A scalable acoustic sensor network for model based monitoring of urban traffic noise

A good understanding of the acoustic environment due to traffic in urban areas is very important. Long term monitoring within large areas provides a clear insight in the actual noise situation. This is needed to take appropriate and cost efficient measures; to asses the effect of measures by comparing before and after implementation and to have an instrument for communication and policy making. A monitoring approach has been developed which is based on two types of low cost acoustic sensor nodes, a basic wireless sensor node and an advanced sensor node. The basic node only measures sound levels, while the advanced node measures time signals and is programmable to compute the sound properties like the spectral content. Both sensor types can be freely combined in an acoustic sensor network. The network is scalable and is built up of subnetworks. Each subnetwork is self organizing and the data from the basic nodes within each subnetwork is sent to an advanced sensor via multi-hopping. Each advanced sensor node sends the data regularly to a central database and these nodes can be accessed by a remote connection. A field test is conducted with the acoustic sensor network in an urban environment, where traffic noise is measured during two weeks with 24 basic nodes and three advanced nodes. Sound is measured close to a busy road, but also in quiet backyards and quiet adjacent streets. Results of this field test will be shown. Network aspects will be discussed and measurements are compared to results from a standard acoustic model. Large differences are found especially in the more quiet areas. Further, it will be shown that more information can be retrieved from the monitoring data, which can be used for the aforementioned purposes.