Reference for adapting the terrain-navigation filters and connecting them to the IEEE TAES paper.
| Component | Source procedure |
|---|---|
| OOSM candidates | OOSM.m reuses stored prior clouds; paper Eqs. (8)–(9) instead condition past-state sampling on accepted and current states. |
| OOSM acceptance | The helper returns its final candidate; the paper adds a covariance-determinant acceptance test. The entry point accepts equal current-update determinants. |
| Covariance | MATLAB cov uses sample normalization; paper Eq. (7) normalizes by particle count. |
| Experiment | The settings block selects a demonstration configuration; use the paper's Section IV and Table I for its experiment settings. |
The APF applies the lookahead likelihood-ratio correction. The MPF uses critical-bandwidth mode estimation and nearest-center assignment, resamples each mode to its existing count, then resets global weights uniformly. Thus mode population counts determine the next step's prior mass; preserve that behavior when comparing historical outputs.
Terrain particles must remain inside the interpolation neighborhood. Likelihood normalization requires a positive finite total, and OOSM requires a nonempty queue. dskensity2d treats coordinate marginals independently; the significance calculation requires positive marginal variances.
FindCriticalBW returns centers from the final density grid when its search reaches the minimum bandwidth. A transition found on the final iteration retains the preceding bandwidth and centers.
Git history and the original entry-point header identify Youngjoo Kim as the author. Use the paper citation and preserve the MIT license and existing third-party notices, including those in HGMeanShiftCluster.m.
The paper describes SRTM terrain near 38° N, 128° E; the data guide describes the supplied DB_part.mat structure. Confirm dataset and third-party reuse terms with the respective rights holders.
APF and bandwidth checks exercise importance weights, ancestor correspondence, and critical-bandwidth termination. The simulation guide specifies dependencies, entry points, and output variables.