Results and Future Work
The robotic system successfully recognizes the gestures and moves accordingly.
Approach 1 is faster but lacks precision ( 1.98 seconds).
Approach 2 is slower but more accurate (2.06 seconds), it takes 3 seconds to read the gesture. It can be used for recognizing complex gestures also.
Also, it is preferable to use Approach 1 generally as it is faster, but when there is some Approach 1 gesture not getting recognized, we can go for Approach 2.
There are some issues with the caster wheels, they turn a bit while moving in a straight line.
Also I thought of one good use of my application – Gesture Controlled Wheel Chair [5].
Another application of this project could its usage for Mapping part of SLAM( Simultaneous localization and mapping ), wherein the robot can initially used for mapping purposes by making it move around the room so as to recognize objects.
Future work
- Optimize code for performance
- Shift processing on board (try to connect wiimote directly to LEGO MINDSTORMS NXT Brick , so as to implement approach 1 on the brick itself ).
- Improve the mechanical design
