Title | ||
---|---|---|
A Robust Genetic Algorithm For Feature Selection And Parameter Optimization In Radar-Based Gait Analysis |
Abstract | ||
---|---|---|
Contactless medical gait analysis plays an important role in assessing health conditions for ambient assisted living. Prior work on radar-based gait analysis has mainly focused on classification of different gaits or detecting asymmetry. We demonstrate that it is possible to estimate medically relevant gait characteristics based on features from the radar back-scatterings. Given a set of radar features, we predict the maximal knee angle during walking for the left and right leg. We present a new robust genetic algorithm (GA) based on a nonlinear regression method that simultaneously performs feature sele ction, parameter optimization for the support vector machine and outlier rejection by encoding these aspects into the chromosome design. Using genetic operations, the proposed algorithm significantly outperforms competing methods on a real-world data set recorded with a 24 GHz continuous-wave radar. |
Year | DOI | Venue |
---|---|---|
2019 | 10.1109/CAMSAP45676.2019.9022515 | CAMSAP |
Keywords | DocType | Citations |
genetic algorithms, support vector machine, gait analysis, Doppler radar, ambient assisted living | Conference | 0 |
PageRank | References | Authors |
0.34 | 0 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Lisa Dawel | 1 | 0 | 0.34 |
Ann-Kathrin Seifert | 2 | 4 | 2.09 |
Michael Muma | 3 | 144 | 19.51 |
Abdelhak M. Zoubir | 4 | 1036 | 148.03 |