The AMANDA data fusion engine consists of two main sub-components that ensure the optimization capabilities required in order to enable the low-power, low-cost and maintenance-free design of the ASSC:

– The pre-processor

– The controller

The pre-processor receives the sensor measurements in raw format and extracts the measurements in time series matrices to take advantage of each data point and use it efficiently in later stages of processing. At the same time, it extracts additional required features such as trend and seasonality of the data to be used by the controller component. The controller receives as an input from the pre-processor the sensor measurements, the time series of these measurements and the calculated metrics and/or features.

The main execution cycle of the ASSC can be summarised to the following:

– Wake-up phase

– Energy availability assessment

– Data acquisition

– Decision phase

– Results transmission

– Sleep phase