I. INTRODUCTION
In most of the robotic projects we have to solve two major control problems: (1) to provide human-robot interfaces that help accurately and intuitively operate the hardware, and (2) to organize the software architecture in the precise though flexible way to host all control algorithms and envision the possible directions of development. In the presented project we deal with manipulation using multi-finger gripper and have faced exactly these challenges. Advanced grippers, recently more and more popular, have enough degrees of freedom to perform difficult gripping tasks in both industrial and home environments. On the other hand, it is difficult and time consuming to operate them and program manipulation tasks by using standard input devices, like robot consoles and joysticks. Therefore, control of anthropomorphic robotic hands by sensor gloves has been extensively studied [1], [3], [6]–[8]. Even though, in our project we have focused on three-finger gripper we decided that, for sake of intuitiveness, it is worth to use sensor glove as an input interface. The detailed description of the system setup is shown in the next section. As the number of outputs from the glove exceeds the number of control inputs for the actuators, the simple master-slave operation with direct output-to-input mapping has to be replaced with additional data processing. Also the early stage of the project forced us to use the simulator instead of real manipulator with the gripper that are being built. Here comes the second of the challenges mentioned above the choice of software architecture. We would like to simply acquire and process data, control actuators both in the simulation environment and real hardware, and visualize results. The Robot Operating System (ROS) seems to be the solution fitting our needs. It is a thin, message-based, tool-based system designed for mobile manipulators. ROS is based on a Unix-like philosophy of building many small tools that are designed to work together [3]. The application of ROS into our project is presented in Section III while test results are shown in Section IV.