Localización y seguimiento de trayectorias con robots caminantes en entornos naturales

Abstract

This thesis is based on the design and experimentation of algorithms for the efficient detection and localization of antipersonnel landmines with legged robots. This task must be made on natural environment (irregular terrain, with different hardness and in the presence of obstacles) and in which forbidden areas can exist. The feet of the robot must avoid these forbidden areas (potential alarms or antipersonnel landmines previously detected by other system). The development of this objective requires to divide it into different independent sub-objectives: (a) gait: planning of the movements of the legs of the robot to move the robot in the desired direction maintaining the stability and adaptation to the terrain takes place, and considering the existence of forbidden areas; (b) location: estimation of the robots position in the environment during the task; and finally, (c) path tracking: definition of the robots trajectory to guarantee the complete exploration of the terrain and generation of algorithms to control the crab angle.The independent solution of these three sub-objectives constitutes the main body of this thesis. Firstly, a free crab gait and two turning gaits have been developed for hexapod robots to avoid forbidden areas. Secondly, a system of location based on different methods has been developed (odometry, dead-reckoning, DGPS and an extended Kalman filter). This system allows knowing, at any time, the position and orientation of the robot with the required precision for this task, considered in ±0.02 m. Finally, an algorithm for path tracking has been generated to move the robot along a predefined trajectory. Thus, the robot can make a complete exploration of the terrain.The presented algorithms have been validated experimentally using the SILO6 hexapod robot.