The objective of this proposal is to thoroughly investigate the deployment of non-isotropic sensors for smart building applications. Various sensing and control applications in smart building require accurate and efficient localization and tracking of human and activities. To achieve this goal, the non-isotropic networked sensor deployment is essential.
This proposal addresses a new set of problems arising from the deployment of four types of sensors: cameras, tripwire, motion sensors and microphones. The set of deployment problems would strive for full sensor coverage and wireless connectivity with a complex floor plan, and involve one or more of the following constraints: (i)Non-isotropic model using visibility (cameras and tripwires), (ii) Non-overlapping sensing range, (iii) Robust 2-coverage for accurate localization. The sensor deployment problem will heavily involve the geometric properties of the environment (building layout), as well as the geometric properties of the sensors (non-isotropic sensing ranges). Ideas from computational geometry will be applied to tackle these problems and approximation algorithms with worst-case guarantee for theoretical rigor as well as practical algorithms suitable for implementations will be developed.
A clear understanding on how to deploy sensors for smart building systems will allows for efficient planning and effective practice of networked sensor deployment to achieve the high quality sensing desired by various indoor applications. For developers, this work will significantly reduce the design and development costs for designing and building smart building systems. For end users, the resultant testbed will lead to exciting applications that will significantly improve the quality of life. It also enriches engineering curriculum to integrate the theory of computational geometry with the system building practice. The PIs are committed in improving female presence in computer science and exposure of research for high school students.
