In Global Navigation Satellite System (GNSS)-denied environments, terrestrial signals of opportunity (SoOP) offer an alternative for positioning, but synchronization impairments such as clock offsets, drift, and multipath limit performance. This paper proposes a receiver-centric multi-channel time-difference-of-arrival (TDOA) localization framework based on Digital Audio Broadcasting (DAB) signals. The method exploits the DAB null symbol for coarse timing and the phase reference symbol (PRS) for fine synchronization, followed by sub-sample time-of-arrival (TOA) estimation. A double-difference formulation removes inter-receiver clock offsets, while a peak-to-sidelobe ratio (PSR)-based weighting improves robustness. A bias correction step mitigates errors due to multipath. Finally, a coordinated-turn extended Kalman filter (CT-EKF) further refines position estimates. Results show improved accuracy over conventional TDOA with Gauss-Newton estimation, especially in challenging conditions.