Smart Radio Environment based on Machine learning Frame work.

              

In this article we would discuss potential future of wireless technology, where Radio Environment can be made configurable / Smart, which is at present not playing any role.

Future wireless networks are expected to constitute of distributed intelligent wireless communications, sensing, and computing platform and all are going to be interconnected, which is going to be challenging as they will have challenging requirement of interconnecting the physical and digital worlds in a seamless and sustainable manner.

The major problem to get these connected is to have seamless connectivity , which is mostly hampered because of obstacles in signal path , those obstacles are generally buildings in urban areas , also there is another problem of power , as more signal means radio unit needs more power to transmit and receive those signal.

Above issues motivate the use of Smart radio environment , which can be configured in such a way that signal which get  blocked by earlier obstacle / environment , can now be reflected to the area where more subscribers are present , so that those Subscribers can be benefitted with services and no need of additional Base Station is needed. This can be achieved with a potential wireless concept, where the environmental objects are coated with artificial thin films of electromagnetic and reconfigurable material (that are referred to as intelligent reconfigurable meta-surfaces), which are capable of sensing the environment and of applying customized transformations to the radio waves. Smart radio environments have the potential to provide future wireless networks with uninterrupted wireless connectivity, and with the capability of transmitting data without generating new signals but recycling existing radio waves.

In a nutshell, this article is focused on discussing how the availability of intelligent reconfigurable meta-surfaces will allow wireless network operators to redesign common and well-known network which can provide seamless connectivity with less antenna devices, by recycling existing radio waves instead of generating them.

 

Let’s Understand  future of wireless communication:

Wireless Futures - Beyond Communications, but Without More Power and Radio Waves

Future wireless networks should have the potential to be turned into a distributed intelligent communications, sensing, and computing platform. Besides connectivity, more specifically, the platform will be capable of sensing the environment to realize the vision of smart living in smart cities by providing context-awareness capabilities, and of locally storing and processing information. Such processing could accommodate the time critical, ultra-reliable, and energy efficient delivery of data, and the accurate localization of people and objects in environments and scenarios where the Global Positioning System (GPS) is not an option. Future wireless networks will have to fulfil the challenging requirement of interconnecting the physical and digital worlds in a seamless and sustainable manner

As in current scenario user traffic is increasing tremendously and to serve user service provider is using different frequency and spectrum, with technology like carrier aggregation etc. , which increases power consumption and emission of radio waves , which is environment.

To limit this this smart radio environment concept can help service provider a lot and also will help service provider to cut the cost as power consumption will go less.

What is Smart Radio Environment ???

In current wireless networks, the radio environment, i.e., the physical objects that alter the propagation of the electromagnetic waves, which is not controllable , and is perceived, in addition, as an adversary to the communication process, i.e., it has usually a negative effect that needs to be counteracted by the transmitters and receivers . By contrast, we define a smart radio environment as a radio environment that is turned into a smart reconfigurable space that plays an active role in transferring and processing information, and that makes more reliable the exchange of data between transmitters and receivers. This can be achieved by with the use of Reconfigurable Meta-Surfaces.

Lets understand what is Meta-Surfaces & Reconfigurable Meta-Surfaces

An example of meta-surface is sketched in Fig. 1, where it is shown that it transforms an incident radio wave into a reflected radio wave and a transmitted (or refracted) radio wave. The specific arrangements of the scattering particles (e.g., full or slotted patches, straight or curved strips, various types of crosses, etc.) determine how the meta-surface transforms the incident wave into arbitrary specified reflected and transmitted radio waves. The major difference between a surface and a meta-surface lies in the capability of the latter of shaping the radio waves.

A reconfigurable meta-surface is a meta-surface in which the scattering particles are not fixed and engineered at the manufacturing phase, but can be modified depending on the stimuli that the meta-surface receives from the external world.

Lets see the example to understand better , in fig.2 which is a conventional wireless network , where smart radio is not present and to serve the user it will be served by BS2 since BS1 signals is blocked by obstacle , which is very far hence the subscriber services will get impacted , but when used Smart radio environment in fig. 2 then we can see that the signal from BS1 is refracted in such a way that the user can be served by BS1 only instead of BS2 , hence his experience will be better.

Now let us understand what we meant by Smart radio environment, since till now there is nothing smart rather than the fact that this meta surface can be configured based on requirement, but the problem to configure / reconfigure such meta surface is very complex.

 

As we know that by leveraging the AI concepts, this kind of complex problem can be solved hence with the help of Artificial intelligence, we can configure this meta-surface in a way that it solves the purpose of seamless coverage with less power and Radio waves.

AI for Smart Radio Environments:

As discussed, smart radio environments are a very complex system to design. This complexity originates from the large number of parameters to be optimized based on the contextual information that is gathered by the intelligent reconfigurable meta-surfaces and that is made available to the network controller. In current era where Machine learning is considered a answer for almost all complex problem, so its better to consider ML in implementing Smart Radio Networks.

We can use Reinforcement Learning Framework, to implement reconfigurable Meta-surface. As we know Reinforcement learning, implements the learning and decision-making procedures by interacting with the environment: Taking actions and receiving feedback on the result of the actions that are taken. By using this approach, we can implement reconfigurable meta-surface.

By following steps.

1)     1)  Interact with the environment through their embedded sensors.

2)      2)  make decisions and take actions, in a distributed way, in order to optimize the wave transformations that they apply to the radio waves

3)     3)  modify the radio waves based on the subsequent response from the environment.

 

Above RL approach seems to be perfect to solve this complex problem and implement smart Radio network, but only problem occurs with RL is that since wireless network behaviour is highly dynamic in nature , to get the model converge will take long time because of exploration and exploitation phenomena of RL model and only using supervised model will not help because they need lot of data to be trained on.

Hence based on some preliminary research done by researcher on this domain, suggest that to use transfer learning approach to solve this complex problem, The approach which researcher have proposed in consists of combining together model-based and data-driven optimization methods.

Let’s understand a bit more how researcher has suggest this approach.

The green box is model based approach where we create a approximate mathematical model of network , which we generate with the help of Deep Neural Network and then feed the output of that model as a input to data driven  model  which is also a neural network , to get the optimal function.

Conclusion and Potential Impact:

Smart radio environments largely expand the concept of network softwarization from the logical domain into the physical domain: The radio environment itself is viewed as a software entity, which can be remotely programmed, configured, and optimized. The concept of smart radio environments is not restricted to enhancing wireless communications, but is aimed at introducing a truly distributed intelligent wireless communications, sensing, and computing platform that interconnects the physical and digital worlds.

By recycling, e.g., the reflections of radio waves and embedding the data of sensors into them at a zero-energy cost, the potential impact of smart radio environments is beyond communications. Imagine a smart radio space where the walls of rooms are coated with sensing meta-surfaces that monitor the health status of people. This will allow us to develop a truly pervasive and preventive e-health system.

It has potential to low the radiation caused by conventional network as service provider need to increase base station and use different frequency to fulfil the raising demand.

the vision of smart radio environments are reality:

• How to integrate the reconfigurable meta-surfaces into wireless networks?

• What are the ultimate performance limits of wireless networks in the presence of reconfigurable meta-surfaces?

 • How to attain such performance limits in practice?.

 

As experiment on this has already been started by NTT DoCoMo and Metawave  who have recently run some experimental tests related to this technology, by using innovative 5G equipment provided by Ericsson and Intel.

Future view of wireless Network with Smart Radio Network in place , where user don’t have to struggle with coverage.

Reference:

1) Smart Radio Environments Empowered by Reconfigurable Intelligent Surfaces: How it Works, State of Research, and Road Ahead Marco Di Renzo, Fellow, IEEE, Alessio Zappone, Senior Member, IEEE, Merouane Debbah, Fellow, IEEE, Mohamed-Slim Alouini, Fellow, IEEE, Chau Yuen, Senior Member, IEEE, Julien de Rosny, and Sergei Tretyakov, Fellow, IEEE

2) A. Zappone, M. Di Renzo, F. Shams, X. Qian, and M. Debbah, “Overhead-aware design of reconfigurable intelligent surfaces in smart radio environments,” arXiv, Mar. 2020. [Online]. Available: https://arxiv.org/abs/2003.02538

3) D. Dardari, “Communicating with large intelligent surfaces: Fundamental limits and models,” arXiv, Dec. 2019. [Online]. Available: https://arxiv.org/abs/1912.01719.

4) NTT DOCOMO, “DOCOMO conducts world’s first successful trial of transparent dynamic metasurface,” Jan. 2020. [Online]. Available: https://www.nttdocomo.co.jp/english/info/media center/pr/ 2020/0117 00.html