SentiSync
The Challenge
Name:
Synchronization of multi-modal sensor data
Domain:
Sensor synchronization, sensor fusion, (Time-Sensitive) Networking
Challenge proposer:
Objectives:
To develop a prototype that integrates a hardware solution for accurate time synchronization of multi-sensor data on an AUTOASSESS drone.
Challenge Addressed:
Multi-sensor robotic systems suffer from temporal misalignment: each sensor timestamps data using its own clock or, worse, the host operating system’s scheduler. For a 500 Hz IMU, even microseconds of jitter compound into centimetres of positional error. SentiSync addressed this by replacing software-based timestamping with deterministic hardware timing. The project also exposed a hidden culprit: Linux embedded systems reading high-frequency IMU data showed severe jitter and packet loss — completely invisible to the user until you look at the delta-time plots.
The Solution
Description:
SentiSync tackled one of robotics’ most underrated headaches: sensors that can’t agree on what time it is. When a LiDAR and an IMU are even slightly out of sync, the resulting maps can look like they were drawn during an earthquake. SentiSystems plugged in their SentiBoard 2 — a hardware timing device with 10 ns precision — and showed that getting sensors to share a common clock isn’t a luxury, it’s a necessity. The project was part of AUTOASSESS Challenge 5, targeting autonomous marine vessel inspection.
Key Results Delivered:
The SentiBoard 2 delivered hardware-synchronized timestamps across LiDAR, IMU, and camera interfaces, distributing a common clock via IEEE 1588v2 PTP to all networked devices. Across 25 evaluation runs, hardware synchronization improved LiDAR–IMU calibration accuracy by 29–35% and rotational stability by ~70% compared to software timestamping. Point cloud deskewing produced sharper, more linear reconstructions. ICP fitness scores were consistently higher for hardware-synced data. The software-only time synchronized Autopilot IMU showed 44.1 cm calibration error — more than five times worse than the best hardware-synced result.
Unique Value and Impact:
SentiSync proves that hardware synchronization is not an optimization — it’s a prerequisite for precision multi-sensor robotics. The SentiBoard 2 is a compact, reconfigurable solution that plugs into any sensor stack (IMUs, LiDARs, DVLs, GNSS, cameras) and ensures every measurement carries a trustworthy timestamp. For the maritime inspection sector, this translates directly into more accurate 3D maps, fewer failed missions, and reduced operational risk. The 70% improvement in rotational stability alone could mean the difference between a robot that maps a ship’s hull reliably and one that delivers a blurry mess.
Solution in Action:
Picture an autonomous drone dispatched to inspect the interior of a ship’s ballast tank — dark, confined, and full of rust and rivets. The drone carries a LiDAR and three IMUs, all ticking to different clocks without SentiSync. The resulting 3D map looks like abstract art: ghosted walls, curved edges, and structures that exist in two places at once. Now plug in the SentiBoard 2. Every sensor pulse is timestamped in hardware to within 10 nanoseconds. The map snaps into focus — crisp walls, sharp corners, reliable geometry. Shipping companies, port authorities, and inspection OEMs can use this to replace dangerous human entry with autonomous survey missions.
The Solution Provider
SentiSystems AS develops cutting-edge synchronisation and sensor fusion technology to enhance autonomy, navigation, and situational awareness. With over a decade of combined research and development, the company’s core innovation, the SentiBoard, enables high-precision timing and unified data handling across sensor platforms. SentiSystems provides scalable solutions that empower robotics, drones, and marine systems to operate with improved precision, safety, and efficiency.
Key Team Members:
The AUTOASSESS Experience
Main contribution of the AUTOASSESS programme:
AUTOASSESS gave SentiSystems the structure, funding, and validation environment to take SentiBoard from a promising prototype to a benchmark-tested solution. The programme’s challenge format pushed the team to define real KPIs, collect rigorous datasets, and compare results head-to-head. Being announced as Open Call #1 winner provided market credibility and opened doors to the maritime robotics community. The consortium partners provided an inspiring peer network and access to ship inspection use cases that shaped the project’s commercial direction.
Testimonial
“We knew hardware synchronization mattered — AUTOASSESS made us prove it. The programme forced us to quantify something the robotics world often hand-waves, and the numbers were more compelling than we expected. A 70% improvement in rotational stability is not a footnote; it’s a headline. We came in with a cool device and left with a validated solution. If you’re building autonomous inspection systems and you’re still relying on software timestamps, call us.”
Frederik Leira, Project Lead