Vision-based Perimeter Defense via Multiview Pose Estimation @KumarLabPenn
Vision-based Perimeter Defense via Multiview Pose Estimation  @KumarLabPenn
Uploaded September 2022 | Updated November 2026, 2 days ago
Paper: arxiv.org/abs/2209.12136

Abstract:
Previous studies in the perimeter defense game have largely focused on the fully observable setting where the true player states are known to all players. However, this is unrealistic for practical implementation since defenders may have to perceive the intruders and estimate their states. In this work, we study the perimeter defense game in a photo-realistic simulator and the real world, requiring defenders to estimate intruder states from vision. We train a deep machine learning-based system for intruder pose detection with domain randomization that aggregates multiple views to reduce state estimation errors and adapt the defensive strategy to account for this. We newly introduce performance metrics to evaluate the vision-based perimeter defense. Through extensive experiments, we show that our approach improves state estimation, and eventually, perimeter defense performance in both 1-defender-vs-1-intruder games, and 2-defenders-vs-1-intruder games.
Vision-based Perimeter Defense via Multiview Pose EstimationCoordinate-Free Dynamics and Differential Flatness of a Class of 6DOF Aerial ManipulatorsCPS: Frontier: Collaborative Research: bioCPS for Engineering Living CellsLearning Local Urban Wind Flow Fields From Range SensinghenrikSemi-Dense Visual-Inertial Odometry and Mapping for Quadrotors with SWAP ConstraintsDORA: Divergence-Oriented Data-Relay Algorithm for Partially Connected Robot TeamsAir-Ground Collaboration with SPOMP: Semantic Panoramic Online Mapping and PlanningMultiple Scarab Ground Robots Manipulating ObjectsRobust Stereo Visual Inertial Odometry for Fast Autonomous FlightTowards Optimizing a Convex Cover of Collision-Free Space for Trajectory GenerationFormation Control of a Swarm of Pico-Quadrotors
Vijay Kumar |

Vision-based Perimeter Defense via Multiview Pose Estimation

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER