Yasushi Takatori’s research while affiliated with Nippon Telegraph and Telephone and other places

... The authors of [7] utilized visual data obtained from multi-view sensing cameras to predict path loss. The authors of [8] utilized side-view images between the transmitter and receiver, along with frequency information, to predict path loss. The work in [9] employed 2D satellite images of the target area as input to a deep neural network to predict path loss distributions at various UAV altitudes. ...

Reference:

Multi-Modal Environmental Sensing Based Path Loss Prediction for V2I Communications

... 5G is being rapidly deployed around the world and brings its share of new features with even more throughput, reduced latencies opening new possibilities (autonomous cars, industry 4.0, fine control at a great distance...) and an infrastructure that starts to be aware of its environment by adapting its capacities in real time according to the needs or emergencies [24,25]. ...

Reference:

5G AND 6G EVOLUTION

... Furthermore, it should also be considered that in the context of network operators, key features such as radio-related key performance indicators, GPS data, and mobility estimation can also be used for finding out about the network infrastructure as well as user mobility data to improve accuracy. The video player can then leverage information provided by the network leading to higher throughput prediction accuracy as described in Cradio [27]. In this paper, the impact of throughput prediction accuracy on the long-term chunk selection mechanism will be shown by comparing six different baselines with different degrees of accuracy: ...

Reference:

Long-Term Adaptive Bitrate Control Mechanism

... As a multicarrier modulation technique, orthogonal frequency division multiplexing (OFDM) become the core technique for 4G and 5G mobile communication systems due to its advantages such as robustness to inter-symbol interference [2], [3]. OFDM is widely used in many modern communication scenarios such as visible light communications [4] and multiple-input multiple-output communications [5], [6]. One of the major drawbacks of OFDM is its high peak-to-average power ratio (PAPR). ...

Reference:

Design and Optimization of Hierarchical Demodulator and Decoder for Layered Coded OFDM-IM with NR-LDPC

... Finally, as consumed energy directly stems from the training process, different DNN architectures and methods should be leveraged. For instance, fully connected layers are known to be more energyefficient than convolutional layers[15]. Besides, DNN compression methods such as energyaware pruning for F-MADRL should be investigated to guarantee a high accuracy while meeting the restricted power-budget of IoT devices.Another improvement direction is to incorporate energy harvesting techniques to enhance communication sustainability. ...

Reference:

Federated Multi-Agent Deep Reinforcement Learning for Intelligent IoT Wireless Communications

... In the literature, references [8][9][10][11][12][13][14][15][16][17][18][19][20][21] discuss dual connectivity enhancement in cellular networks. Reference [8] introduces a radio resource scheduling method to increase spectrum efficiency and achieve fairness for dual connectivity networks. ...

Reference:

A genetic algorithm based flow control scheme for LTE-NR dual connectivity networks

... The RIS with its elements grouped in blocks assisting multiple users has been formulated, and two heuristics have been proposed to solve them in [7]. The user association strategy with the user mobility for RIS-aided multi-beam transmission systems have been considered in [8], and the ON/OFF selection of RIS in a massive RIS aided wireless communications has been proposed in [9]. The resource allocation of RIS-aided dual connectivity has been presented in [10]. ...

Reference:

Resource Allocation and Sharing Methodologies When Reconfigurable Intelligent Surfaces Meet Multiple Base Stations

... NTT laboratories are researching and developing Cradio ® to solve this complicated problem and achieve a broader range of optimization by linking with various industries and applications other than wireless [4,5]. Cradio ® is shown in Fig. 5, which, as mentioned above, consists of wireless sensing/visualization technology, wireless-network-quality prediction/estimation technology, and wireless-networkdynamic design/control technology. ...

Reference:

Wireless Technologies for Future Network Services Tailored to the Requirements of Various Users and Applications

... This forward link is considered essential to ensure the coherence and high-speed of the technologies needed for the safety and accuracy of these systems and processes. This research is aimed at achieving these stringent communication requirements by enhancing the network performance to support these applications and advance future technologies in these areas [17]. ...

Reference:

Enhancing beyond 5G connectivity and security: optimizing user-to-multiple AP associations with hybrid deep learning and innovative optimization techniques

... In [16], [17], the authors extended the approach presented in [15] and showed that in addition to satellite, images or vector data obtained from OpenStreetMap, a hybrid approach incorporating a simple path loss model helps the NN with learning and offer improved predictive performance and generalisation as compared to pure data-driven approaches (i.e., without the inclusion of path loss models). The work presented in [18] and [19] also employs a combination of CNN and FCNN for path loss predictions whilst [20] integrates both images and environment related parameters as input features. In [21], the work presented in [22], [23] is extended and a hybrid approach is proposed to improve the path loss estimation accuracy. ...

Reference:

Application of Machine Learning for Radiowave Propagation Modeling Below 6 GHz