VOICE RECOGNITION BASED HOME AUTOMATION SYSTEM USING ANDROID AND ARDUINO

Transcription

1 VOICE RECOGNITION BASED HOME AUTOMATION SYSTEM USING ANDROID AND ARDUINO Dennis Brown Germiniasi da Silveira 1 dennisbrownger@yahoo.com.br Marlon José do Carmo 2 marlonreiff@yahoo.com.br 1 The University of Adelaide South Australia, Australia 2 Centro Federal de Educação Tecnológica de Minas Gerais Leopoldina, MG, Brazil Abstract. This paper proposes the development of a home automation voice command system. The system was modelled using some of the UML elements in a way to be flexible, and adaptable towards any residence structure. That means that, any user is able to dynamically add and remove electrical devices into the system without the help of a technician. It was designed an Android cellphone application, which communicates in a local network with an Arduino board. In order to assess the system, it was built a miniaturized residence to test it in a real case scenario. The findings suggest that the system works relatively well; however, there were a few instances that the system could not recognize the content properly. Keywords: Home Automation, Arduino, Android, Speech Recognition, Artificial Neural Network 1. INTRODUCTION Home automation can be understood as the uptake of technology in residential environment in order to enhance the quality of life of its inhabitants by providing services such as telehealth, multimedia entertainment, and power saving (Gurek et al., 2013). In this context, this technology focuses on investigating the means to improve the quality of life of inhabitants, in particular disabled people. The home automation industry has been growing rapidly, which is mainly fuelled up by the need to provide systems to support elderly and disabled people, mostly the ones that live alone. Furthermore, the world population is becoming older. Thus, home automation must adapt to the needs of these citizens (AlShu eili, Gupta, & Mukhopadhyay, 2011). The population structure of industrialized countries is changing; there will be less people to take care of the growing number of elderly citizens. This problem has encouraged a large number of studies involving robotics and automation. Rather than live in an elderly care institution, it is more feasible to inhabit home (Harmo et al. 2015). In addition, disabled people find difficult to switch on and off devices such as fans, air conditioners, lights, etc. They need help from someone to do these tasks. In the absence of someone to help, their lives become highly difficult. Therefore, the development of a system to assist them to turn on and off their home devices is crucial. The breakthrough in cutting edge voice command technology has an enormous potential in solving this problem (Gnanasekar, Jayavelu, & Nagarajan, 2012).

2 2. PURPOSE This paper focuses on the development of a voice command system based on the Offline Google Speech Recognition Engine. The system s user interface is a cellphone Android Application, which interacts with an Arduino board to execute the commands. 3. METHODS The system was designed and built to be as simple as possible providing all needed functionalities to the users. In this sense, a friendly user interface was developed to run on an Android phone. Then, a house model was built to test out the system operation. 3.1 System Overview By using voice commands, a user shall be able to switch on and off home devices such as lighting, fan, multimedia, etc. A key feature of this system is its flexibility to adapt to different scenarios, and devices, as the user is able to add, delete, and edit rooms, user accounts, and devices. For security purposes, the systems requires login to access it, and there are two kinds of users, the administrator user and the common user. Each one has a different scope of access in the system, for instance, the administrator has permission to add, delete and edit rooms, user account, and devices, whereas the common user is limited to only give voice commands. From a technical point of view, the system utilizes a local network in which the UI and the Arduino have different roles. The UI is a cellphone Android application client in the wifi network, and its main role is to provide the user the means to interact with the system. It also processes the user s voice commands through a third party artificial neural network system from Google and the outcome of that is sent to the Arduino via socket in the local wifi network. At last, it has a database to store data from the server, so the server does not need to be consulted every time, and it has a simple cryptography system to protect the data integrity in the system. The Arduino is the server in the wifi network, and its main role is to process and manage all the client requests sent to it, execute the requests and send a reply back to the client. 2.2 The User Interface (UI) The user interface was developed based on the following case use diagram which highlights and summarizes all the permissions granted to each user.

3 Figure 1: Case Use Diagram. As the user interface is an Android application, the main programming language to implement it is Java, which is object oriented. Another significant feature of this programming language is its capability to reutilize code. Thus, a class diagram that represents the UI was designed.

4 Figure 2: Class Diagram. Based on the class diagram displayed before, the user interface was coded using Eclipse IDE, and the outcome of that is showed by Fig. 3. These screenshots are three instances of the UI. Firstly, it is requested the user to login, secondly the UI downloads data from the server, and at last the main screen is displayed. On the main screen, the four main buttons vertically centralized provide, from the top to the bottom, access to devices, rooms, user account, and voice command. The other three buttons provide, from the left to the right, access to login/logout, data synchronization with the server and exit.

5 Figure 3: Some of The Android Application s Screen. 3.3 The Network Server The clients request processing is assigned to the Arduino server which includes turning on or off devices, storing or consulting data from a SD memory card. The SD memory card is the way the system stores essential data. Inside the card, there are four files utilized by the system: users.txt, rooms.txt, devices.txt, and usersblocked.txt. These four files are responsible for storing the users, the rooms, the devices and the blocked users of the system, respectively. The software responsible to manage all these functions is implemented on the Arduino Uno which has an Ethernet Shield to connect to the local wifi network. 2.4 The House Model In order to evaluate the system behaviour in a real case scenario, it was developed a miniaturized residence with five rooms as illustrated by the Fig. 4.

6 Figure 4: House Model. This house model is filled with seven devices symbolically represented by led s, apart from the fan which is represented by a twelve volts electric motor. The Table 1 details the rooms with the devices within. Table 1: Room and Devices. Room Kitchen Bedroom Garage Lounge Toilet Device Light, fan Light Light Light, television Light The led s are all connected into the Arduino s digital pins along with resistors to limit the electric current on them. The electric motor has an external power source and a transistor trigged by the Arduino is set as a switch to turn on or off the motor. The circuit is showed by Fig. 5.

7 Figure 5: Electric Circuit. The methodology adopted to simulate the system involved a user speaking a sequence of commands, and the result of that is photographed. 4. RESULTS The system was tested by a user, and four instances of its operation were randomly selected and displayed by the following figures. Turning on the fan in the kitchen (left) and turning off the television in the lounge (right): Figure 6: System Being Tested DISCUSSION The proposed system reached the expectations, by demonstrating a relevant flexibility in adapting to different scenarios and properly responding to the user voice requests. However, when it comes to processing time, the Google Speech Recognition Engine seemed to take a while to deal with the voice requests. In addition, there were a few instances that the user speech could not be recognized.

8 A similar work has been published in which the authors developed a bluetooth based home automation system using cellphone ( Piyare & Tazil, 2011). However, their system utilizes Bluetooth as a way to connect devices in the network which limits the range of connection. In addition, speech recognition was not considered in their work. 6. CONCLUSIONS As the world population is becoming older, the need to develop systems to assist elderly and disabled citizens is crucial. The project discussed in this paper suggests the development of a voice command system based on the Offline Google Speech Recognition Engine. For that, it was developed a cellphone Android application working along with an Arduino Uno board. The proposed system was tested and the findings suggest that the system works reasonably well, although some improvements shall be taken in regard with the voice command recognition tool. In some instances, the Google Speech Recognition was not able to recognize the speech. Therefore, that reveals that this tool is not fully reliable, and the uptake or development of a more reliable speech recognition tool may be considered. Acknowledgements I thank my supervisor, Marlon José do Carmo, for all the assistance towards me and my family for being always supportive. REFERENCES AlShu eili, H., Gupta, G.S. and Mukhopadhyay, S. (2011), Voice Recognition Based Wireless Home Automation System, IEEE, School of Engineering and Advanced Tehnology Massey University. Gnanasekar, A.K., Jayavelu, P. and Nagarajan, V. (2012), Speech Recognition Based Wireless Automation Of Home Loads With Fault Identification For Physically Challenged, IEEE, Gurek, A., Gur, C., Gurakin, C., Akdeniz, M., and Metin, S.K. (2013), An Android Based Home Automation System, IEEE, Harmo, P., Taipalus, T., Knuuttila, J., Vallet, J., and Halme, A. (2015), Needs and Solutions - Home Automation and Service Robots for the Elderly and Disabled, Helsinki University of Technology. Piyare, R., and Tazil, M. (2011), Bluetooth Based Home Automation System Using Cellphone, IEEE,