iman sharifi's Projects
PID, LQR, Feedback Linearization, Backstepping, Sliding Mode, and Model Reference Adaptive Control for 6-DoF Robot Control
Control Quadcopter using PID Controller in Vrep CoppeliaSim Software with Python
Lets practice some fuzzy examples according to Wong's Book, including Gradient Descend Algorithm, Recursive Least Square, ...
SIFT, SURF, and ORB detector to extract the keypoints like corners. Fundumental Matrix and Homographical Transformation for Panorama Images.
We use Convelutional Neural Networks (CNN), as a Deep Learning paradigm, to detect the house numbers using the MNIST dataset.
System Identification and Self-Tuning PID Control using NN
Extract features from the Stanford Dataset using Histogram of Oriented Gradients (HOG) and use Supported Vector Machine (SVM) to map the features to the assigned labels. Next, use Non-Maximum Suppression and Heatmap method to find the best bounding box on the face.
Config files for my GitHub profile.
This is my personal website.
Extract Symbolic Rules using Background Knowledge and Environement Setting to find the State Transition rules in the Maze gridworld.
Modeling a Mobile Robot using Advanced Dynamics Methods, Design & Simulation using Matlab and SolidWorks, Control using PID Controller, and Implemetation using Arduino UNO, Kalman Filter, and DC Motors..
Offline Tuning PID gains in a given system using Heuristic Methods, including Genetic Algorithm andParticle Swarm Optimization
Use Symbolic Logical Programming to find Safe actions in each state and help Reinforcment Learning to ensure safety in the exploration phase. We use it to make decision in Autonomous Highway Driving.
Novel adaptive tuning of the PID gains using an Actor-Critic-based Neural Network for Attitude Control of a 6-Dof, 4-motor Robot
Extract geveral, symbolic rules in the environment to help Reinforcement Learning find the best, safe policy in the exploration phase. This method is interpretable, explainable, data-efficient, and enjoy symbolic reasoning.
Safe Q-Learning via a Symbolic Logical Programming (Prolog) paradigm in Maze Gridworld.
Tiltrotor Control and System Identification using Fuzzy C-Means Clustering.
Detection, Counting, Tracking, and Summarization of the Vehicles using Computer Vision algorithms in a video camera recorded on a highway.