Fusion-SLAM™ is our fully proprietary and patented core technology. It originated from fundamental research in visual perception and deep learning, led by our co-founder, Dr. Xuanhao Zhou. Through over a decade of dedicated in-house R&D, the technology has been rigorously refined and validated across hundreds of real-world deployments, most notably serving as the core positioning engine for AMRs and automated forklifts within our parent company, Lanxin Robotics.

This same, field-proven core now powers the MRDVS V2 Pro module. It achieves tightly-coupled sensor fusion of 2D LiDAR, visual ceiling features, and IMU data. Powered by our domain-trained deep learning models, it delivers centimeter-level absolute positioning that is significantly more stable and reliable than single-sensor solutions.

We have now productized this industrial-grade robustness into the V2 Pro, a self-contained hardware module, making it directly available for retrofitting manual forklifts and enabling AGV conversions. This brings proven, reliable autonomy to a wider range of vehicles and industrial applications.

The algorithm is not open-sourced and remains a key part of our intellectual property portfolio, ensuring continuous innovation and reliable performance for our customers.

Our solution offers three fundamental advantages over conventional systems:

1. Marker-Free Core & Rapid Deployment: At its core, the system requires no pre-installed markers, beacons, or infrastructure (e.g., UWB anchors, reflector walls), and enabling deployment in days rather than weeks. For scenarios where customers prefer an additional safeguard for rapid re-localization, a limited number of QR codes can be optionally deployed without compromising the system’s fundamental infrastructure-free advantage.
   
   
2. Radically Different & Scalable Cost Model: Unlike UWB systems whose cost scales with warehouse area (requiring extensive anchor networks that can cost hundreds of thousands of USD for large sites), our solution scales primarily with the number of vehicles. This translates to dramatically lower upfront investment and far more predictable, linear cost scaling for large-scale or expanding operations.
3. Proven Robustness in LiDAR-Degraded Environments: Unlike systems reliant on 2D LiDAR or VSLAM alone, our Fusion-SLAM™ technology ensures stable, centimeter-accurate operation in challenging conditions – including steel structures, open spaces, dynamic storage floors, and high-traffic aisles – through tight fusion of ceiling vision and inertial data.

No. The system uses absolute SLAM based on natural features, not relative wheel odometry. This eliminates cumulative drift entirely. There is no need for periodic manual correction – it provides stable, drift-free positioning in the global map.

The system is designed for optimal performance within ceiling heights up to 15 meters, which is the effective range for the visual component to ensure robust, feature-rich localization. Within this range, positioning accuracy is ±1-3 cm and remains within ±4 cm at 15 meters.

The integrated 2D LiDAR supports this with a maximum detection range of up to 40 meters for high-reflectivity (90%) targets, and about 15 meters for standard low-reflectivity (10%) targets, providing reliable planar mapping data as part of the Fusion-SLAM™ solution.

A comprehensive set of hardware parameters and environmental tolerances is provided in Appendix A: V2 Pro Hardware Specification.