An improved indoor location system by hybridapplication of multilayer perceptron and radial basis function networks, together with variable region of interest geometry

Adli, Amin Solehuddin (2022) An improved indoor location system by hybridapplication of multilayer perceptron and radial basis function networks, together with variable region of interest geometry. [Project Paper] (Submitted)

Text FS 2022 10.pdf Download (1MB)

Abstract

The development of precise wireless localization algorithms, a critical enabler technology for future Location-Based Services (LBS), is currently generating much buzz. As interest in universal computing and location-aware techniques develops, so does the necessity for a viable and precise solution to localization. In various location-based services (LBS) applications, GPS designs are standard. Due to the weak GPS signals from satellites that cannot penetrate buildings and structures or through the earth, GPS performance in urban canyons is worse than in open areas, mainly in-house and indoor, as well as underground circumstances or surroundings. As a result, GPS is no longer valid for interior navigation. Due to this issue, Indoor Positioning Systems (IPS) has been developed. Due to impediments in the indoor environment, such as walls and partitions. Positional issues are exacerbated by these irregularities, which occur naturally in any indoor space. IPS encompasses a wide range of techniques. This method, however, generates a large amount of RSS sample data. The problem with using a big data sample is that the data is highly nonlinear. Artificial Neural Networks (ANNs) can help solve these problems. Hence, in this study, the work focuses on reducing the error for indoor location systems using an artificial neural network. A total of 576 data were collected for this experiment. This study used a radial basis function network to process the data. The software used in this study is MATLAB. The data consist of 96 points and six signal strengths. Each of the 96 points has six signal strengths.

Item Type:	Project Paper
Faculty:	Fakulti Sains
Depositing User:	Ms. ROHANA ALIAS
Date Deposited:	25 Jun 2024 10:12
Last Modified:	25 Jun 2024 10:12
URI:	http://psaspb.upm.edu.my/id/eprint/1976

Actions (login required)

View Item