NANO-SLAM : Natural Gradient Gaussian Approximation for Vehicle SLAM
Tianyi Zhang, Wenhan Cao, Chang Liu, Feihong Zhang, Wei Wu, Shengbo Eben Li
TL;DR
This work targets accurate vehicle localization in GPS-denied environments by enhancing RBPF-based SLAM. It introduces NANO-SLAM, which replaces linearization-based sampling with a variational Gaussian optimization solved via natural gradient descent on the Gaussian manifold, while landmarks continue to be estimated with EKF and final poses are obtained through importance sampling. The key contributions are formulating the sampling distribution as a KL-based objective, deriving natural gradient updates for Gaussian parameters, and validating the approach on real-world data with substantial RMSE improvements (over 50%) and only modest computational overhead. The results demonstrate that NANO-SLAM provides more accurate, robust vehicle localization suitable for real-time autonomous driving in challenging urban scenarios.
Abstract
Accurate localization is a challenging task for autonomous vehicles, particularly in GPS-denied environments such as urban canyons and tunnels. In these scenarios, simultaneous localization and mapping (SLAM) offers a more robust alternative to GPS-based positioning, enabling vehicles to determine their position using onboard sensors and surrounding environment's landmarks. Among various vehicle SLAM approaches, Rao-Blackwellized particle filter (RBPF) stands out as one of the most widely adopted methods due to its efficient solution with logarithmic complexity relative to the map size. RBPF approximates the posterior distribution of the vehicle pose using a set of Monte Carlo particles through two main steps: sampling and importance weighting. The key to effective sampling lies in solving a distribution that closely approximates the posterior, known as the sampling distribution, to accelerate convergence. Existing methods typically derive this distribution via linearization, which introduces significant approximation errors due to the inherent nonlinearity of the system. To address this limitation, we propose a novel vehicle SLAM method called \textit{N}atural Gr\textit{a}dient Gaussia\textit{n} Appr\textit{o}ximation (NANO)-SLAM, which avoids linearization errors by modeling the sampling distribution as the solution to an optimization problem over Gaussian parameters and solving it using natural gradient descent. This approach improves the accuracy of the sampling distribution and consequently enhances localization performance. Experimental results on the long-distance Sydney Victoria Park vehicle SLAM dataset show that NANO-SLAM achieves over 50\% improvement in localization accuracy compared to the most widely used vehicle SLAM algorithms, with minimal additional computational cost.