Ivo Müürsepp
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Ivo Müürsepp has a total of 13 co-authors. Here are the top co-authors by number of shared publications:
- Muhammad Mahtab Alam
Tallinn University of Technology
8 shared publications - Mihkel Tommingas
Thomas Johann Seebeck Department of Electronics, Tallinn University of Technology, Tallinn, Estonia
3 shared publications - Marika Kulmar
Thomas Johann Seebeck Department of Electronics, Tallinn University of Technology, Tallinn, Estonia
2 shared publications - Sander Varbla
Thomas Johann Seebeck Department of Electronics, Tallinn University of Technology, Tallinn, Estonia
2 shared publications - Alar Kuusik
Tallinn University of Technology
1 shared publication - Anum Umer
Thomas Johann Seebeck Department of Electronics, Tallinn University of Technology, Tallinn, Estonia
1 shared publication
Latest Publications
- Open Access
DL-AoD Estimation-Based 5G Positioning Using Directionally Transmitted Synchronization Signals
This paper introduces a method for estimating the Downlink Angle of Departure (DL-AoD) of 5G User Equipment (UE) from measured signal strengths of directionally transmitted synchronization signals. Based on estimated DL-AoD values, from two or more anchor nodes, the position of the UE was estimated. Unlike most prior work, which is simulation-based or relies on custom testbeds, this study uses real measurements from an operational 5G network in an industrial factory environment. A deterministic estimator was derived, but multipath and unknown beam characteristics limit its accuracy. To address this, machine learning was applied to automatically adapt to the environment. Previous simulation studies reported 90th-percentile DL-AoD estimation errors below 2°, while experimental works achieved best-case accuracies of 5–6°. In this study, the experimental DL-AoD estimation error remained below 4° for 90% of the measurements, indicating improved real-world performance. Reported positioning errors in the literature range from 3.8 m to 140 m, whereas the 13.2 m error obtained here lies near the midpoint of this range, confirming the practicality of the proposed method in industrial environments. Compared to existing approaches, this work demonstrates high angular accuracy using only sub-6 GHz beams in a realistic industrial scenario without detailed knowledge of antenna beam patterns and channel state. The findings demonstrate that standard 5G signals can provide accurate indoor localization without additional infrastructure, offering a practical path toward cost-effective positioning in industrial IoT and automation.
Open AccessDL-AoD Estimation-Based 5G Positioning Using Directionally Transmitted Synchronization Signals
This paper introduces a method for estimating the Downlink Angle of Departure (DL-AoD) of 5G User Equipment (UE) from measured signal strengths of directionally transmitted synchronization signals. Based on estimated DL-AoD values, from two or more anchor nodes, the position of the UE was estimated. Unlike most prior work, which is simulation-based or relies on custom testbeds, this study uses real measurements from an operational 5G network in an industrial factory environment. A deterministic estimator was derived, but multipath and unknown beam characteristics limit its accuracy. To address this, machine learning was applied to automatically adapt to the environment. Previous simulation studies reported 90th-percentile DL-AoD estimation errors below 2°, while experimental works achieved best-case accuracies of 5–6°. In this study, the experimental DL-AoD estimation error remained below 4° for 90% of the measurements, indicating improved real-world performance. Reported positioning errors in the literature range from 3.8 m to 140 m, whereas the 13.2 m error obtained here lies near the midpoint of this range, confirming the practicality of the proposed method in industrial environments. Compared to existing approaches, this work demonstrates high angular accuracy using only sub-6 GHz beams in a realistic industrial scenario without detailed knowledge of antenna beam patterns and channel state. The findings demonstrate that standard 5G signals can provide accurate indoor localization without additional infrastructure, offering a practical path toward cost-effective positioning in industrial IoT and automation.
- Open Access
- 2 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of the Communications Society
- Vol. 6, 4383
Erratum to “UWB and GNSS Sensor Fusion Using ML-Based Positioning Uncertainty Estimation”
Presents corrections to the paper, (Erratum to “UWB and GNSS Sensor Fusion Using ML-Based Positioning Uncertainty Estimation”).
- 2 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of the Communications Society
- Vol. 6, 4383
Open AccessErratum to “UWB and GNSS Sensor Fusion Using ML-Based Positioning Uncertainty Estimation”
Presents corrections to the paper, (Erratum to “UWB and GNSS Sensor Fusion Using ML-Based Positioning Uncertainty Estimation”).
- Open Access
- 26 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of the Communications Society
- Vol. 6, 2177-2189
UWB and GNSS Sensor Fusion Using ML-Based Positioning Uncertainty Estimation
This article presents a novel machine learning (ML) augmented sensor fusion scheme for seamless indoor-outdoor positioning using ultra-wideband (UWB) and global navigation satellite system (GNSS) sensors. Dilution of precision (DOP) is a common metric that describes the level of geometrical uncertainty and is often applied in sensor fusion schemes. However, it does not fully reflect other factors that may influence positioning integrity, such as signal quality, pseudorange error, or the number of servicing nodes. An incorrect estimation of sensor uncertainty may significantly affect the precision and robustness of individual sensors and their fused positioning solution, especially in indoor-outdoor transition zones, where positioning is most challenging. Therefore, this article proposes a sensor fusion scheme augmented with two distinct extreme gradient boosted (XGBoost) ML models for estimating UWB and GNSS positioning uncertainties. Trained using real-life datasets, these models have the advantage of considering an ensemble of features rather than a single parameter to estimate the uncertainty of the current coordinate. In contrast to sensor fusion solutions, which implement only highly accurate RTK fixed solution mode, the GNSS model can operate with different correction qualities as well. The proposed scheme was applied on a moving testbed with UWB and GNSS sensors while relying on the ML models to enhance the coordinate filtering. The results show that the ML-based approach can improve seamless transition between indoor and outdoor areas with almost no sensor dropouts with a mean positioning error of 0.16 m and a maximum error of approximately 0.5 m.
- 26 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of the Communications Society
- Vol. 6, 2177-2189
Open AccessUWB and GNSS Sensor Fusion Using ML-Based Positioning Uncertainty Estimation
This article presents a novel machine learning (ML) augmented sensor fusion scheme for seamless indoor-outdoor positioning using ultra-wideband (UWB) and global navigation satellite system (GNSS) sensors. Dilution of precision (DOP) is a common metric that describes the level of geometrical uncertainty and is often applied in sensor fusion schemes. However, it does not fully reflect other factors that may influence positioning integrity, such as signal quality, pseudorange error, or the number of servicing nodes. An incorrect estimation of sensor uncertainty may significantly affect the precision and robustness of individual sensors and their fused positioning solution, especially in indoor-outdoor transition zones, where positioning is most challenging. Therefore, this article proposes a sensor fusion scheme augmented with two distinct extreme gradient boosted (XGBoost) ML models for estimating UWB and GNSS positioning uncertainties. Trained using real-life datasets, these models have the advantage of considering an ensemble of features rather than a single parameter to estimate the uncertainty of the current coordinate. In contrast to sensor fusion solutions, which implement only highly accurate RTK fixed solution mode, the GNSS model can operate with different correction qualities as well. The proposed scheme was applied on a moving testbed with UWB and GNSS sensors while relying on the ML models to enhance the coordinate filtering. The results show that the ML-based approach can improve seamless transition between indoor and outdoor areas with almost no sensor dropouts with a mean positioning error of 0.16 m and a maximum error of approximately 0.5 m.
- Open Access
- 12 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in TechRxiv
DL-AoD Estimation Based 5G Positioning Using Directionally Transmitted Synchronization Signals
This paper first presents a novel method for estimating Downlink Angle of Departure (DL-AoD) of the 5G User Equipment (UE) based on measured signal strengths of directionally transmitted synchronization Signals. Results obtained with described method can be then used for mobile positioning purposes. Machine learning was employed to automatically find and implement unknown values of parameters necessary for precise DL-AoD estimation. Experiments conducted in an industrial environment showed that proposed method can have median value of estimation error below half a degree. Estimated angles were later used for indoor positioning. Median and the mean of obtained positioning errors were 5 and 4.1 m correspondingly.
- 12 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in TechRxiv
Open AccessDL-AoD Estimation Based 5G Positioning Using Directionally Transmitted Synchronization Signals
This paper first presents a novel method for estimating Downlink Angle of Departure (DL-AoD) of the 5G User Equipment (UE) based on measured signal strengths of directionally transmitted synchronization Signals. Results obtained with described method can be then used for mobile positioning purposes. Machine learning was employed to automatically find and implement unknown values of parameters necessary for precise DL-AoD estimation. Experiments conducted in an industrial environment showed that proposed method can have median value of estimation error below half a degree. Estimated angles were later used for indoor positioning. Median and the mean of obtained positioning errors were 5 and 4.1 m correspondingly.
- Open Access
- 11 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 187290-187301
Empirical Evaluation of NB-IoT and CAT-M Coverage for Underground Water System
Narrow-band Internet of Things (NB-IoT) and LTE CAT-M are part of low-power wide-area network (LPWAN) access technologies that provide long-range and energy-efficient network access to IoT devices. In this study, we conducted empirical measurements on both NB-IoT and LTE CAT-M networks inside a street manhole. Street manholes are an integral part of smart city structures, where sensor implementation is important. Our study demonstrates how to derive the network parameters from system information (SI) messages. More precisely, as we measured three commercial operators, we show the parameters and their influence on enhanced coverage level (ECL) repetitions. We give special consideration to signal strength and quality measurements at different locations inside the manhole. Our results indicate that when designing sensor networks for urban infrastructure monitoring, NB-IoT and LTE CAT-M should have different coverage margins in the implementation. Robust and low cost NB-IoT construction leads to the consideration of up to 2.4 dB higher coverage margins compared to LTE CAT-M. In this work, we analyze various environmental conditions that are relevant when evaluating and understanding underground signal propagation.
- 11 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 187290-187301
Open AccessEmpirical Evaluation of NB-IoT and CAT-M Coverage for Underground Water System
Narrow-band Internet of Things (NB-IoT) and LTE CAT-M are part of low-power wide-area network (LPWAN) access technologies that provide long-range and energy-efficient network access to IoT devices. In this study, we conducted empirical measurements on both NB-IoT and LTE CAT-M networks inside a street manhole. Street manholes are an integral part of smart city structures, where sensor implementation is important. Our study demonstrates how to derive the network parameters from system information (SI) messages. More precisely, as we measured three commercial operators, we show the parameters and their influence on enhanced coverage level (ECL) repetitions. We give special consideration to signal strength and quality measurements at different locations inside the manhole. Our results indicate that when designing sensor networks for urban infrastructure monitoring, NB-IoT and LTE CAT-M should have different coverage margins in the implementation. Robust and low cost NB-IoT construction leads to the consideration of up to 2.4 dB higher coverage margins compared to LTE CAT-M. In this work, we analyze various environmental conditions that are relevant when evaluating and understanding underground signal propagation.
- Open Access
- 25 July 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 103082-103095
Management Closed Control Loop Automation for Improved RAN Slice Performance by Subslicing
Network slicing applied to a radio access network (RAN) offers the potential to enhance the satisfaction of service requirements and optimize resource utilization, particularly in scenarios where radio resources are limited. The dependence of RAN slice performance on the allocated bandwidth part (BWP) size has not been investigated. In this paper, we use the acquired knowledge of RAN slice performance dependence on the number of allocated resource blocks (RB) for automatic subslicing to improve slice performance on a fixed BWP size. Subslicing means that the slice user equipments (UE) are clustered, slice BWP is subpartitioned, and slice RBs are allocated to the UE groups. Based on six key performance indicators (KPI), it is decided for each subslice whether it should be split, merged, or not changed for its performance improvement. Performance improvement means that the slice utilization and block error ratio (BLER) are reduced and goodput (application-level throughput) per one resource block (RB) is increased. The monitor-analyze-plan-execute-knowledge (MAPE-K)-type management closed control loop (MCCL) was implemented for slice performance improvement by subslicing. A realistic 5G new radio (NR) band serving a set of UEs of different values of their BLER and requested rates was a setup of the RAN slice simulated using MATLAB R2021b. By automatic subslicing, the slice utilization can be reduced by up to 7%, goodput per one RB by up to 38%, and slice BLER by up to 60%. This effect was greater for the uplink. The best slice performance improvement is achieved by automatic subslicing if the UEs are clustered by their achieved BLER.
- 25 July 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 103082-103095
Open AccessManagement Closed Control Loop Automation for Improved RAN Slice Performance by Subslicing
Network slicing applied to a radio access network (RAN) offers the potential to enhance the satisfaction of service requirements and optimize resource utilization, particularly in scenarios where radio resources are limited. The dependence of RAN slice performance on the allocated bandwidth part (BWP) size has not been investigated. In this paper, we use the acquired knowledge of RAN slice performance dependence on the number of allocated resource blocks (RB) for automatic subslicing to improve slice performance on a fixed BWP size. Subslicing means that the slice user equipments (UE) are clustered, slice BWP is subpartitioned, and slice RBs are allocated to the UE groups. Based on six key performance indicators (KPI), it is decided for each subslice whether it should be split, merged, or not changed for its performance improvement. Performance improvement means that the slice utilization and block error ratio (BLER) are reduced and goodput (application-level throughput) per one resource block (RB) is increased. The monitor-analyze-plan-execute-knowledge (MAPE-K)-type management closed control loop (MCCL) was implemented for slice performance improvement by subslicing. A realistic 5G new radio (NR) band serving a set of UEs of different values of their BLER and requested rates was a setup of the RAN slice simulated using MATLAB R2021b. By automatic subslicing, the slice utilization can be reduced by up to 7%, goodput per one RB by up to 38%, and slice BLER by up to 60%. This effect was greater for the uplink. The best slice performance improvement is achieved by automatic subslicing if the UEs are clustered by their achieved BLER.
- Open Access
- 24 June 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Indoor and Seamless Positioning and Navigation
- Vol. 2, 205-218
UWB Positioning Integrity Estimation Using Ranging Residuals and ML Augmented Filtering
This article investigates the use of Ultra-Wideband (UWB) ranging residuals for coordinate integrity estimation and their use in a filtering scheme. Typically, UWB system accuracy is improved using channel statistics to detect and mitigate non-line of sight (NLOS) effects between UWB sensors and the object to be located, potentially improving the end coordinate solution. However, in practice when considering UWB system with a high positioning update rate, this is not a feasible approach, as gathering and processing CS data takes too much time. In contrast to this approach, this article proposes a set of features based on UWB ranging residuals that could be used as an alternative in integrity assessment. By using machine learning (ML), the most important features were extracted from the initial set and then used to train and validate a model for UWB coordinate error prediction. Lastly, the prediction was applied in an adaptive Kalman filtering (AKF) scheme as an input for measurement uncertainty. Model testing was done using UWB measurement test dataset gathered at an industrial site. The overall results showed significant improvement in 2D and 3D positioning metrics of ML-augmented filtering when compared to non-ML filtering. On average, the end coordinates in the test set had ca. 10 cm smaller mean location error and ca. 40 cm smaller dispersion in 2D positioning. Additionally, the presence of outliers was reduced significantly as the maximum error offset decreased by several meters. Although ML augmented filtering is computationally slower than non-ML filtering (e.g., ordinary and extended Kalman filter), it is still faster than using channel statistics for UWB integrity estimation. The results show that using the proposed residual features in an ML model provides a feasible approach to predict UWB positioning integrity and use it as a measure of uncertainty in a coordinate filtering scheme.
- 24 June 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Indoor and Seamless Positioning and Navigation
- Vol. 2, 205-218
Open AccessUWB Positioning Integrity Estimation Using Ranging Residuals and ML Augmented Filtering
This article investigates the use of Ultra-Wideband (UWB) ranging residuals for coordinate integrity estimation and their use in a filtering scheme. Typically, UWB system accuracy is improved using channel statistics to detect and mitigate non-line of sight (NLOS) effects between UWB sensors and the object to be located, potentially improving the end coordinate solution. However, in practice when considering UWB system with a high positioning update rate, this is not a feasible approach, as gathering and processing CS data takes too much time. In contrast to this approach, this article proposes a set of features based on UWB ranging residuals that could be used as an alternative in integrity assessment. By using machine learning (ML), the most important features were extracted from the initial set and then used to train and validate a model for UWB coordinate error prediction. Lastly, the prediction was applied in an adaptive Kalman filtering (AKF) scheme as an input for measurement uncertainty. Model testing was done using UWB measurement test dataset gathered at an industrial site. The overall results showed significant improvement in 2D and 3D positioning metrics of ML-augmented filtering when compared to non-ML filtering. On average, the end coordinates in the test set had ca. 10 cm smaller mean location error and ca. 40 cm smaller dispersion in 2D positioning. Additionally, the presence of outliers was reduced significantly as the maximum error offset decreased by several meters. Although ML augmented filtering is computationally slower than non-ML filtering (e.g., ordinary and extended Kalman filter), it is still faster than using channel statistics for UWB integrity estimation. The results show that using the proposed residual features in an ML model provides a feasible approach to predict UWB positioning integrity and use it as a measure of uncertainty in a coordinate filtering scheme.
- Open Access
Heuristic Radio Access Network Subslicing with User Clustering and Bandwidth Subpartitioning
In 5G and beyond, the network slicing is a crucial feature that ensures the fulfillment of service requirements. Nevertheless, the impact of the number of slices and slice size on the radio access network (RAN) slice performance has not yet been studied. This research is needed to understand the effects of creating subslices on slice resources to serve slice users and how the performance of RAN slices is affected by the number and size of these subslices. A slice is divided into numbers of subslices of different sizes, and the slice performance is evaluated based on the slice bandwidth utilization and slice goodput. A proposed subslicing algorithm is compared with k-means UE clustering and equal UE grouping. The MATLAB simulation results show that subslicing can improve slice performance. If the slice contains all UEs with a good block error ratio (BLER), then a slice performance improvement of up to 37% can be achieved, and it comes more from the decrease in bandwidth utilization than the increase in goodput. If a slice contains UEs with a poor BLER, then the slice performance can be improved by up to 84%, and it comes only from the goodput increase. The most important criterion in subslicing is the minimum subslice size in terms of resource blocks (RB), which is 73 for a slice that contains all good-BLER UEs. If a slice contains UEs with poor BLER, then the subslice can be smaller.
Open AccessHeuristic Radio Access Network Subslicing with User Clustering and Bandwidth Subpartitioning
In 5G and beyond, the network slicing is a crucial feature that ensures the fulfillment of service requirements. Nevertheless, the impact of the number of slices and slice size on the radio access network (RAN) slice performance has not yet been studied. This research is needed to understand the effects of creating subslices on slice resources to serve slice users and how the performance of RAN slices is affected by the number and size of these subslices. A slice is divided into numbers of subslices of different sizes, and the slice performance is evaluated based on the slice bandwidth utilization and slice goodput. A proposed subslicing algorithm is compared with k-means UE clustering and equal UE grouping. The MATLAB simulation results show that subslicing can improve slice performance. If the slice contains all UEs with a good block error ratio (BLER), then a slice performance improvement of up to 37% can be achieved, and it comes more from the decrease in bandwidth utilization than the increase in goodput. If a slice contains UEs with a poor BLER, then the slice performance can be improved by up to 84%, and it comes only from the goodput increase. The most important criterion in subslicing is the minimum subslice size in terms of resource blocks (RB), which is 73 for a slice that contains all good-BLER UEs. If a slice contains UEs with poor BLER, then the subslice can be smaller.