An important task performed by many robots is detecting, localizing, and tracking objects in the environment. All sorts of robots"”from humanoid robots to autonomous cars and drones"”need to be able to find objects around them and track their location. KOLT delivers a drop-in solution to this problem. A software package build for ROS, KOLT consists of a deep neural network for object detection in RGBD images coupled with a Kalman filter for tracking and filtering of detecting objects. The ultimate goal is to develop a drop-in solution for most vision tasks that roboticists encounter.

• This report represents the work of one or more WPI undergraduate students submitted to the faculty as evidence of completion of a degree requirement. WPI routinely publishes these reports on its website without editorial or peer review.

Creator

• Rosenberg, Nathan Austin

Publisher

• Worcester Polytechnic Institute

Identifier

• E-project-042419-170210

Advisor

• Michalson, William R.

Year

• 2019

Date created

• 2019-04-24

Resource type

• Major Qualifying Project

Major

• Robotics Engineering

Rights statement

• In Copyright