| Student Name | School | Degree | Year | |
|---|---|---|---|---|
| Muhammad Hanzala Iqbal | BS Electrical Engineering | Undergraduate | Fourth | [email protected] |
| Muhammad Usama | Undergraduate | Fourth | [email protected] |
| Inter School Idea ? | No |
| Do you need expertises from another area: | No |
| If Yes please provide details of expertises you need: |
| Idea Name: | Cellular Traffic Prediction usinG Federated Learning |
| Slogan: | Anticipating the Future with FL |
| Supervisor Name: | Dr. Sajjad Ur Rehman |
| Supervisor Designation: | Head of Electrical Department |
| Supervisor School: | SEECS |
| Supervisor Department: | |
| Contact number: | +92 333 9123390 |
| Email ID: | [email protected] |
| Abstract: | |
| In the modern age of mobile communication, cellular networks are crucial to the success of communication systems supporting a large number of mobile users and devices. With the development of AI (Artificial Intelligence) techniques, cellular traffic prediction is becoming more and more vital for the realization of various applications. Machine Learning based wireless traffic prediction techniqu | |
| What is the unmet need in society that your idea will fulfill ? | |
| In traditional machine learning techniques, the users are not utilizing all resources efficiently and facing privacy issues. New wireless traffic prediction approaches are needed for efficient cellular traffic prediction. | |
| Who needs it ? How many would benefit ? | |
| Telecommunication industries will have the greatest interest in our proposed wireless traffic prediction technique. Because the current traffic prediction techniques are not as much efficient and effective than ours. | |
| How will the solution works | |
| In our proposed system we will implement the federated learning (FL) algorithms to predict the cellular traffic (SMS, calls, and Internet) on base stations. The cloud server will deploy the models in local base stations and those models will be trained by local data. And only the hyper-parameters from this locally trained model will be sent back to the server for further training, making the model of the cloud server a master model. In this way, we will be reducing the load on main server. | |
| Who are your competitors ? How is your solution different | |
| Till now we do not have any competitors as it is researched based project so atleast in Pakistan we do not have any competitors but yes some other universities like King Abdullah University is doing research on it in PHD program. | |
| Status: | new |
| Entry Date & Time: | 2022-12-12 (1703) |