Abstract
This paper presents a new environmental model for real time robot navigation, called T-Net (Target Network) . The T-Net, which consists of straight paths, represents rigid environmental structure and keeps relations between the local areas. In this model, the robot memorizes the path with a pair of feature points and can move precisely by tracking these feature points. The most important feature of the T-Net acquisition is that the robot can approximate a local area uniquely with a straight path. We have applied this idea into our mobile robot that has Multiple Vision Agents, and verified it through simulations and experimentation in a real world.
