Robot planning. Navigation
Head: Enrique J. Bernabeu Soler
The main objective of this research is to innovate, implement and devise a navigation system for an mobile robot. The robot is able to move automatically in both indoors and outdoors environments. Collision are avoided as they are predicted. A planning algorithm generates a set of collision-free trajectories for the mobile as frequently as new information from the sensor is received. A trajectory verifying the robot kinodynamics constraints is selected.
In order to obtain such a objective some problems have to be previously solved:
- Data fusion. Information from sensors, specially a laser and a camera, is processed in order to get a description of the world “in sight”
- Actuation of the acceleration and steering systems.
- Collision prediction and avoidance with the obstacles in sight. Obstacles are considered both static and dynamic.
- Generation of several collision-free maneuver (steering angle and acceleration) for the vehicle.