Faculty Mentor Name
Mehran Andalibi
Location
Hazy 132
Start Date
5-7-2016 5:50 PM
End Date
5-7-2016 6:00 PM
Abstract
The process of simultaneously localizing a mobile robot in an unknown environment while building a map of the environment, known as Simultaneous Localization and Mapping (SLAM), has been a central research topic in robotics in recent years. SLAM in dynamic and complex environments remains an open research problem. If properly designed, a team of robots could significantly increase the speed of mapping and make the system more robust since overlapping information can be used to verify proper functionality of the individual robots and the failure of a single robot does not hinder the overall mission. The purpose of this project is to implement SLAM in a dynamic indoor environment using multiple ground robots. This task requires the combination of inertial and visual sensors as well as active range-finders such as a Kinect Infrared Sensor. The first stage of this project focused on gaining a familiarity with the robotic platforms and software that were to be used by developing methods of data acquisition, sensor fusion, and map building. Also, a camera was implemented in order to detect moving objects and remove them from the map. Future steps of the project include combining local maps from single robots into a global map and gaining a familiarity with localization given the environment map, depth information, and on-board sensor measurements ultimately leading to the implementation of cooperative SLAM.
Simultaneous Localization and Mapping by Cooperative Robots
Hazy 132
The process of simultaneously localizing a mobile robot in an unknown environment while building a map of the environment, known as Simultaneous Localization and Mapping (SLAM), has been a central research topic in robotics in recent years. SLAM in dynamic and complex environments remains an open research problem. If properly designed, a team of robots could significantly increase the speed of mapping and make the system more robust since overlapping information can be used to verify proper functionality of the individual robots and the failure of a single robot does not hinder the overall mission. The purpose of this project is to implement SLAM in a dynamic indoor environment using multiple ground robots. This task requires the combination of inertial and visual sensors as well as active range-finders such as a Kinect Infrared Sensor. The first stage of this project focused on gaining a familiarity with the robotic platforms and software that were to be used by developing methods of data acquisition, sensor fusion, and map building. Also, a camera was implemented in order to detect moving objects and remove them from the map. Future steps of the project include combining local maps from single robots into a global map and gaining a familiarity with localization given the environment map, depth information, and on-board sensor measurements ultimately leading to the implementation of cooperative SLAM.