Autonomous Mobile Robotics

Autonomous Mobile Robotics
(ME 597)

Topics

Sensor modeling and vehicle state estimation with Bayes, Kalman, and Particle Filters
Onboard localization and mapping, motion modeling and control with ROS, Gazebo, and MATLAB simulations
SLAM, RANSAC, as well as Trajectory Rollout, Wavefront, A*, and probabilistic roadmaps for pathfinding.
Labs using Turtlebots for occupancy grid mapping, particle filters, and RRT exploration and navigation.

Labs

The robot used was a Turtlebot with an attached Kinect for laser scans. Visualizations are from rviz. The code was first simulated in the Gazebo environment, then tried on the actual robot.

Occupancy Grid Mapping Lab

Using the Kinect laser scans and known position data, an occupancy grid map is generated as the robot drives around the environment.

Gazebo Simulation

Physical Environment

Rapidly Exploring Random Tree (RRT) Lab

Using RRTs, a path is planned between two points on the map. The robot is then controlled to follow the planned path.

Gazebo Simulation

Physical Environment

24 Hour Final Exam

Independently written exam, with code and report to be submitted within 24 hours, received a 97%.

MATLAB Simulation

Question 2 from exam, navigation of an aerial racer around 3 waypoints in a forest. Inputs for the racer are accelerations in the x and y directions in the robot body frame. An occupancy grid is dynamically built using 4 sensors with a 30 degree field of view mounted at the front of the racer. Trajectory rollout is used to select the best set of inputs while avoiding obstacles.