An accelerated signal tracking module using a heterogeneous multi-GPU platform for real-time GNSS software receiver

Abstract

This paper presents the design of an accelerated signal tracking module using a heterogeneous multi-graphics processing unit (GPU) platform for real-time global navigation satellite system (GNSS) software receiver. We also propose a load balancing method for the efficient use of the multi-GPU. The proposed method allocates the number of channels to each GPU, and the GPU generates replica signals and performs correlation with the allocated channels. Through experiments, we showed that the proposed load balancing method achieves near optimal performance for GNSS signal tracking and reduces the operation time by 60% compared to that obtained using a single GPU.