New adaptive filter enhances state estimation for nonlinear systems

Researchers have developed a new adaptive split-combine Gaussian mixture filter (AMF) designed to accurately estimate the probability density function (PDF) of states in nonlinear and multimodal systems. This filter addresses limitations in existing Gaussian mixture filters by adaptively splitting and combining Gaussian particles without requiring online numerical optimization. The AMF demonstrates superior performance compared to baseline filters on benchmarks involving Van der Pol oscillators and the Lorenz attractor, with a parallel implementation enhancing its computational efficiency for high-fidelity PDF estimation. AI

RANK_REASON Academic paper detailing a new filtering technique. [lever_c_demoted from research: ic=1 ai=0.7]

Read on arXiv stat.ML →

• adaptive split-combine Gaussian mixture filter
• Gaussian mixture filter
• Kalman-type filters
• Lorenz attractor

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →