4:20 PM - 4:40 PM
[1L3-J-11-04] Self-location by a 360-degree camera mounted on a flying-robot
Keywords:Flying Robot, AI, Deep leaning
In this research, we try to derive the three-dimensional coordinates and orientation of a flying robot equipped with a 360-degree camera.
There are various means such as GPS, direction sensor, etc. as means for knowing the coordinates and orientation.
However, it is impossible to know exact orientation and coordinates in places where indoors or magnetic fields are unstable.
In this research, we aimed to acquire the direction and coordinates of the flying robot using a 360 degree camera.
As a concrete method, we installed two color objects at the place we wanted to be a reference, aimed at achieving the purpose by acquiring the relative coordinates of the color object and the flying robot.
Currently the average of the errors of x, y and z coordinates is less than 400 mm, and the average of the error of the direction is about 30 degrees.
There are various means such as GPS, direction sensor, etc. as means for knowing the coordinates and orientation.
However, it is impossible to know exact orientation and coordinates in places where indoors or magnetic fields are unstable.
In this research, we aimed to acquire the direction and coordinates of the flying robot using a 360 degree camera.
As a concrete method, we installed two color objects at the place we wanted to be a reference, aimed at achieving the purpose by acquiring the relative coordinates of the color object and the flying robot.
Currently the average of the errors of x, y and z coordinates is less than 400 mm, and the average of the error of the direction is about 30 degrees.