient and enjoyable public transportation system. As we close this chapter of the project, we look forward to the broader impact this innovation can have on urban mobility, environmental sustainability, and the well-being of passengers and transportation companies alike.

Ethical Considerations

Ensuring that users are well-informed about data collection, processing, and the purpose of the appli-cation, and obtaining explicit consent. Implementing techniques to anonymize and protect user identity in collected geolocation data to preserve privacy. Ensuring fairness in passenger allocation and service delivery without bias or discrimination. Maintaining transparent communication with users about the functionalities and objectives of the application.

Future Work

The project has laid the foundation for significant advancements in public transportation efficiency. While the current implementation addresses many critical aspects of real-time vehicle monitoring and prediction, there remain several avenues for future work and enhancements:

Real-time Passenger Engagement: Extend the Android application to include features that engage passengers during their journey, such as providing entertainment options, information about points of interest, and real-time service updates. This can enhance the overall passenger experience and reduce boredom. 

Advanced Traffic Analysis: Integrate more sophisticated traffic analysis methods, including LSTM models, to improve traffic condition forecasts. Consider factors like historical traffic patterns, weather conditions, and special events that can impact traffic. 

Predictive Maintenance: Implement predictive maintenance features for the vehicle fleet based on real-time data. Predict when maintenance is required, reducing the risk of breakdowns and ensuring vehicle reliability. 

User Customization: Allow passengers to customize their journey experience by selecting preferences for factors like route options, arrival time notifications, or onboard services. This personalization can enhance satisfaction and engagement. 

Integration with Public Transport Networks: Extend the system to integrate with other modes of public transportation, such as trains or trams, to provide passengers with comprehensive multi-modal journey planning. 

Energy Efficiency: Consider eco-friendly factors, such as optimizing routes to minimize fuel consumption, and eventually integrating electric or hybrid vehicles into the fleet to reduce the environmental impact. 

LSTM Model Refinement: Continuously improve the models accuracy by collecting and utilizing more data over time. Employ more advanced techniques and algorithms for prediction. By pursuing these avenues of future work, the project can evolve into a more comprehensive and effective solution for improving public transportation efficiency, passenger satisfaction, and environmental sustainability. It represents an ongoing commitment to the advancement of transportation services in the modern age.

Limitations

While the project holds substantial promise for enhancing public transportation, it is crucial to acknowledge several inherent limitations:

Data Accuracy and Reliability: The accuracy and reliability of real-time geolocation data are contingent on various factors, including the quality of GPS signals, network connectivity, and potential signal loss in areas with limited coverage. Inaccuracies in data may lead to imprecise predictions. 

Data Privacy and Security: Collecting and storing real-time geolocation data necessitates a comprehensive approach to data privacy and security. Ensuring user consent, data anonymization, and protection against data breaches and unauthorized access is essential. 

Resource Requirements: Running a real-time geo-location Android application with LSTM models demands substantial computational resources, both on the device and server-side. Adequate hardware and bandwidth are crucial for smooth operation. 

LSTM Model Accuracy: The accuracy of predictions and forecasting heavily relies on the quality and quantity of historical data used for training the LSTM model. An insufficient or biased dataset may lead to less accurate predictions. 

Predictive Challenges: Predicting future vehicle locations and traffic conditions is inherently challenging due to unforeseen events, such as accidents, road closures, or inclement weather, which can disrupt traffic patterns. 

Traffic Data Sources: Real-time traffic forecasting relies on external data sources, and the availability and accuracy of these sources can vary, affecting the systems effectiveness. Acknowledging these limitations is essential for a comprehensive under-standing of the projects scope and potential constraints. Addressing these challenges will be critical to the successful implementation and deployment of the proposed system.

ACKNOWLEDGEMENT

We would like to express our heartfelt gratitude to the almighty Allah who offered upon our family and us kind care throughout this journey until the fulfilment of this research. Also, we express our sincere respect and gratitude to our supervisor Md. Shahiduzzaman, Assistant Professor, Department of Computer Science and Engineering, Bangladesh University of Business and Technology (BUBT). Without his guidance, this research work would not exist. We are grateful to him for his excellent supervision and for putting his utmost effort into developing this project. We owe him a lot for his assistance, encouragement and guidance, which has shaped our mentality as a researcher. Finally, we are grateful to all our faculty members of the CSE department, BUBT, to make us compatible to complete this research work with the proper guidance and supports throughout the last four years.

CONFLICTS OF INTEREST

There is no conflict of interest in this research.

Supplemental Materials:

| 4.00 KB

Article References:

  1. Ahmed, F.; Bayazid, A.Z.M.; and Al Muntasir, M.F. (2024). The terrible air pollution in Dhaka city is getting worse. GSC Adv. Res. Rev. 19, 42–52. https://doi.org/10.30574/gscarr.2024.19.1.0133 
  2. Bansal, Kashish, et al. (2020). “DeepBus: Machine Learning Based Real Time Pothole Detection System for Smart Transportation Using IoT.” Internet Technology Letters. https://doi.org/10.1002/itl2.156   
  3. Besse P.C. Guillouet B. Loubes J.-M.M. et al. (2018). ‘Destination prediction by trajectory distribution-based model, IEEE Trans. Intell. Transp. Syst., 19, (8), pp. 2470- 2481. 
  4. Besse, P. C., Guillouet, B., & Royer, F. (2018). Destination Prediction by Trajectory Distribution-Based Model. IEEE Transactions on Intelligent Transportation Systems, 19(8), 2470-2481. https://10.1109/tits.2017.2749413  
  5. Choudhury, A., Shanmugavadivelu, S., and Velpuri, B. (2019). Real Time Speed Trend Analysis and Hours of Service Forecasting Using LSTM Network. 2019 IEEE 5th International Conference for Convergence in Technology (I2CT).
  6. D. Dhinakaran, M. R. Khanna, S. P. Panimalar, A. T. P, S. P. Kumar and K. Sudharson, (2022). "Secure Android Location Tracking Application with Privacy Enhanced Technique", Fifth International Conference on Computational Intelligence and Communication Technologies (CCICT), pp. 223-229.
  7. Gui, Zhipeng, et al. (2021). “LSI-LSTM: An Attention-Aware LSTM for Real-Time Driving Destination Prediction by Considering Location Semantics and Location Importance of Trajectory Points.” Neurocomputing, 440, pp. 72-88. https://doi.org/10.1016/j.neucom.2021.01.067 
  8. Huang qiu, et al. (2020). “ConvLSTM Based Transportation Mode Learning from Raw GPS Trajectories.” IET Intelligent Transport Systems. https://doi.org/10.1049/iet-its.2019.0017  
  9. Jisha, R.C., et al. (2017). “Iot Based School Bus Tracking and Arrival Time Prediction.” IEEE Xplore, 1. https://ieeexplore.ieee.org/abstract/document/8125890  
  10. Julio, Nikolas, et al. (2016). “Real-Time Prediction of Bus Travel Speeds Using Traffic Shockwaves and Machine Learning Algorithms.” Res. in Transportation Economics, 59, pp. 250-257. https://doi.org/10.1016/j.retrec.2016.07.019   
  11. Kamble, Shridevi Jeevan, and Manjunath R Kounte. (2020).“Machine Learning Approach on Traffic Congestion Monitoring System in Internet of Vehicles.” Procedia Computer Science, 171, pp. 2235-2241. https://doi.org/10.1016/j.procs.2020.04.241 
  12. Krause, Cory M., and Lei Zhang. (2019). “Short-Term Travel Behavior Prediction with GPS,   Land Use, and Point of Interest Data.” Transportation Research Part B:       Methodological, 123, pp. 349-361. https://doi.org/10.1016/j.trb.2018.06.012 
  13. Liu, H., Xu, H., & Li, W. (2020). Bus Arrival Time Prediction Based on LSTM and Spatial-Temporal Feature Vector. IEEE Access, 8, 11917-11929. https://10.1109/access.2020.2965094  
  14. Meng, X., Fu, H., & Chen, E. (2020). D-LSTM: Short-Term Road Traffic Speed Prediction Model Based on GPS Positioning Data. IEEE Transactions on Intelligent Transportation Systems, 1-10. https://10.1109/tits.2020.3030546 
  15. Mikhailov, Sergei, and Alexey Kashevnik. (2021). “Car Tourist Trajectory Prediction Based on Bidirectional LSTM Neural Network.” Electronics, 10(12). p. 1390. https://doi.org/10.3390/electronics10121390  
  16. Sami HM, and Arifuzzaman SM. (2021). Comparing pure stock portfolio with stock and crypto-currency mixed portfolio through LSTM to compare & analyze investment opportunities for portfolio performance measurement, Aust. J. Eng. Innov. Technol., 3(3), 45-56. https://doi.org/10.34104/ajeit.021.045056 
  17. Skhosana, M., Ezugwu, A.E. (2021). A real-time machine learning-based public transport bus-passenger information system. Inter J. of Sustainable Development and Planning, 16(7), pp. 1221-1238. https://doi.org/10.18280/ijsdp.160703 
  18. Skhosana, Menzi. (1970). “Irenbus - a Real-Time Machine Learning Based Public Transport Management System.” ResearchSpace Home, researchspace. https://doi.org/ukzn.ac.za/handle/10413/21011 
  19. Skhosana, Menzi, et al. (2020). “An intelligent machine learning-based real-time public transport system.” Computational Science and Its Applications – ICCSA 2020, pp. 649-665. https://doi.org/10.1007/978-3-030-58817-5_47 
  20. Sun, Jie, and Jiwon Kim. (2021). “Joint Prediction of next Location and Travel Time from Urban Vehicle Trajectories Using Long Short-Term Memory Neural Networks.” Transportation Research Part C: Emerging Technologies, 128, p. 103114. https://doi.org/10.1016/j.trc.2021.103114  
  21. Souza, Vinicius M.A., et al. (2018). “Asfault: A Low-Cost System to Evaluate Pavement Conditions in Real-Time Using Smartphones and Machine Learning.” Pervasive and Mobile Computing, 51, pp. 121-137. https://doi.org/10.1016/j.pmcj.2018.10.008  
  22. Seitbekova, Yerkezhan, et al. (2020). “The Bus Arrival Time Prediction Using LSTM Neural Network and Location Analysis.” J. of Advances in Technology and Engineering Research, 6(2), pp. 46-57. https://ideas.repec.org/a/apb/jaterr/2020p46-57. html 
  23. Zhao, J., Gao, Y., & Sun, H. (2018). Travel Time Prediction: Based on Gated Recurrent Unit Method and Data Fusion. IEEE Access, 1-1. https://10.1109/access.2018.2878799   
  24. Zhao, J., Gao, Y., & Lu, S. (2019). Traffic Speed Prediction Under Non-Recurrent Congestion: Based on LSTM Method and Bei Dou Navigation Satellite System Data. IEEE Intelligent Transportation Systems Magazine, 11(2), 70-81. https://10.1109/mits.2019.2903431 

Article Info:

Academic Editor

Dr. Wiyanti Fransisca Simanullang Assistant Professor Department of Chemical Engineering Universitas Katolik Widya Mandala Surabaya East Java, Indonesia.

Received

September 1, 2024

Accepted

October 22, 2024

Published

October 29, 2024

Article DOI: Aust. J. Eng. Innov. Technol., 2024; 6(5), 104-121

Coresponding author

Md. Al-Amin Hossain Bijoy*

Department of Computer Science and Engineering, Bangladesh University of Business and Technology, Dhaka, Bangladesh.

Cite this article

Bijoy MAH, Sarker P, Siam NS, Ahsan MN, Hossain M, and Shahiduzzaman M. (2024). Real-time vehicle monitoring and route prediction using geolocation through LSTM. Aust. J. Eng. Innov. Technol., 6(5), 104-121. https://doi.org/10.34104/ajeit.024.01040121

Views
24
Download
4
Citations
Badge Img
Share