A Proprioceptive-Only Benchmark for Quadruped State Estimation: ATE, RPE, and Runtime Trade-offs Between Filters and Smoothers
Ylenia Nisticò, João Carlos Virgolino Soares, Joan Solà, Claudio Semini
THE PROBLEM
This paper focuses on Perception & SensingSensorA device that provides information about the robot or its environment. fusion. This paper benchmarks three Perception & SensingState estimationCombining noisy sensor data to estimate the robot’s true state. algorithms (MUSE, IEKF, Invariant Smoother) for quadruped robots using only proprioceptive sensors, giving you concrete accuracy vs. Simulation & Sim-to-RealLatencyDelay between input, computation, and action. trade-offs to pick the right filter for your legged Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions.'s compute constraints. The study shows IEKF and IS are more accurate long-term while all three perform similarly short-term, with reproducible code released open-source. Read the paper by tracking the Core ConceptsTaskThe job the robot is supposed to complete, such as pick-and-place, navigation, or drawer opening. definition, the Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. or data assumptions, and the evidence that supports the claimed improvement.
HOW IT WORKS
Task framing
Core method
Data and supervision
Evaluation evidence
FIGURES
KEY RESULTS
This paper benchmarks three Perception & SensingState estimationCombining noisy sensor data to estimate the robot’s true state. algorithms (MUSE, IEKF, Invariant Smoother) for quadruped robots using only proprioceptive sensors, giving you concrete accuracy vs. Simulation & Sim-to-RealLatencyDelay between input, computation, and action. trade-offs to pick the right filter for your legged Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions.'s compute constraints. The study shows IEKF and IS are more accurate long-term while all three perform similarly short-term, with reproducible code released open-source.
WHY DEVELOPERS SHOULD CARE
This paper benchmarks three Perception & SensingState estimationCombining noisy sensor data to estimate the robot’s true state. algorithms (MUSE, IEKF, Invariant Smoother) for quadruped robots using only proprioceptive sensors, giving you concrete accuracy vs. Simulation & Sim-to-RealLatencyDelay between input, computation, and action. trade-offs to pick the right filter for your legged Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions.'s compute constraints. The study shows IEKF and IS are more accurate long-term while all three perform similarly short-term, with reproducible code released open-source.
LIMITATIONS
The main limitation to check is whether the claimed behavior holds outside the paper's reported setup. That means testing across different Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. embodiments, scenes, objects, and data distributions.
WHAT COMES NEXT
The practical next step is independent reproduction with clear baselines, ablations, and stress tests. For a developer, the useful follow-up is to map the paper's Perception & SensingSensorA device that provides information about the robot or its environment. fusion assumptions onto a concrete Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. stack, then test the smallest version of the method that could run end to end.