Browsing by Author "A A M Ismail"
  • Publication
    Prediction model for spectroscopy using python programming
    ( 2021-12)
    A A M Ismail
    ;
    Norshamsuri Ali @ Hasim
    ;
    Mohd Shafiq Amirul
    ;
    Rosdisham Endut
    ;
    Syed Alwee Aljunid Syed Junid
    This paper is motivated by searching for the perfect pattern for the spectroscopy spectra using artificial neural networks (ANN) using python programming coding. The pattern from the spectroscopy is based on the absorption and emission of light and other radiation by materials in relation to the wavelength dependence of these processes. Spectral equipment such as spectrometers, spectral analysers, spectrographs, or spectrophotometers is utilised to determine spectrum values. The problem in this spectroscopy is to identify the sample or analyte, which can be solved by a prediction model for spectroscopy using Python. These problems occur when finding the best algorithm of pre-processing techniques that can predict any model accurately into an understandable format for prediction models. Various types of pre-processing techniques have been used, such as Multiplicative Scatter Correction (MSC), Inverse MSC, Extended MSC (EMSC), Extended Inverse MSC, de-trending, Standard Normal Variate (SNV) and normalisation in order to get a better r2 value. In this project, we find the r2 and the root mean square error (RMSE) to evaluate the prediction values and the actual values. First, choosing pre-processing techniques and then finding the best statistical method for constructing predictive models that produce high accuracy. We used ANN in this project as a prediction model. Based on the results, we managed to achieve our objective, which is that the prediction model has more than 90% of accuracy. Furthermore, the results show that our prediction model has 1.0 accuracy at 100 Epoch with a 0.3 learning rate. Finally, we can conclude that our prediction model can be used to predict the spectroscopy-based data format.
  • Publication
    Two dimensional (2D) OCDMA encoder-decoder for various industrial application
    ( 2021-12)
    Syed Alwee Aljunid Syed Junid
    ;
    Mohd Arif Anuar Mohd Salleh
    ;
    A A M Ismail
    ;
    ND Keraf
    ;
    R Matem
    ;
    Mohd Rashidi Che Beson
    ;
    Norshamsuri Ali @ Hasim
    ;
    Mohd Shafiq Amirul
    ;
    Rosdisham Endut
    ;
    Nor Azura Malini Ahmad Hambali
    Future telecommunication systems and networks are expected to provide a variety of integrated broadband services to the customers. There has been a tremendous interest in applying Code Division Multiple Access (CDMA) techniques to fiber optic communication systems. This technique is one of the multiple access schemes that is becoming popular because of the flexibility in the allocation of channels, ability to operate asynchronously, enhanced privacy and increased capacity in bursty networks. This project is focusing on designing 2D OCDMA system with the hardware implementation of design using FPGA. The coding techniques in OCDMA are time versus wavelength and amplitude versus phase. 2D OCDMA coding incorporates both wavelength selection and time distribution. The data bit would be encoded as consecutive chips with various wavelengths. The code architecture seeks to produce codes with high autocorrelation and low cross-correlation properties. Code length is an essential aspect of code and device architecture for coding characteristics. The hardware implementation of the system is designed by using FPGA De1-SoC. The FPGA have the abilities to enhance the transmission of data to the receiver in a short period of time. The performance of 2D OCDMA system is expected to surpass 1D OCDMA system in terms of BER and the number of simultaneous users that can be supported. The system encoder and decoder were designed using optical switch, splitter, combiner and modulator. The performance of 2D OCDMA system in terms of time spreading and wavelength spreading is also compared which shown a huge difference in the results. The best performance of the system is when the number of wavelength (M) is fixed to 18 and the number of times spreading (N) is 31 which can support 350 more simultaneous users.
      1  4