ECML PKDD 2008

Embedded Machine Learning: Using Support Vector Machines in Wireless Sensor Networks using TinyOS and Lego Mindstorms NXT

The tutorial on embedded machine learning will present a case study of implementing and using a binary support vector machine in wireless sensor networks. It will use a very popular operating system for wireless sensor networks called TinyOS and the new exciting open hardware/software platform Lego Mindstorms NXT from LEGO. Outline of the tutorial (structured list of topics):
The tutorial provides an overview of embedded machine learning and an Overview of wireless sensor networks, TinyOS and the programming language nesC. It provides an introduction to LEGO MINDSTORMS and the main hardware items needed to engage the problem in a meaningful way. The mapping of the binary support vector machine to the constraints of the embedded machine learning problem is given (memory, battery, little CPU).