Course-assignments for the course ESE 650 - Learning in Robotics (Spring 2022) that involve State Estimation (Kalman Filter and its variants) and Mapping (Occupany Grid). The theory behind these concepts is described in these blog posts: Deriving Kalman Gain, Filtering Algorithm, EKF, UKF, Particle Filter, Mapping.
The problems are described in more detail in Problems.pdf and the solutions in Solutions.pdf.
The aim of this problem is to use filtering to estimate an unknown parameter of a non-linear dynamical system. Collected a dataset of observations using the ground truth value of the system parameter and developed the EKF equations that uses this dataset to estimate it.
Implemented a quaternion-based 7-DOF UKF for tracking the orientation of a drone in three-dimensions using 6-axis IMU data. The Vicon data is used as ground truth for calibration and tuning of the filter.
Coded a particle filter based SLAM using odometry and lidar data collected on THOR-OP humanoid robot to build a 2D occupancy grid map of indoor environments.
