Time Synchronization in GNSS/INS Systems | Positioning Engine Deep Dive (Part 1)
Автор: Point One Navigation
Загружено: 2026-03-17
Просмотров: 593
Описание:
What's actually happening inside a sophisticated GNSS/INS positioning system? There's a lot more than just "GPS + IMU = position."
In Part 1 of this technical deep-dive series, Adam Shapiro (Navigation Software Lead, Point One Navigation) and Tom Weeks (COO) walk through the sensor fusion pipeline that transforms raw sensor data into centimeter-accurate position estimates multiple times per second—including the critical infrastructure that most teams building positioning systems get wrong.
CHAPTERS
00:00 Introduction: what's inside a GNSS/INS navigation system
00:45 Input bus: sensor integration, logging, and adapters
03:32 Synchronized Time: why P1 time exists (and why most sensor fusion systems get this wrong)
07:20 The parking garage problem: when syncing GPS time can fail
08:40 IMU pre-processing: coarse alignment and downsampling
10:30 Sensor data rates for GNSS, IMUs, & Computer Vision - equipment & filtering decisions
13:15 GNSS pre-processing: multipath detection, environment detection, and cycle slip
15:35 The 19-centimeter problem: carrier phase ambiguity & cycle slip in RTK systems
17:14 Navigation engine: Time alignment: reordering out-of-sequence measurements
18:15 Variable latency: why GNSS measurements can arrive 120ms late
20:00 Inside the navigation filter: estimator, dynamics model, and measurement models (preview for Part 2)
WHY THIS MATTERS FOR GNSS/INS SYSTEM DESIGN
Building a production-grade sensor fusion system requires more than implementing a Kalman filter. The infrastructure—time synchronization, pre-processing, sensor error detection—is what separates research code from systems that work reliably in autonomous vehicles, drones, construction equipment, and agricultural machinery.
This session covers the input pipeline, time synchronization, and pre-processing stages of a production GNSS/INS system—the foundation of how modern INS (Inertial Navigation System) software handles multi-sensor integration for autonomous navigation, robotics localization, and precision positioning applications.
TOPICS COVERED
-Sensor integration & logging: Why logging raw sensor data verbatim is critical for debugging GNSS/INS systems
-Sensor adapters: Handling different GNSS receiver and IMU brands with a common downstream format
-Time synchronization (P1 time): The synthetic time base that unifies GNSS, IMU, wheel odometry, and camera data without GPS dependency
-The #1 mistake: Why time synchronization is what most teams do wrong when building sensor fusion systems
-Variable latency in sensor fusion: Why measurements arrive 10-120ms after capture and how to handle it
-IMU integration: Coarse alignment, orientation-agnostic installation, and high-rate data processing
-Downsampling strategies: Filtering engine vibration (3,000Hz → 100Hz) without losing critical dynamics
-GNSS signal processing: Multipath detection for urban environments and challenging GNSS conditions
-Environment detection: Inferring context (urban canyon, open sky, tunnel) from signal quality metrics
Carrier phase processing: Cycle slip detection to prevent 19cm errors in RTK positioning
-Time alignment in navigation filters: Putting asynchronous, out-of-order measurements into correct sequence for Kalman filter processing
-Plus: A preview of the navigation filter architecture (Extended Kalman Filter, measurement models, and RTK ambiguity resolution) covered in Parts 2 and 3.
🚀 Get started with Point One's Precise Navigation platform — hardware, software, and RTK network:
Free Trial: https://app.pointonenav.com/trial
Technical Questions: https://pointonenav.com/contact-us
Coming in Part 2: Measurement modeling, RTK corrections integration, and carrier phase ambiguity resolution—how GNSS/INS systems achieve centimeter-level accuracy with Real-Time Kinematic positioning.
Coming in Part 3: The Extended Kalman Filter estimator, platform-specific dynamics models, sensor error learning, and multi-protocol outputs—how navigation systems adapt to different vehicles and improve over time.
Related topics: GNSS/INS integration, sensor fusion for autonomous navigation, RTK positioning systems, Kalman filter implementation, INS architecture, IMU integration, localization systems, state estimation, navigation filter design
Повторяем попытку...
Доступные форматы для скачивания:
Скачать видео
-
Информация по загрузке: