How SLAM Powers Modern Robot Vacuums: From Sparse Maps to Vision‑Based Navigation

Introduction

In early 2018, 360 launched a robot vacuum that uses SLAM (Simultaneous Localization and Mapping) technology to automatically build maps of unknown environments and navigate intelligently. SLAM is widely applied in autonomous vehicles, robots, and AR/VR devices.

What is SLAM?

SLAM enables a device to determine its position while constructing a map of the surrounding area. It is essential for AR glasses to overlay virtual objects correctly and for robots to understand and move within their environment.

Two Forms of SLAM

Sparse SLAM creates maps containing only a limited set of environmental points, suitable for applications such as rocket or submarine navigation.

Dense SLAM generates maps with complete environmental information, used in AR/VR and advanced robotics.

Robot Vacuum Solutions

The article compares four navigation schemes:

ESLAM (Blindfolded Elephant) – combines wheel odometry and IMU; low cost but prone to missed areas.

LSLAM (Narrow‑eye) – uses a single‑line laser scanner; good in narrow passages and dense obstacle scenes.

VSLAM (Vision‑based) – relies on a camera to build a feature‑sparse map; offers larger field of view but struggles in low‑light or feature‑poor environments.

Hybrid approaches – newer models integrate laser and camera sensors for improved performance.

Testing and Results

Tests were conducted in a 9 m² conference room with three scenarios (simple, medium, complex). Results show:

VSLAM and laser SLAM perform similarly in simple scenes.

Laser SLAM outperforms vision in medium and complex scenes with many obstacles.

ESLAM lacks mapping capability and has poor path planning.

Market Analysis

The domestic robot‑vacuum market exceeded ¥5 billion in 2017, with rapid growth and high profit margins. 360’s laser‑based LSLAM solution offers higher efficiency and user satisfaction compared with competitors such as Xiaomi and iRobot.

Conclusion

Among the three schemes, LSLAM provides the best balance for narrow passages and dense obstacles, making it the preferred choice for 360’s current robot vacuum models.