Smart Connectivity for Teleoperated Vehicles Using a Predictive LTE Multi-Operator Approach
Ahmed Maher Mohamed Hamed
Smart Connectivity for Teleoperated Vehicles Using a Predictive LTE Multi-Operator Approach /Ahmed Maher Mohamed Hamed - 2024 - 64p. ill. 21 cm.
Supervisor:
Nashwa AbdelBaki
Thesis (M.A.)—Nile University, Egypt, 2024 .
"Includes bibliographical references"
Contents:
DEDICATION................................................................................................................i
ACKNOWLEDGEMENTS...........................................................................................ii
TABLE OF CONTENTS..............................................................................................iv
LIST OF FIGURES ......................................................................................................vi
LIST OF TABLES.......................................................................................................vii
ACRONYMS..............................................................................................................viii
PUBLICATIONS..........................................................................................................xi
ABSTRACT................................................................................................................xiii
Chapter 1 Introduction ...................................................................................................1
1.1 Connectivity-Based Services in Business............................................................1
1.2 Teleoperated Services..........................................................................................2
1.3 Automotive Industry Assistive Driving ...............................................................4
1.4 Connectivity Service............................................................................................5
1.5 Connectivity Demand Growth .............................................................................6
1.6 The need behind Multi-Operator Connection ....................................................10
1.7 Thesis Objectives...............................................................................................11
1.8 Thesis Outline ....................................................................................................12
Chapter 2 Related Teleoperating Idea & Prediction Model Data ................................13
2.1 Teleoperative Supporting Devices.....................................................................13
2.2 Teleoperating with Cellular ...............................................................................18
2.3 Prediction Model and Correlation of the Environment Parameters...................21
Chapter 3 Accurate Data Collection for Modeling ......................................................25
3.1 Collecting Equipment ........................................................................................25
3.2 Selected Test Area Criteria ................................................................................25
3.3 Preforming Data Collection ...............................................................................27
Chapter 4 Feature Engineering for Improved Predection ............................................36
4.1 Data Cleansing & Labelling...............................................................................37
4.1.1 Labelling Level 1: Basic Suitability .......................................................................... 37
4.1.2 Labelling Level 2: Prioritization with Weighted Scores .......................................... 37
4.1.3 Tie-Breaking Mechanism ........................................................................................ 38
4.1.4 Addressing Unsuitable Data for Model Training .................................................... 39
4.1.5 Data Features: Building the Model's Foundation................................................... 40
4.2 Predictive Modelling..........................................................................................43
4.2.1 Dividing the Data: Training and Testing Sets.......................................................... 45
4.2.2 Addressing Data Imbalance: Ensuring a Fair Playing Field for Operators .............. 46
4.2.3 Evaluating the Impact of Balanced Data: Confusion Matrix and Optimal Choice.. 48
4.2.4 Best Model Results ................................................................................................. 51
Chapter 5 Conclusion and Future Work ......................................................................55
5.1 Conclusion .........................................................................................................55
5.1.1 Building Upon Existing Work: Comparing Approaches to Throughput Prediction 56
5.1.2 Model Selection and Verification ........................................................................... 56
5.1.3 Success Without Private Data................................................................................. 57
5.1.4 Moving Forward: Beyond Throughput ................................................................... 57
5.1.5 Road Sample Distribution and Localized Impact.................................................... 58
5.1.6 Impact on Model Performance............................................................................... 58
5.2 Future Work .......................................................................................................59
5.2.1 Enhancing Data Streams and Sensor Fusion .......................................................... 59
5.2.2 Investigating More Complex Models: Beyond Simple ANNs.................................. 60
5.2.3 Data Intensity Through Crowd-Sourced Collection................................................ 61
Abstract:
The demand for reliable connectivity-driven services is on the rise, leading to
increased sensitivity in technologies like Advanced Driver-Assistance Systems
(ADAS). ADAS represents a prevalent technological advancement in modern
vehicles, with the aspiration of achieving trustworthy autonomous driving as the
ultimate goal. From both end-user and manufacturer perspectives, we are assessing
Teleoperated Driving as a promising feature to address emerging needs for traffic
management and health and safety precautions. Human-to-human interaction and
perception have been found to be superior to human-to-machine interaction in
handling tasks, such as human driving compared to machine driving.
As this entire service relies on sensors which it is already implemented by various
car manufacturers, and connectivity. It is obvious that the quality of the connectivity
is expected to change in terms of coverage and capacity. We studied the feasibility
of predicting the most suitable market operator within a specific area. This is based
on previously established criteria, to serve as the primary data connection before
engaging in a new measurement delay. To this end, an extensive measurement
period was conducted. Establishing error margins in data collection and refining the
data filtering process were meant to accurately capture the specifically targeted road
conditions. These steps were completed prior to the commencement of our analysis
of different models.
Smart Connectivity for Teleoperated
Vehicles Using a Predictive LTE
Multi-Operator Approach
Ahmed Maher Mohamed Hamed
xiv
These efforts have strengthened our study's focus on confirming the feasibility of
seamless transitions between different operators, ensuring the fulfilment of the
necessary teleoperation conditions.
Keywords: Cellular networks, teleoperated vehicle, teleoperated driving, latency,
throughput, 4G, LTE, predication models
Text in English, abstracts in English and Arabic
0000-0003-3128-9559 ORCID
informatics
Dissertation, Academic
610
Smart Connectivity for Teleoperated Vehicles Using a Predictive LTE Multi-Operator Approach /Ahmed Maher Mohamed Hamed - 2024 - 64p. ill. 21 cm.
Supervisor:
Nashwa AbdelBaki
Thesis (M.A.)—Nile University, Egypt, 2024 .
"Includes bibliographical references"
Contents:
DEDICATION................................................................................................................i
ACKNOWLEDGEMENTS...........................................................................................ii
TABLE OF CONTENTS..............................................................................................iv
LIST OF FIGURES ......................................................................................................vi
LIST OF TABLES.......................................................................................................vii
ACRONYMS..............................................................................................................viii
PUBLICATIONS..........................................................................................................xi
ABSTRACT................................................................................................................xiii
Chapter 1 Introduction ...................................................................................................1
1.1 Connectivity-Based Services in Business............................................................1
1.2 Teleoperated Services..........................................................................................2
1.3 Automotive Industry Assistive Driving ...............................................................4
1.4 Connectivity Service............................................................................................5
1.5 Connectivity Demand Growth .............................................................................6
1.6 The need behind Multi-Operator Connection ....................................................10
1.7 Thesis Objectives...............................................................................................11
1.8 Thesis Outline ....................................................................................................12
Chapter 2 Related Teleoperating Idea & Prediction Model Data ................................13
2.1 Teleoperative Supporting Devices.....................................................................13
2.2 Teleoperating with Cellular ...............................................................................18
2.3 Prediction Model and Correlation of the Environment Parameters...................21
Chapter 3 Accurate Data Collection for Modeling ......................................................25
3.1 Collecting Equipment ........................................................................................25
3.2 Selected Test Area Criteria ................................................................................25
3.3 Preforming Data Collection ...............................................................................27
Chapter 4 Feature Engineering for Improved Predection ............................................36
4.1 Data Cleansing & Labelling...............................................................................37
4.1.1 Labelling Level 1: Basic Suitability .......................................................................... 37
4.1.2 Labelling Level 2: Prioritization with Weighted Scores .......................................... 37
4.1.3 Tie-Breaking Mechanism ........................................................................................ 38
4.1.4 Addressing Unsuitable Data for Model Training .................................................... 39
4.1.5 Data Features: Building the Model's Foundation................................................... 40
4.2 Predictive Modelling..........................................................................................43
4.2.1 Dividing the Data: Training and Testing Sets.......................................................... 45
4.2.2 Addressing Data Imbalance: Ensuring a Fair Playing Field for Operators .............. 46
4.2.3 Evaluating the Impact of Balanced Data: Confusion Matrix and Optimal Choice.. 48
4.2.4 Best Model Results ................................................................................................. 51
Chapter 5 Conclusion and Future Work ......................................................................55
5.1 Conclusion .........................................................................................................55
5.1.1 Building Upon Existing Work: Comparing Approaches to Throughput Prediction 56
5.1.2 Model Selection and Verification ........................................................................... 56
5.1.3 Success Without Private Data................................................................................. 57
5.1.4 Moving Forward: Beyond Throughput ................................................................... 57
5.1.5 Road Sample Distribution and Localized Impact.................................................... 58
5.1.6 Impact on Model Performance............................................................................... 58
5.2 Future Work .......................................................................................................59
5.2.1 Enhancing Data Streams and Sensor Fusion .......................................................... 59
5.2.2 Investigating More Complex Models: Beyond Simple ANNs.................................. 60
5.2.3 Data Intensity Through Crowd-Sourced Collection................................................ 61
Abstract:
The demand for reliable connectivity-driven services is on the rise, leading to
increased sensitivity in technologies like Advanced Driver-Assistance Systems
(ADAS). ADAS represents a prevalent technological advancement in modern
vehicles, with the aspiration of achieving trustworthy autonomous driving as the
ultimate goal. From both end-user and manufacturer perspectives, we are assessing
Teleoperated Driving as a promising feature to address emerging needs for traffic
management and health and safety precautions. Human-to-human interaction and
perception have been found to be superior to human-to-machine interaction in
handling tasks, such as human driving compared to machine driving.
As this entire service relies on sensors which it is already implemented by various
car manufacturers, and connectivity. It is obvious that the quality of the connectivity
is expected to change in terms of coverage and capacity. We studied the feasibility
of predicting the most suitable market operator within a specific area. This is based
on previously established criteria, to serve as the primary data connection before
engaging in a new measurement delay. To this end, an extensive measurement
period was conducted. Establishing error margins in data collection and refining the
data filtering process were meant to accurately capture the specifically targeted road
conditions. These steps were completed prior to the commencement of our analysis
of different models.
Smart Connectivity for Teleoperated
Vehicles Using a Predictive LTE
Multi-Operator Approach
Ahmed Maher Mohamed Hamed
xiv
These efforts have strengthened our study's focus on confirming the feasibility of
seamless transitions between different operators, ensuring the fulfilment of the
necessary teleoperation conditions.
Keywords: Cellular networks, teleoperated vehicle, teleoperated driving, latency,
throughput, 4G, LTE, predication models
Text in English, abstracts in English and Arabic
0000-0003-3128-9559 ORCID
informatics
Dissertation, Academic
610