GPU-based real-time synthetic aperture sonar processing on-board autonomous underwater vehicles

This paper describes a hybrid CPU/GPU implementation of the synthetic aperture sonar (SAS) processing algorithm developed under the Autonomous Naval Mine Counter Measures (ANMCM) programme at the NATO STO Centre for Maritime Research and Experimentation (CMRE). This new implementation allows processing of SAS data in real time onboard an autonomous underwater vehicle (AUV). This is a key component to enable the development of a higher level of autonomy, allowing the AUV to adapt the mission profile to environmental conditions and sonar performance in order to meet high level objectives. The software has been developed focusing on its embedded implementation on-board an AUV with a low-end NVIDIA GPU, but can also be used in post-processing taking advantage of a higher grade GPU. The first results obtained at sea demonstrated the capability to produce SAS images in real time.