M B

Udacity

Autonomous Systems, Self-Driving Car engineer - Nanodegree•  October 2018 - April 2019

The program comprised of the following projects: - Computer Vision - Designed a data pipeline that used CV libraries such as OpenCV and images from car's front mounted camera to detect lane lines, road curvature, car's position in lane and paint a projected path onto the lane as the car travels. - Deep Learning - Designed a DNN that used Keras and TensorFlow to train a car in a simulation, based on only images from its front mounted camera, to drive around a track. - Sensor Fusion - Designed an object tracker that used a Kalman filter to processed both Lidar and Radar data from car sensors to track a bicycle as it moved around the car. - Localization - Implemented a 2-D Particle filter to localize a car, in a simulation using sparse land marks, as it moved within a mapped space. - Path Planning- Generated jerk minimizing trajectories to facilitate efficient lane changes for a car on a three lane track with other cars, in a simulation. - Control- Implemented PID controllers to automate the controls of a car (throttle, steering angle, brake) that successfully navigated a track within a simulation. - System Integration- Programmed a real Self-Driving car using ROS. Implemented modules of the Perception, Planning and Control sub-systems for the car to successfully navigate a closed track with traffic lights.