Development of hand gesture recognition sensor based on accelerometer and gyroscope for controlling arm of underwater remotely operated robot

Hand Gesture Recognition sensor based on accelerometer and gyroscope is a sensor for capturing the positions of operator hand while controlling underwater remotely operated vehicle equipped with an arm. The proposed system has an advantage in its convenience by means of no training or exercise needed for operator before using it. The key issue here is how beginner operator could use easily the underwater remotely operated robot arm without any specific training. The conventional one uses a joystick for controlling the underwater system and it is inconvenience for beginner user as well as less precision. The proposed system consists of two main part: (1) ground station and (2) underwater remotely operated robot arm. This paper proposes the development of hand gesture recognition sensor used by operator at the ground station for controlling robot arm at the underwater robot. The proposed sensor uses accelerometers and gyroscopes installed in elbow, forearm, and wrist. These devices measure 3D position of each joints for constructing 3D position of hand. We design sensor's casing for its convenience of use by using CAD software. Each sensor is connected by Arduino Nano microcontroller having compact circuit and embedded it into sensor's casing. The sensors are connected to a microcontroller acting as master connected to microcontroller slave (sensor part). These sensors value are converted to 3D position by using forward kinematic. The forward kinematic values are sent to the underwater robot by using a wire utilizing Pulse Position Signal. Then, it converted again to servo's movement by using inverse kinematic. The result is operator able to control the underwater remotely robot arm by utilizing hand gesture directly. The last, operator could control the robot gripper based on flex sensor installed in operator's fingers. The accuracy of the sensor has been tested under laboratory condition, it has 98% of accuracy.