YUCHONG YE

Build a 3D perception pipeline using multiple filtering and segmentation techniques that recognize objects in RGB-D images.

Train a convolutional neural net to drive a car in the simulator.

Built a pipeline using distortion correction, image rectification, color transformation and gradient thresholding to identify lane lines and their curvature.

Detect and track vehicles using color and gradient feature and a support vector machine classifier.

Model Predictive Control to drive the car around the track even with additional latency between commands.

Implemented a particle filter that estimates the car’s position in the simulator based on LIDAR measurements and 2D landmark map when a rough GPS location is given.

Built an extended kalman filter to estimate the state of car in a 2D Simulator with noisy lidar and radar measurements. 

Control a KUKA KR210 Arm with six degrees of freedom to perform pick and place actions using inverse kinematics.