Motivation:
The estimation of diffusion fields is a common problem in environmental monitoring applications. Wireless sensor networks (WSNs) are often used which measure the diffusion field, e.g. a temperature field distributed over a certain area. Diffusion fields are highly nonlinear such that an estimation of the field itself is a challenging task. For this task, we focus on Kernel adaptive filters, e.g. the Kernel Least-Mean Square (KLMS), which are able to learn nonlinear functions. Furthermore, since we consider a WSN scenario we make use of distributed signal processing algorithms where information is exchanged among the sensor nodes in order to estimate the complete diffusion field.

Goal:
In this project, known distributed KLMS algorithms from the literature shall be applied to both simulated temperature fields and measurements of real temperature fields by sensor networks. The algorithms from the literature need to be understood and adapted to the provided data of the temperature field. If possible, the algorithms should be optimized to gain further improvement in their performance.

Requirements:

• Basic Matlab programming skills
• Knowledge from DSP Advanced is helpful