Page 221 - Kaleidoscope Academic Conference Proceedings 2021

Session 6: Machine learning for next generation wireless network

S6.1 Wireless channel scenario recognition based on neural networks
Xiaojing Xu, Ruimei Li, Hua Rui, Wei Lin and Xiangfeng Liu (State Key Laboratory of Mobile
Network and Mobile Multimedia Technology, ZTE Corporation, China); Wei Cao (ZTE
Corporation, China)

Wireless channel scenario recognition plays a key role in improving the performance of mobile
communication systems. This paper combines wireless channel characteristics extracted using
expert experience and neural networks, and proposes a wireless channel scenario recognition
framework based on neural networks. Firstly, the wireless propagation environment is analyzed,
and some wireless channel characteristics are extracted, such as the frequency domain fading
factor, multipath power delay distribution, time domain energy peak response ratio and time
correlation characteristics. Secondly, the combined algorithm model using the wireless channel
characteristics and neural networks are proposed. Finally, after simulation verification, the new
method has a greater improvement in the recognition accuracy than the traditional threshold
algorithm.
S6.2 A review of network slicing in 5G and beyond: Intelligent approaches and challenges*
Ghazal Rahmanian, Hadi Shahriar Shahhoseini and Amir Hossein Jafari Pozveh (Iran University
of Science and Technology, Iran)

Artificial intelligence has the ability to provide simple solutions to complex problems by analyzing
huge volumes of data in a short amount of time with appropriate accuracy. With the emergence of
slicing technology, Fifth Generation communication networks(5G) are becoming more complex
due to supporting a large number of new connected devices and new types of services. Concerning
this complexity, intelligent techniques can become beneficial in these networks. In this paper after
a brief review of network slicing through a new functionality model including design, deployment,
monitoring and management of slices, the need for automation of network operations in network
slices is discussed. Our investigation of recent research has shown that artificial intelligence and
machine learning have a clear potential to become one of the most important enablers in the 5G
network and beyond.

S6.3 Reinforcement learning for scheduling and MIMO beam selection using CAVIAR simulations*
João Paulo Tavares Borges, Ailton Pinto de Oliveira, Felipe Henrique Bastos e Bastos, Daniel
Takashi Né do Nascimento Suzuki and Emerson Santos de Oliveira, Jr. (Universidade Federal do
Pará, Brazil); Lucas Matni Bezerra (Universidade Estácio de Sá, Brazil); Cleverson Veloso
Nahum (Universidade Federal do Pará, Brazil); Pedro dos Santos Batista (Ericsson Research,
Sweden); Aldebaro Barreto da Rocha Klautau, Jr. (Universidade Federal do Para, Brazil)

This paper describes a framework for research on Reinforcement Learning (RL) applied to
scheduling and MIMO beam selection. This framework consists of asking the RL agent to schedule
a user and then choose the index of a beamforming codebook to serve it. A key aspect of this
problem is that the simulation of the communication system and the artificial intelligence engine
is based on a virtual world created with AirSim and the Unreal Engine. These components enable
the so-called CAVIAR methodology, which leads to highly realistic 3D scenarios. This paper
describes the communication and RL modeling adopted in the framework and also presents
statistics concerning the implemented RL environment, such as data traffic, as well as results for
three baseline systems.

– 159 –