Optimal control of building systems in major public places requires knowledge about the number and distribution of people within the building. The “S4ECoB” system being developed in the project uses a network of spatially-distributed microphones to pick up sounds in their vicinity. The system is capable of determining the number of people in parts of a building based on acoustic data and control building systems accordingly.

More specifically, eight microphones and their associated electronics for digitising and encoding signals form a microphone array. Up to three of these arrays can be connected to an energy-efficient embedded hardware and software platform. It will process audio data streams, classify acoustic events, and, from them, estimate the number of people in the rooms being monitored by the arrays. The computed data is transferred to a central unit, where it is processed further and building control is adapted to the current situation, thus minimising energy demands.

The system presented numerous challenges. To meet them, IMMS has developed novel solutions within the project, developing the embedded platform that processes the acoustic signals. This involved designing and building adequate hardware and implementing software components. The Institute has also designed and implemented the communications architecture between the components. The innovative architecture and the new methods adopted for the hardware and software implementation have proved to be the key to an embedded signal processing platform with low energy usage that offers not only very high performance but also the advantages of flexibility and low price.