Deploying AI To Achieve Seamless Data, Voice Connection In Nigeria

Deploying AI To Achieve Seamless Data, Voice Connection In Nigeria

The advent of the Global System for Mobile communications (GSM) in Nigeria has revolutionized the country‘s mobile communication sector. Consequently, the GSM telecommunication industry has recorded tremendous growth since its debut in Nigeria.

Needless to say that mobile communication has become a powerful tool for liberating Nigerians from the shackles of the once powerful, but now extinct, national telecoms monopoly, popularly known as the Nigerian Telecommunication (NITEL); it has significantly contributed positively in boosting economic activities in all sectors in Nigeria.

The technology provided by the currently existing GSM service providers in Nigeria ranges from the progress in the network, from 2G, 3G, 4G to 5G. In addition, because mobile communication has become essential for easing communication, our country has over the years, been flooded with different grades of mobile phones from different manufacturers, leading to poor QoS.

According to the head, Department of Mechatronics Engineering FUT Minna and coordinator, Advanced Engineering Innovation Research Group, Professor Abiodun Musa Aibinu, “Not quite long ago, we came to the knowledge during our research that, despite the obvious benefits of mobile cellular communication, the increasing poor QoS experienced by subscribers do manifest itself in the forms of: Call failure, call drops, and low received signal strength. These problems have continuously been raising a great deal of worries and uncertainty among the populace and the network providers.”

In making efforts to solve the problem of poor QoS coupled with the high demand for quality and efficient on-the-go communication capabilities, MNOs have resulted into deployment of mobile cellular network stations in form of Base Transceiver Station (BTS).

But the increase in the number of BTS has not resolved the problems, Aibinu stated, adding that. “On the other hand, as subscribers, we have resulted to the use of mobile phones with multiple SIM slots in providing temporary solution to these non-abating problems. Some of us even have more than one mobile phone with the hope of enjoying good QoS using the appropriate SIM or phone where necessary. In some instances, “porting” of mobile numbers is suggested. “Porting” as used in the community simply means retaining the mobile identity number but changing the network or service provider.


Way Forward:

A team of researchers from the Federal University of Technology Minna has developed a Multiple Operator-Enabled SIM (MOES) Card, to tackle the issue of poor QoS in Nigeria. The team led by Aibinu, made of 12 researchers, eight research fellows, and four students, started the research work in 2013.

Titled: “Seamless Data and Voice Connection Using Multiple Operators Enabled SIM (MOES) card”, and funded by Nigerian Communications Commission (NCC), with the sum of N2,961,561.15, the aim of the project is to develop an Artificial Intelligence (AI) based portable device, in form of mobile phones, capable of providing seamless data and voice connections over the existing Mobile Network Operators (MNOs), which is envisioned as the solution to the QoS-induced problems which are experienced by mobile communication subscribers in Nigeria.

According to Aibinu, the proposed device is expected to provide subscribers with the privilege of enjoying good services from available operators without the added expenses of having more than one phone or a phone with multiple SIM slots.

“Furthermore, the proposed cognitive phone is expected to guarantee all time access to networks with the best available signal strength and QoS, even while on transit on Nigerian highways and roads. The seamless data and voice connection involves the development and fabrication of a miniaturised embedded system integrating the existing mobile communication SIM card information and mobile communication hardware with artificial intelligent based MNOs handover software applications,” he explained.

Aibinu disclosed that the project kicked off in 2013, with the development of Subscriber Identity Module (SIM) card information extraction software and switching interface system. “The work started from ground zero in 2013 with a team of about 12 researchers comprising eight research fellows and four students. Our objectives included: finding a method of extracting appropriate information from SIM card, determining method of categorizing SIM card parameters, and determining relevant pin configuration for SIM Card switching.

He explained that SIM is a logical module that runs on an Integrated Circuit Card (ICC) type of smart card called Universal Integrated Circuit Card (UICC), adding that “The UICC and the logical application/module running it are commonly referred to as a SIM card. The main purpose of the SIM card is to provide a compact and secure storage of the components required for the GSM/UMTS authentication scheme. This is usually achieved by the authentication and key generation algorithms which make use of the International Mobile Subscriber Identity (IMSI) and the Subscriber Authentication key (Ki).

The next step was the construction of testbeds and SIM cards with programmable Micro-controller having various Graphical User Interfaces, with several months of test running it, Aibinu stated.

“From our findings, we observed that information contained in a typical SIM card can be categorised into two broad categories namely MNO dependent and MNO independent parameters. Only few of these parameters in the MNO independent cluster are appropriate for MOES work. We also observed that irrespective of the SIM card type and size, the pins‘ configuration in a SIM always remains the same; however their functions may differ.

“From testbed experiments conducted, it was identified that switching from one SIM card to another can be best achieved by switching the clock pin of the SIM card. In addition to received signal strength, Channel availability and GPS coordinate can be used for handover criteria in the proposed cognitive phone. This forms the basis of our subsequent research and endeavour in this field,” he explained.

After achieving this, the researchers then focused on developing hybrid AI system for handover in MOES system and cognitive phone Handover (also known as Handoff). This process refers to the process of transferring the point of attachment of mobile phone to the network, from a BTS to another BTS.

Explaining further, Aibinu stated that this occurs as the mobile station moves from the region of coverage of the initial BTS to the coverage region of the target BTS. “The process is expected to be seamless, thus ensuring that any ongoing mobile activity is not dropped, and we the users do not experience poor QoS,” he asserted.

In developing this algorithm, the researchers made use of their knowledge of parametric signal modeling. “We adopted the use of a two-stage approach with the first stage involving the development of mathematical steps, which in turn involved the k _ step ahead ANN based prediction algorithm. The second stage involved the development of Fuzzy logic based Handover decision making algorithm.

“The k _ step ahead ANN based RSS prediction model coefficients are estimated from the synaptic weights. They are also estimated with the coefficients of the adaptive activation function of a properly trained two-layer ANN. The extracted coefficients were used to form a matrix from which appropriate prediction level can be easily obtained via simple matrix multiplication.”

Since 2015 till date, the researchers have been fabricating cognitive mobile phone and evaluating its performance.

“Since 2015, we have been working tirelessly in fabricating laboratory scale model of this work. We produced our first MOES card in 2015 and evolving over time, we are presently in version 4fourof the work. In addition, we have applied the concepts from this work successfully to the design and development of Emergency Phone booth system and Tracking system. I am happy to inform us all that the laboratory models of these two designs have been produced and are presently being optimised for production of market ready prototype.”

Speaking on the next step, Aibinu said with these findings, the next step is to complete the patent filing of this work, while commending NCC for funding the project which has received accolades from far and wide. “For instance, in 2017, we exhibited this research work at the Federal University of Technology, Minna 26 Convocation Research and Development (R&D) Exhibition in 2017 and were given the best research award. I am also pleased to inform us that till date, the work has been used to train over 16 students in Nigeria at both undergraduate and post graduate levels,” he stated.

Be the first to comment

Leave a Reply

Your email address will not be published.