Enhanced RF Sequence Recovery: Comparative Analysis of AoA-Only vs AoA+TDoA Fusion

We present a comparative study of grid-based
trajectory recovery using angle-of-arrival (AoA) alone versus
fused AoA and time-difference-of-arrival (TDoA) observations.
Building on our prior AoA-only framework, we integrate TDoA
measurements into a discrete beam-search inference pipeline and
quantify the resulting gains under sparse and noisy conditions.
Across Monte Carlo trials, fusion yields extbf25–45% error
reduction relative to AoA-only in regimes with limited observation fractions (ho ≤ 0.5) and high AoA noise (σheta ≥ 10◦
),
while maintaining robustness to TDoA noise up to extbf100 ns
(≈30 m range). A geometry-based dilution of precision (GDOP)
analysis confirms that augmenting AoA with TDoA reduces
uncertainty ellipse eccentricity, particularly for non-ideal sensor
layouts. Comparisons against baseline extended Kalman and
particle filters highlight fusion’s advantages in discrete multihypothesis tracking, with similar accuracy but lower complexity
at modest beam widths. Synthetic experiments (100-step trajectories, 3-sensor triangle) demonstrate that AoA+TDoA consistently
achieves <300 m mean error under stress conditions where AoAonly exceeds 500 m. These results underscore the operational
relevance of multi-modal fusion for electronic warfare and
passive geolocation, while motivating future work on real-world
validation, synchronization costs, and adaptive fusion strategies.