Sensor Network Localization

[HOME] [EDUCATION] [PUBLICATIONS] [RESEARCH] [SOFTWARE] [INTERNSHIPS] [PROJECTS]

LOCALIZATION OF DISTRIBUTED ACOUSTIC SENSORS/ACTUATORS

Position Calibration of Microphones and Loudspeakers in Distributed Computing Platforms Vikas C. Raykar, Igor Kozintsev, and Rainer Lienhart, IEEE Transactions on Speech and Audio Processing, Volume 13, Issue 1, pp. 70-83, Jan. 2005.

Arrays of audio/video sensors and actuators (such as microphones, cameras, speakers and displays) along with array processing algorithms offer a rich set of new features for emerging multimedia applications. Until now, array processing was mostly out of reach for consumer applications perhaps due to significant costs of dedicated hardware and complexity of processing algorithms. On the other hand, several mobile computing and communication devices like laptops, PDAs and tablets are equipped with multiple audio/video sensors and actuators. An ad-hoc network of such devices can be used to form a distributed sensor network. A prerequisite for using distributed audio-visual I/O capabilities is to put the sensors and actuators into a common time and space.

This work focused on providing a common space by automatically determining the relative 3D positions of audio sensors and actuators. A closed form approximate solution is derived, which is further refined by minimizing a non-linear error function. The formulation and solution accounts for the lack of temporal synchronization among different platforms. An approximate expression for the mean and covariance of the implicitly defined estimator is derived using the implicit function theorem and approximate Taylors' series expansion. The theoretical performance limits for the sensor positions are derived via the Cramer-Rao bound and analyzed with respect to the number of sensors and actuators as well as their geometry. Extensive simulation results and the practical details of implementing our algorithms in a real-life system are discussed.

This work was performed when I was working as an Intern at Intel Corporation, Santa Clara. The ACM Multimedia paper formulates the problem of position calibration of multiple microphones and speakers on a distributed computing platform. It does not discuss in detail the issue of getting a good initial guess for the minimization routine to converge. The ICCV workshop paper descibes an analytical method to get a good initial guess. For the most complete paper which addresses all the issues refer to the journal paper in IEEE Transactions on Speech and Audio Processing special issue. This work resulted in filing of two patents.

[Sensor Network Position Calibration papers] [DDSP project at Intel Research]

Related publications

Automatic Position Calibration of Multiple Microphones Vikas C. Raykar and Ramani Duraiswami, In Proceedings of International Conference on Acoustics, Speech and Signal Processing (ICASSP 2004), Montreal, Canada, May 2004, vol. IV, pp. 69-72 [ Slides ]

Position Calibration of Audio sensors and actuators in a distributed computing platform Vikas C. Raykar, Igor Kozintsev and Rainer Lienhart (ACM Multimedia 2003, Berkeley, CA, USA, November 2003, pp. 572-581) (acceptance rate of 17%) [ Slides ]

Self Localization of acoustic sensors and actuators on distributed platforms Vikas C. Raykar, Igor Kozintsev and Rainer Lienhart (ICCV 2003 International Workshop on Multimedia Technologies in E-Learning and Collaboration, Nice, France, October 2003) [ Slides ]

Providing Common Time and Space in Distributed AV-Sensor Networks by Self-Calibration R. Lienhart, I. Kozintsev, D. Budnikov, I. Chikalov, and V. C. Raykar in Intelligent Multimedia Processing with Soft Computing Series: Studies in Fuzziness and Soft Computing, Vol. 168 Tan, Yap-Peng; Yap, Kim H.; Wang, Lipo (Eds.) 2005.

Method for 3-Dimensional position calibration of audio sensors and actuators on a distributed computing platform. Vikas C. Raykar, Rainer W. Lienhart, and Igor V. Kozintsev, US Patent number 6,941,246, 2005.

Three-Dimensional Position Calibration of Audio Sensors and Actuators on a Distributed Computing Platform. Vikas C. Raykar, Igor V. Kozintsev, and Rainer W. Lienhart, US Patent number 7,035,757, 2005.