A Multimode SoC FPGA-Based Acoustic Camera for Wireless Sensor Networks

Acoustic cameras allow the visualization of sound sources using microphone arrays and beamforming techniques. The required computational power increases with the number of microphones in the array, the acoustic images resolution, and in particular, when targeting real-time. Such computational demand leads to a prohibitive power consumption for Wireless Sensor Networks (WSNs). In this paper, we present a SoC FPGA based architecture to perform a low-power and real-time accurate acoustic imaging for WSNs. The high computational demand is satisfied by performing the acoustic acquisition and the beamforming technique on the FPGA side. The hard-core processor enhances and compresses the acoustic images before transmitting to the WSN. As a result, the WSN manages the supported configuration modes of the acoustic camera. For instance, the resolution of the acoustic images can be adapted on-demand to satisfy the available network's BW while performing real-time acoustic imaging. Our performance measurements show that acoustic images are generated on the FPGA in real time with resolutions of 160x120 pixels operating at 32 frames-per-second. Nevertheless, higher resolutions are achievable thanks to the exploitation of the hard-core processor available in SoC FPGAs such as Zynq.