World Applied Sciences Journal 19 (11): 1563-1567, 2012 ISSN 1818-4952 © IDOSI Publications, 2012 DOI: 10.5829/idosi.wasj.2012.19.11.2470 Corresponding Autor: Shaukat Ali, Department of Information Technology, University of Haripur, Haripur, Pakistan. 1563 Natural Language Interface for Sensor Networks Shaukat Ali, Huma Javed, Azhar Rauf, Saeed Mahfooz and Shah Khusro 1 2 2 2 2 Department of Information Technology, University of Haripur, Haripur, Pakistan 1 Department of Computer Science, University of Peshawar, Peshawar, Pakistan 2 Abstract: Sensor networks have a number of applications in military, industries, health and other commercial and non-commercial institutes. Many people working in these different areas may not be technical persons to the query or other similar type languages. The user of these environments may be interested to retrieve data of their interest, but they cannot query the sensor networks because they do not know the query language. This paper proposes a Natural Language Interface to Sensor Networks (NLISNs) so that a non-technical person may be able to query sensor’s data. Since most of the users of sensor networks are non-technical for example Biologist, Military person etc. Therefore, this work will enable these persons to query and analyze the data of sensor networks. The accuracy result of the proposed system is 74%. Key words: Sensor Networks Natural Language Query INTRODUCTION The problem being addressed in this paper is to Sensors have a variety of application in today’s sensors. One of potential solution to this issue is ubiquitous computing environment. It has a number of to design a Natural language interface for sensor applications in Military, Environmental studies, Health, networks. The area of NLP (Natural Language Home appliances and other commercial applications [1]. Processing) is an active research area. An area in It provides many useful functions by sensing the which NLP systems are powered enough to be used in condition of physical, biological, chemical and database query systems [4]. Research is going on meteorological phenomena in order to build expert Natural Language interfaces for databases, but little work systems in areas, such as, animal tracking, precision has been done in the field of sensor database. Literature agriculture, risk monitoring, or hazard management. [5-7], are using natural language interface to data, but Therefore, it becomes challenge for the external users or they are focussing on relational database as their target systems to discover and invoke the sensor-derived data source. [2]. The work of this paper focuses on the query based Three different approaches are used to get data from technique to retrieve data from sensor network, but there sensor, are mentioned in the existing literature [3]: are many Non-technical persons, which might face Event Based Approach: Events are registered with sensors with query languages, for example, the persons who are which return data when that event occurs. These events working in Military, Environmental studies, Health and are pre-programmed in the sensor chip. Home appliances etc. Therefore, this paper concentrates Query Based Approach: In this approach the sensor order to enable a common person to query sensor’s data. returns data whenever a query is sent to the sensor The experimental results show 74% of success rate for network from the base station. the translation of natural question to SQL query. Periodic Sensing: The sensor always senses data and it from user interface, which is provided in the is returned to base station on periodic basis, after some implementation of system. Rest of the paper is structured fixed interval. as follow: Literature review is discussed in section 2. make it easier for a non-expert to interact with problem in query writing because of not being familiar on the Natural Language Interface for sensor networks in The remaining suggested queries need minor changes