From the lab into the wild
Brano Kusy
CSIRO in Brisbane (Australia)
DEIB - PT1 Room
July 26th, 2016
11.00 am
Contact:
Luca Mottola
Research Line:
Advanced software architectures and methodologies
CSIRO in Brisbane (Australia)
DEIB - PT1 Room
July 26th, 2016
11.00 am
Contact:
Luca Mottola
Research Line:
Advanced software architectures and methodologies
Abstract
From the lab into the wild: addressing challenges of long-term, autonomous deployments of wireless sensor networks
Deploying wireless sensing systems that are expected to last months or years with limited user interaction presents a unique set of challenges that need to be overcome. In majority of the cases, the main challenge boils down to maximizing the information gain within the limited energy budget available to the sensing platform. In this talk, I will present several strategies that we identified through a decade of sensing in agricultural, wildlife monitoring, and forestry applications.
The key factor to the successful deployment of sensing technology is adaptiveness to the specific, often dynamically changing, context of the sensing platform. Solar harvested energy and activity patterns of the tracked phenomena can change on a weekly or even daily basis and the sensing algorithms need to adjust to keep the information gain at maximum. Another key factor is cooperation between sensor nodes operating in a similar context: statistical algorithms can be used to reduce sampling rates of a group of nodes to save energy, to repair corrupted data, or to infer close-to-optimal sampling parameters for newly deployed sensor nodes. I will finish the talk discussing several emerging areas of interest for pervasive computing, including long range wireless communication, augmented reality, and emerging sensing modalities.
Deploying wireless sensing systems that are expected to last months or years with limited user interaction presents a unique set of challenges that need to be overcome. In majority of the cases, the main challenge boils down to maximizing the information gain within the limited energy budget available to the sensing platform. In this talk, I will present several strategies that we identified through a decade of sensing in agricultural, wildlife monitoring, and forestry applications.
The key factor to the successful deployment of sensing technology is adaptiveness to the specific, often dynamically changing, context of the sensing platform. Solar harvested energy and activity patterns of the tracked phenomena can change on a weekly or even daily basis and the sensing algorithms need to adjust to keep the information gain at maximum. Another key factor is cooperation between sensor nodes operating in a similar context: statistical algorithms can be used to reduce sampling rates of a group of nodes to save energy, to repair corrupted data, or to infer close-to-optimal sampling parameters for newly deployed sensor nodes. I will finish the talk discussing several emerging areas of interest for pervasive computing, including long range wireless communication, augmented reality, and emerging sensing modalities.
Short Bio
Brano Kusy received the B.S. (2000) and M.S. (2002) degrees from Comenius University in Bratislava, Slovakia, and the Ph.D. from the Vanderbilt University, Nashville, USA (2007), all in Computer Science. He was a postdoctoral scholar at Stanford University, California, between 2007 and 2009. Since 2009, he has been at CSIRO in Brisbane, Australia, where he works as a research scientist in the Cyber-Physical Systems program. He is also an adjunct lecturer with the University of Queensland, Australia where he has been teaching a course on Advanced Embedded systems since 2010. At CSIRO, he leads the Pervasive Computing team, which performs research and works on projects in embedded distributed sensing, wireless communications, and embedded learning. Brano's research interests are in coordination and communication algorithms for wireless sensor networks, radio channel measurement processing for networking and localization, and embedded machine learning.