


Our edge
detection algorithm consists of searching in 20 small windows. The windows
are opened along the predicted edge positions which were given by dynamical
model. In each window we find a most likely edge and finally linearfit these
data to get the whole stripes.


GOAL:
Develop a navigation system for the robot to recognize and walk along the
building corridors.


•Analyzed the geometric relation of the system,
developed the algorithm to estimate robot's distance and angle position.


•Implemented edge detection algorithm for image
information processing.


•Designed the motion control strategy. The robot
can walk stably along the hallway from arbitrary position in tens of seconds.


•Robot walks at it mechanical speed 0.5m/s in
the corridor.




