Distributed Software Defined Radio Testbed for Real-time Emitter Localization and Tracking

摘要

We present a real-time TDOA localization testbed utilizing software defined radio and GPS based synchronization. A distributed software stack enables the localization of unknown signals by transporting the recorded baseband samples of several sensors over a backhaul network for centralized signal processing in a fusion center. To overcome the relatively large drifts of synchronization signals provided by standard GPS receivers, the system is permanently re-calibrated using a stationary reference beacon or alternatively a signal of opportunity. In a further step, we introduce a new generation of low-cost GPS disciplined oscillators to simultaneously achieve single digit meter accuracy and high update rates necessary for target tracking. The tracking is performed using a Kalman filter running on the output of the localization algorithm. Therefore, we also study the dilution of precision effect and adopt the filter accordingly. Further optimizations tackling the ambiguity problem in a three sensor system and optimum reference selection are introduced. Finally we present outdoor measurements to characterize the performance of the system.