Factor Graph
A probabilistic graphical model composed of alternating variable nodes and factor nodes, representing the factorization structure of a multivariate joint probability distribution. Factor graphs provide the foundational mathematical framework for belief propagation algorithms.
Timeline (last 90 days)
Factor graphs represent all historical observations and motion constraints as factor nodes in a graph, jointly solving for the optimal state trajectory through nonlinear least-squares optimization, with GTSAM and Ceres Solver being the most commonly used open-source libraries.
Both EKF and factor graphs assume that the timestamps of input data are accurate, and time offsets will be treated as spatial errors by these algorithms.
All Facts (2)
Factor graphs represent all historical observations and motion constraints as factor nodes in a graph, jointly solving for the optimal state trajectory through nonlinear least-squares optimization, with GTSAM and Ceres Solver being the most commonly used open-source libraries.
50%UnverifiedBoth EKF and factor graphs assume that the timestamps of input data are accurate, and time offsets will be treated as spatial errors by these algorithms.
50%