🤖 AI Summary
This work addresses the challenge of fairly comparing Bayesian log-odds and Dempster’s rule of combination in two-dimensional occupancy grid mapping by proposing a unified matching framework based on the pignistic transformation. By mapping the outputs of diverse fusion methods into a consistent decision probability space, the framework effectively isolates the influence of sensor-specific parameters, thereby enabling a direct comparison of the fusion rules themselves. For the first time, this approach facilitates systematic evaluation across paradigmatically distinct methods, with validation demonstrated through simulations, real LiDAR data, and path planning tasks. Experimental results reveal that under pignistic probability (BetP) matching, Bayesian methods significantly outperform Dempster-based approaches (15/15 consistent outcomes, p = 3.1e-5, effect size 0.001–0.022), whereas the conclusion reverses when using normalized belief matching—highlighting the strong dependence of fusion performance on the matching criterion and underscoring the necessity and novelty of the proposed framework.
📝 Abstract
We introduce a pignistic-transform-based methodology for fair comparison of Bayesian log-odds and Dempster's combination rule in occupancy grid mapping, matching per-observation decision probabilities to isolate the fusion rule from sensor parameterization. Under BetP matching across simulation, two real lidar datasets, and downstream path planning, Bayesian fusion is consistently favored (15/15 directional consistency, p = 3.1e-5) with small absolute differences (0.001-0.022). Under normalized plausibility matching, the direction reverses, confirming the result is matching-criterion-specific. The methodology is reusable for any future Bayesian/belief function comparison.