Navigation is a basic problem in robotics. Many succesful algorithms exist for navigation, but these solutions generally rely on information such as exact coordinates, exact distances, and orientation. I-Bug is an intensity-based bug algorithm which uses none of these and still guarantees that the robot will attain its goal, to reach the tower.
In our first I-Bug paper, accepted to IEEE ICRA 2009, I-Bug is in an unknown 2D environment with possibly infinite but locally finite piecewise-analytic obstacles. The tower emits a signal, and the robot's plan is based on sensing the intensity of this signal. A proof of convergence and bounds are provided for both the case of the level sets being radially symmetric and asymmetric. The final version of the paper, with proofs ommitted due to length constraints, can be found here, and the complete version can be found here. The presentation from ICRA 2009 is available here.