Natural language search of sensor data

As sensors become more affordable, sensor networks are increasingly deployed to monitor diverse environments. However, these sensor network deployments often utilize different standards for communication and data storage. As a result, it is challenging to build large-scale pervasive systems able to find, query, and analyze information across a diverse set of sensor networks. Additionally, aggregating sensor data from various sources is difficult because data can be sampled using different levels, units, rates, and resolutions. To address these challenges, we have developed a pervasive sensor network search environment based on natural language processing and the semantic web. By using natural language, we can understand the context of the query and use semantic rules about the data to aggregate and transform data to more useful results. To demonstrate the system, we have deployed the environment in two application domains. In each domain, the system successfully answers domain-specific natural language queries.