System Implementation of a Gesture Detection Glove for Human-Computer Interaction Felipe Antunes Quirino, Marcelo Romanssini, Mathias Baldissera and Alessandro Girardi Computer Architecture and Microelectronics Group - Federal University of Pampa Alegrete, Brazil felipeantunesquirino@alunos.unipampa.edu.br Abstract—The use of portable electronic devices, such as smartphones and tablets, is diffused around the world. However, because of the reduced size, the interaction with the human being is difficult, especially the data input by typing on virtual keyboards in the screen. Also, it is necessary to develop new interfaces that allow a more agile human-computer interaction with less effort to the user. This report describes the development of a gesture capture glove capable of recognizing patterns of hand movement and transform them into useful information. The proposed system consists of six accelerometers and six gyroscopes located at the fingertips and over the posterior part of the hand, allowing precise detection of movements in three- dimensional space. A series of applications were developed, such as the recognition of gesture patterns of sign languages, allowing people with some kind of special needs (blind or deaf-mute, for example) to interact efficiently with a computer or smartphone. The glove may replace conventional joysticks and mice devices with the advantage of inserting more degrees of freedom to the control. We also developed a 3D virtual hand model controlled by the glove. The proposed system is cheap and can be implemented with conventional electronics devices available in the market. Index Terms—glove, HCI, gesture capture I. I NTRODUCTION The relationship between man and machine is becoming more natural and less complex with the evolution of technol- ogy. Devices capable of automatically recognizing user intent facilitate human-machine interaction. This type of technology in recent years has been gaining space in both market and research fields. The reduced size of portable electronic devices, such as smartphones, smartwatches and tablets have turning it difficult for human beings to interact with small input keyboards [1]. The implementation of a gesture capture glove has the objective of improving the interaction with these devices, since, through gestures, the user can transmit the intention of performing a certain activity or command. This paper describes the development of a gesture detection glove capable of recognizing patterns of hand movement and transform them into useful information. The proposed system consists of six accelerometers and six gyroscopes [2] located at the fingertips and over the posterior part of the hand allowing precise detection of movements in three-dimensional space. The sensors are read by a microcontroller, which processes the values of linear and angular acceleration and sends the information to the final device (a computer, a smartphone or a tablet). Fig. 1 depicts the system structure. The development of gesture capture devices has been re- ported previously in the literature. The Smart Glove project [3] uses both three-axes accelerometers and pressure sensors to detected the curves between the fingers and capture move- ments of the hand. There are also other devices whose purpose is to facilitate human-machine interaction, like Kinect, that consists of a camera that captures three-dimensional images in conjunction with an infrared projector and a monochrome CMOS sensor to create the 3D environment and detect its modifications [4]. The disadvantage of this method is that it depends on the space and luminosity of the environment. The Leap Motion is another interesting recognition device. It recognizes fingers and hand movements by means of infrared LEDs [5]. This equipment does not need any contact with the hand and it is portable. However, it is possible to verify that the gesture capturing by means of accelerometers and gyroscopes presents advantages in relation to other approaches, mainly because of the robustness. There is no interference according to the luminosity, skin color or anatomy from different users. In addition, the ease of use, intuitive operation and freedom of movement in any environment makes a glove a good alternative for automatic hand movement recognition. II. SYSTEM I MPLEMENTATION The hand gesture is captured by six accelerometers and gyroscopes of type MPU 6050, providing linear and angular acceleration data for each finger and for the dorsal area of the hand. Fig. 2 shows the location of the sensors on the glove. The PAMPIUM microcontroller is used for control, performing configuration and capturing sensor data. Each MPU 6050 has a three-axes accelerometer, responsible for measuring the linear acceleration on the x, y and z-axes, as well as a three-axes gyroscope that measures the angular acceleration over the x, y and z-axes, thus obtaining a good notion of three-dimensional space. The HC05 is a Bluetooth SPP (Serial Port Protocol) module, designed for transparent wireless serial communication, which makes it easy to interface with the final device. PAMPIUM is a fully configurable microcontroller designed by our research group. This microcontroller has a general- purpose 16-bit RISC architecture described in System Verilog.