Figure 2.

Hyperparameters used in graph trilateration. (A) $\tau$ represents the size of the time window to consider the temporal context in the proposed method. (B) The step method uses non-overlapping sliding windows, where the window size and sliding interval are the same. The sliding method indicates overlaps in sliding windows to induce more temporally smooth BLE localization. (C) The weighting factor applies temporal smoothing to BLE locations estimated over $T=3$ timesteps using the weighted average. We explore various weighting values with $s_1<s_2<s_3$.