POLARIS - Position Orientation Localization ARTag Recognition Indoor System
The goal of this project is to provide a method for the indoor localization of generic robots. By the end of this project, the team aspires to make the third and fourth floors of George Mason University's Nguyen Engineering Building robot friendly. The motivation for making the third and fourth floors of the Engineering Building robot friendly is to introduce a delivery system. The delivery system will be a generic robot that utilizes the indoor localization module, POLARIS, that we intend to build.
The creation of a robot friendly building requires a non-salient method for reliable indoor localization. Non-salient tags will be placed on the ceilings of the third and fourth floors of the Nguyen Engineering building and will be recognized by the indoor localization module, POLARIS, through the use of computer vision. The non-salient tags will correspond to particular X,Y positions for a given floor, and will enable the robots to localize themselves within that floor. The resultant localization will provide for the ability of the robot to navigate.
The module POLARIS we intend to build will provide indoor localization functionality to a generic robot in the robot friendly building. The project will be executed with the use of two robots. The Pioneer 3-DX, and a generic robot. The Pioneer will be used first to map out the third and fourth floors of the Nguyen Engineering building while simultaneously localizing itself within the map being created through Simultaneous Localization And Mapping (SLAM). The ground truth map data gained from SLAM will be utilized to build a lookup table (LUT) of positions corresponding to recognized glyphs found using a ceiling-facing camera on POLARIS. The LUT will be stored in POLARIS, which will be usable by any generic robot that has a USB 2.0 port and proximity sensors. POLARIS will be connected to the generic robot, which will then have the ability to navigate the floors with the acquired LUT and current glyph recognized by the ceiling-facing camera of POLARIS. The generic robot will be able to retrieve commands from a user to know to which destination to navigate to. POLARIS allows for fulfillment of the requirements for the delivery system requested by faculty for transporting objects in a robot friendly building.
Global Positioning System (GPS) has allowed people to navigate with assurance, it has numerous applications for the military, and it is efficiently used in emergency response. However, GPS has its limits. Indoors, the line of sight of satellite signals tend to be blocked. A system is needed to provide reliable indoor localization so that the most can be made of positioning indoors. An indoor localization module would allow us to increase our coverage of areas where we can localize. The applications of extending the capabilities of localization to indoor areas are numerous. One very important use of this type of system would be for emergency response in an area unsafe to humans. Autonomous robots could navigate utilizing the localization system through the unsafe area and provide relief without risking the lives of humans.
- Requirements of POLARIS and Current Progress
- Accuracy of POLARIS
- Localizing with AprilTags
- Non-Salient AprilTags
- SLAM-assisted creation of a Look-Up Table Overview
- Using the Localization Module
- Manual Creation of a Look-Up Table (LUT)
- SLAM-assisted creation of a Look-Up Table (LUT)
- Using the AprilTags Glyph Recognition System
- POLARIS Serial Communication Protocol
- Creation of special Kernel event handling (USB auto connection for serial communication)
From left to right: Humaira Shah, Gloria Rodriguez, Faculty Supervisor Daniel M. Lofaro, Lucian Mardy, and Jason Fogle.