Minimizing area costs in GPS applications on a programmable DSP by code compression

Piia Saastamoinen1,  Jari Nurmi1,  Ilkka Saastamoinen2,  Mikko Laiho2
1Tampere University of Technology, Department of Computer Systems, 2Atheros Communications, Tampere, Finland


The amount of applications requiring personal satellite based navigation is growing rapidly at the moment. Complexity of the GPS (Global Positioning System) navigation algorithms and thus the memory requirements for the systems are growing at the same pace as the demands from customers. The large program memory footprint can be efficiently reduced by code compression. In this paper we describe in detail the analysis and compression procedures of typical GPS functions, as well as the on-chip decompression flow. For the GPS functions, our compression scheme achieves compression ratio of 55% at best.