1. |
- Azari, Amin, 1988-, et al.
(författare)
-
Machine Learning assisted Handover and Resource Management for Cellular Connected Drones
- 2020
-
Ingår i: Proceedings of the IEEE Vehicular Technology Conference. - : Institute of Electrical and Electronics Engineers (IEEE).
-
Konferensbidrag (refereegranskat)abstract
- Cellular connectivity for drones comes with a wide set of challenges as well as opportunities. Communication of cellular-connected drones is influenced by 3-dimensional mobility and line-of-sight channel characteristics which results in higher number of handovers with increasing altitude. Our cell planning simulations in coexistence of aerial and terrestrial users indicate that the severe interference from drones to base stations is a major challenge for uplink communications of terrestrial users. Here, we first present the major challenges in co-existence of terrestrial and drone communications by considering real geographical network data for Stockholm. Then, we derive analytical models for the key performance indicators (KPIs), including communications delay and interference over cellular networks, and formulate the handover and radio resource management (H-RRM) optimization problem. Afterwards, we transform this problem into a machine learning problem, and propose a deep reinforcement learning solution to solve HRRM problem. Finally, using simulation results, we present how the speed and altitude of drones, and the tolerable level of interference, shape the optimal H-RRM policy in the network. Especially, the heat-maps of handover decisions for different altitudes/speeds of drones have been presented, which promote a revision of the legacy handover schemes and boundaries of cells in the sky.
|
|
2. |
- Azari, Amin, et al.
(författare)
-
Risk-Aware Resource Allocation for URLLC : Challenges and Strategies with Machine Learning
- 2019
-
Ingår i: IEEE Communications Magazine. - : Institute of Electrical and Electronics Engineers (IEEE). - 0163-6804 .- 1558-1896. ; 57:3, s. 42-48
-
Tidskriftsartikel (refereegranskat)abstract
- URLLC) is a major challenge of 5G wireless networks. Stringent delay and reliability requirements need to be satisfied for both scheduled and non-scheduled URLLC traffic to enable a diverse set of 5G applications. Although physical and media access control layer solutions have been investigated to satisfy only scheduled URLLC traffic, there is a lack of study on enabling transmission of non-scheduled URLLC traffic, especially in coexistence with the scheduled URLLC traffic. Machine learning (ML) is an important enabler for such a coexistence scenario due to its ability to exploit spatial/temporal correlation in user behaviors and use of radio resources. Hence, in this paper, we first study the coexistence design challenges, especially the radio resource management (RRM) problem and propose a distributed risk-aware ML solution for RRM. The proposed solution benefits from hybrid orthogonal/non-orthogonal radio resource slicing, and proactively regulates the spectrum needed for satisfying delay/reliability requirement of each URLLC traffic type. A case study is introduced to investigate the potential of the proposed RRM in serving coexisting URLLC traffic types. The results further provide insights on the benefits of leveraging intelligent RRM, e.g. a 75% increase in data rate with respect to the conservative design approach for the scheduled traffic is achieved, while the 99.99% reliability of both scheduled and non-scheduled traffic types is satisfied.
|
|
3. |
- Azari, Amin, et al.
(författare)
-
Serving Non-Scheduled URLLC Traffic: Challenges and Learning-Powered Strategies
-
Annan publikation (övrigt vetenskapligt/konstnärligt)abstract
- Supporting ultra-reliable low-latency communications (URLLC) is a major challenge of 5G wireless networks. Whilst enabling URLLC is essential for realizing many promising 5G applications, the design of communications' solutions for serving such unseen type of traffic with stringent delay and reliability requirements is in its infancy. In prior studies, physical and MAC layer solutions for assuring the end-to-end delay requirement of scheduled URLLC traffic have been investigated. However, there is lack of study on enabling non-scheduled transmission of urgent URLLC traffic, especially in coexistence with the scheduled URLLC traffic. This study at first sheds light into the coexistence design challenges, especially the radio resource management (RRM) problem. It also leverages recent advances in machine learning (ML) to exploit spatial/temporal correlation in user behaviors and use of radio resources, and proposes a distributed risk-aware ML solution for RRM. The proposed solution benefits from hybrid orthogonal/non-orthogonal radio resource slicing, and proactively regulates the spectrum needed for satisfying delay/reliability requirement of each traffic type. A case study is introduced to investigate the potential of the proposed RRM in serving coexisting URLLC traffic types. The results further provide insights on the interplay between the reliabilities of coexisting traffic, uncertainties in users' demands and channel conditions, and amount of required radio resources.
|
|
4. |
- Brunello, Davide, et al.
(författare)
-
Low Latency Low Loss Scalable Throughput in 5G Networks
- 2021
-
Ingår i: 2021 IEEE 93rd vehicular technology conference (VTC2021-spring). - : Institute of Electrical and Electronics Engineers (IEEE).
-
Konferensbidrag (refereegranskat)abstract
- Low Latency Low Loss Scalable Throughput (L4S) is a technology intended to reduce queue delay problems, ensuring low latency to Internet Protocol flows with a high throughput performance. To reach this goal, it relies on Explicit Congestion Notification (ECN), a mechanism that marks packets to signal congestion in the network avoiding packets to be dropped. The congestion signals are managed at the sender and receiver sides thanks to scalable congestion control algorithms. In this paper, the challenges to implement L4S in a 5G network are analyzed. Using a proprietary state-of-the-art network simulator, the L4S marking strategy has been implemented at the Packed Data Convergence Protocol layer. To evaluate the benefits of the implementation, L4S has been adopted to support Augmented Reality (AR) video gaming traffic while using the IETF experimental standard Self-Clocked Rate Adaptation for Multimedia (SCReAM) for the congestion control. The results show that the video gaming traffic experiences lower delay when supported by L4S. Moreover, in all the cases analyzed, L4S provides an average application layer throughput above the minimum requirements of a high-rate latency-critical application, even at high system loads. Furthermore, the packet loss rate has been significantly reduced thanks to L4S. If it is used in a combination with a Delay Based Scheduler (DBS), a packet loss rate very close to zero has been reached.
|
|
5. |
- Cheng, Sijia, et al.
(författare)
-
A Hardware-in-the-Loop Simulator for mmWave Massive MIMO Using PYNQ Framework
- 2023
-
Ingår i: 2023 IEEE 66th International Midwest Symposium on Circuits and Systems, MWSCAS 2023. - : Institute of Electrical and Electronics Engineers (IEEE). - 9798350302103 ; , s. 698-702
-
Konferensbidrag (refereegranskat)abstract
- Massive multiple-input multiple-output (MIMO) plays an important role in beyond 5G systems to support high data rate and low latency, especially operating at millimeter-wave frequency bands. In traditional software-based simulators, the simulation time increases exponentially with the number of base station antennas and served users. We propose a hardware-in-loop simulator that combines the flexible software and fast and accurate hardware with Python Productivity for Zynq (PYNQ) framework. The simulation of massive MIMO detection is accelerated by 4 orders of magnitudes with bit-accurate result. With the help of the developed simulator, we explore the design trade-offs between system performance and hardware cost for in-flight-cabin use cases.
|
|
6. |
- Cheng, Sijia, et al.
(författare)
-
MmWave Massive MIMO Processing in Demanding Environments - An Aircraft Cabin Deployment Study
- 2023
-
Ingår i: Conference Record of the 57th Asilomar Conference on Signals, Systems and Computers, ACSSC 2023. - 1058-6393. - 9798350325744 ; , s. 1631-1635
-
Konferensbidrag (refereegranskat)abstract
- Massive multiple-input multiple-output (MIMO) technology will continue playing an important role in beyond 5G wireless systems. Fully digital beamforming at the millimeter wave (mmWave) spectrum enables applications that require high data rates and low latency. This paper investigates the development and deployment of mmWave massive MIMO systems in an in-flight cabin environment, from both system performance and hardware implementation aspects. Ray-tracing methods with realistic geometry coefficients captured from commercial airplane models are used to represent this in-door dense environment. A hardware-in-the-loop simulator running on the AMD/Xilinx PYNQ FPGA is leveraged to conduct fast design space exploration. The outcomes of our investigation provide insights and guidelines in strategies for implementing and deploying mmWave massive MIMO systems in-flight cabins, for instance, how to place and distribute access points to enhance coverage and serve quality and the appropriate quantization strategy balancing hardware cost and signal processing quality.
|
|
7. |
- Garcia, A. E., et al.
(författare)
-
Direct air to ground communications for flying vehicles : Measurement and scaling study for 5g
- 2019
-
Ingår i: IEEE 5G World Forum, 5GWF 2019 - Conference Proceedings. - : IEEE. - 9781728136271 ; , s. 310-315
-
Konferensbidrag (refereegranskat)abstract
- Broadband connectivity in the sky for unmanned aerial vehicles (UAVs) and commercial aircraft is the next step toward the ambitious goal of connectivity anywhere. To achieve it, direct air-to-ground communications (DA2GC) is a more promising solution than satellite communications due to its lower latency and higher throughput capabilities. In this paper, we develop a DA2GC system with tilted up directional antennas at ground base station and perform a measurement campaign using a UAV at low altitudes to study the DA2GC link at a carrier frequency of 3.7 GHz in terms of signal-to-noise ratio and throughput. Our designed DA2GC system is a general setup that can enable new frequency spectrum and new applications for 5G implementation. We also scale the obtained results from the measurements to a scenario of an aircraft flying at an altitude of 13 km. We study the physical limits of current cellular networks and analyze link budget for the scaling. A maximum of 14 Mbps is achieved during the measurement campaign. By scaling the measurement results with existing communications techniques such as carrier aggregation and techniques toward 5G such as higher antenna gains due to beamforming, extrapolated throughput could reach up to 257 Mbps with a maximum inter-site distance of 200 km due to limitations of current cellular networks. These results show that cellular networks with communication techniques toward 5G can be utilized for providing high-capacity connectivity for current aircraft at high altitudes as well as next-generation flying vehicles at low altitudes.
|
|
8. |
- Lee, Woong-Hee, et al.
(författare)
-
Geometric Sequential Learning Dynamics
- 2021
-
Ingår i: IEEE Communications Letters. - : IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC. - 1089-7798 .- 1558-2558. ; 25:2, s. 542-545
-
Tidskriftsartikel (refereegranskat)abstract
- In this letter, we introduce a novel dynamic model for predicting the exact strategies of the opponents without message exchange, namely geometric sequential learning dynamics (GSLD). The intuition is twofold; first, the utility function is widely modeled by arbitrary exponential varieties; second, the equidistant sampled exponential function comprises a geometric sequence. To validate GSLD, we model the exponential variety game (EVG) and prove its convergence by showing that it is a continuous quasi-concave game. The proposed scheme enables the construction of the exact individual utility function, which results in a faster convergence and a high utility value.
|
|
9. |
- Lee, Woong-Hee, et al.
(författare)
-
Noise Learning-Based Denoising Autoencoder
- 2021
-
Ingår i: IEEE Communications Letters. - : IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC. - 1089-7798 .- 1558-2558. ; 25:9, s. 2983-2987
-
Tidskriftsartikel (refereegranskat)abstract
- This letter introduces a new denoiser that modifies the structure of denoising autoencoder (DAE), namely noise learning based DAE (nlDAE). The proposed nlDAE learns the noise of the input data. Then, the denoising is performed by subtracting the regenerated noise from the noisy input. Hence, nlDAE is more effective than DAE when the noise is simpler to regenerate than the original data. To validate the performance of nlDAE, we provide three case studies: signal restoration, symbol demodulation, and precise localization. Numerical results suggest that nlDAE requires smaller latent space dimension and smaller training dataset compared to DAE.
|
|
10. |
- Li, Wenhao, et al.
(författare)
-
Cost aware service selection in a mobile edge marketplace
- 2022
-
Ingår i: Computer Networks. - : Elsevier BV. - 1389-1286 .- 1872-7069. ; 205
-
Tidskriftsartikel (refereegranskat)abstract
- A marketplace plays an important role that bridges between Mobile Edge Infrastructure Services (EISs) providers and their customers, also manages relations between actors in a mobile edge ecosystem. One of the key services of the marketplace is service selection where not only a list of EISs matching the customers' demands is provided but also enables service selection based on the customers' requirements to fully automate the process. In this paper, we first investigate important attributes of EISs (such as coverage area, latency models, pricing models, etc.), and requirements of edge-based applications (such as latency and reliability) for EIS selection in a marketplace. We then formulate an optimization problem to choose the right set of EISs among available services in the marketplace. We propose two service selection algorithms, i.e., Best Fit (BF) and an Improved version of BF (IBF) to minimize the monetary cost of selected services subject to latency, reliability constraints and customer requirements. The evaluation shows that IBF has 4% improvement in monetary cost as compared to the BF. IBF has only 1% deviation from the optimal solution generated by a brute force algorithm, while it is 189 times faster than the brute force. Accordingly, IBF not only outperforms BF in terms of monetary costs but also achieves the optimal solution as compared to the brute force algorithm in significantly lower execution time. Furthermore, IBM CPLEX Optimizer is also implemented to solve the considered problem to have more concrete evaluation. The results indicate that although CPLEX can also solve the problem with the optimal result, its computing time is still dramatically worse than IBF.
|
|