Factor Graphs + TagSLAM For Efficient Map and Pose Optimization

How do robots really know where they are — not just guess? In this video, we dive deep into the world of Factor Graphs, tagSLAM, and the PennCOSYVIO dataset, exploring how real-world SLAM systems are built, tested, and refined. Chapters: 0:00 Welcome 0:43 2D SLAM 6:52 SLAM As Bayes Net 9:50 SLAM As Factor Graph 11:43 PennCOSYVIO Dataset Case Study 13:30 Google Project Tango 15:30 SLAM Evaluation 16:55 SLAM Metrics 19:10 PennCOSYVIO Dataset 21:42 Camera Calibration Factor Graph 24:24 Pose Estimation Factor Graph 27:00 TagSLAM 27:42 iSAM2 (Incremental Smoothing and Mapping) 29:40 Libraries 30:17 Understanding Check We’ll cover: ✅ Bayesian Optimization for Non-linear Least Squares ✅ Factor Graphs and how they are general compared to Bayesian Optimization ✅ What factor graphs are and how they represent relationships between measurements, poses, and landmarks ✅ How optimization over these graphs refines both the map and robot motion over time ✅ Using trajectory evaluation tools (ATE, RPE) to compare estimated vs ground truth motion and measure SLAM performance ✅ The story behind the PennCOSYVIO dataset — how it was captured at the University of Pennsylvania using fiducial tags and multiple cameras ✅ How tagSLAM fuses visual data using factor graphs to estimate accurate trajectories This video is a case study in how research datasets are created and used — from experimental setup and calibration to multi-sensor fusion and optimization. You’ll see how factor graphs power real SLAM pipelines, turning noisy data into precise, robust localization. Whether you're new to robotics or an AI enthusiast, this video will give you a clear and fun introduction to the world of robots! 🔔 Subscribe for demystifying and deeper dives into perception, computer vision, AI, and robotics! 👍 Like this video if you enjoy learning about intelligent machines! 📩 Have questions? Drop them in the comments! #robotics #ai #computervision #automation #WhatIsARobot #technology #innovation #sensing #autonomy #artificialintelligence #embodiedintelligence #robot #robotics #computervision #SLAM #PennCOSYVIO #factorgraphs #tagSLAM #visualinertialodometry #trajectoryevaluation #perception #ai #robotvision #mapping #optimization #embodiedintelligence #artificialintelligence #visualSLAM #researchcase