Self-Assessment for Single-Object Tracking in Clutter Using Subjective Logic

Toward Precise Ambiguity-Aware Cross-Modality Global Self-Localization

“…Sect. III-B), we will explain how we deal with clutter measurements and misdetections using the singleobject-tracking formulation of the measurement likelihood [52], [53], [54] (cf. Sect.…”

“…it is primarily obscure if the measurement corresponds to the true pose or clutter. To model misdetections and clutter among unknown data association, we employ the single-object tracking (SOT) formulation of the complete measurement likelihood [52], [53], [54]…”

“…As opposed to histogram or particle filters, this formulation enables us to simultaneously evaluate our sensor-to-map matching network at different regions in the map that may be arbitrarily distant to each other while determining the most likely hypothesis over time through the fusion of odometry measurements and map-based pose observations while keeping the computational demand bounded. To account for misdetections and clutter measurements we employ the complete measurement likelihood formulation for single-object tracking (SOT) [52], [53], [54] which not only provides means for outlier rejection, but also augments the probability calculation of the hypotheses.…”

Online Performance Assessment of Multi-Sensor Kalman Filters Based on Subjective Logic

“…Regarding the application of SL for SA in filtering, our earlier work [7] introduced a novel SA technique for linear Kalman filtering that includes explicit modeling of statistical uncertainty. Based on this, we proposed a SA framework for single-object tracking in clutter using linear Kalman filtering and nearest neighbor association in [11]. This approach considers various aspects of reliability assessment, such as the correctness of the assumed measurement noise, the detection probability, and the clutter rate.…”

“…The presented SL SA for Kalman filtering is mainly based on [7], [11]. A detailed algorithm and its comprehensive explanation of the SA method for the linear case can be found in [7].…”