Distinctive Image Features from Scale-Invariant Keypoints
International Journal of Computer Vision
Navigation in large-scale environments using an augmented model of visual homing
SAB'06 Proceedings of the 9th international conference on From Animals to Animats: simulation of Adaptive Behavior
Three 2D-warping schemes for visual robot navigation
Autonomous Robots
Landmark vectors with quantized distance information for homing navigation
Adaptive Behavior - Animals, Animats, Software Agents, Robots, Adaptive Systems
Analyzing the effect of landmark vectors in homing navigation
Adaptive Behavior - Animals, Animats, Software Agents, Robots, Adaptive Systems
Hi-index | 0.00 |
Several methods can be used for a robot to return to a previously visited position. In our approach we use the average landmark vector method to calculate a homing vector which should point the robot to the destination. This approach was tested in a simulated environment, where panoramic projections of features were used. To evaluate the robustness of the method, several parameters of the simulation were changed such as the length of the walls and the number of features, and also several disturbance factors were added to the simulation such as noise and occlusion. The simulated robot performed really well. Randomly removing 50% of the features resulted in a mean of 85% successful runs. Even adding more than 100% fake features did not have any signi.cant result on the performance.