Browsing by Author "Usama Masood"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Intelligent load and power sources management using wireless sensor network
    (University of Management and Technology, 2014) Usama Masood; Asaad Masood; Haseeb Mushtaq; Muneeb Kayani
    Smart Homes, it seems like future but it’s not. It’s already prevailing in the western world. The idea of our homes becoming smart is something everybody aspires. Homes that can automatically sense human presence and switches utilities accordingly, note temperature of the room and set thermostat of air conditioner automatically, use power sources such as solar power at times when solar intensity is at its peak and also save this power for backup purposes in times of load-shedding. All this seems like future technology but it’s not, our project shows a ray of light to ordinary citizens who are not familiar with latest research in electrical engineering. Our project is product-oriented not research-oriented. It basically focuses on the needs of customer in our country Pakistan which is suffering from major energy crisis, where load-shedding is at its peak, where our energy consumption far exceeds our energy production. So our project focuses on this problem and try to solve it to some extent. In this project we have two power sources, AC mains and solar power. It senses solar power intensity and utilizes it smartly by charging the batteries. As high backup energy is required in our country and not every user can purchase costly solar panels so if low power solar panels are used batteries can also be charged through AC mains. This is the power sources management and it is based on smart algorithm. In times of load-shedding power automatically switches from mains to backup utilizing stored solar power hence saving electricity bills smartly. Our project also consists of a wireless sensor network based on intelligent algorithm for detecting human presence in all areas of a room for automatic switching of appliances. This cuts off major misuse of energy.
  • Thumbnail Image
    Item
    Prediction of autism spectrum disorder using DC-CNN based LSTM
    (UMT Lahore, 2022-06) Rana Muhammad Awais; Usama Masood
    Autism Spectrum disorder (ASD) is a psychological disorder that affects peoples of all ages. Machine learning is used to diagnose ASD. Machine learning based prediction models were built on chronological dataset of ASD and the local patterns are extracted from the features to detect the ASD. A novel dilated causal convolutional neural network based long-short term memory model is proposed to improve the accuracy and precision for the prediction of ASD. Dilate causal convolutional neural network (DC-CNN) is used to extract the crucial features and long-short term memory (LSTM) is used to predict the ASD. The ASD screening dataset is used in this study to analyze and forecast likely cases in adults and children. The different evaluation parameters such as F1-score, precision, Recall are used to validate the efficiency of proposed model with other machine learning models such as KNN, SVC, Artificial Neural Networks, Random Forest, Logistic Regression and Decision Tree.