Application of Contact-Based Sensors for Self-Localization and Object Recognition in Humanoid Robot Navigation Tasks

This paper presents the application of a six-axis force sensor and a novel optical three-axis tactile sensor to humanoid robot navigation system which is based on contact interaction towards supporting visual-based navigation. The force sensors are mounted on humanoid robot arms to perform grasping in self-localization task to define the robot´s position and orientation. The grasping results guided the robot locomotion and avoid it from collision. Meanwhile the optical three-axis tactile sensors are mounted on cooperative two-finger system for object handling tasks. The tactile sensor is capable of acquiring normal force and shearing force. Experiment with hard and soft objects are performed which results revealed good performance of the integrated robotic fingers and tactile sensor system to recognize and grip the objects. The presented control algorithms for both sensors are capable of preventing the probability of damage to the sensors and objects during robust grasping and object handling tasks.