Navigation & Mapping

Navigation failures — robots spinning in circles, running in straight lines, ignoring rooms, or refusing to start — almost always trace back to one of three systems: the primary navigation sensor (LiDAR or camera), the wheel encoders, or the floor-contact bumper sensors. Knowing which navigation technology your robot uses narrows the cause immediately.

LiDAR-based robots (Roborock, Dreame, Xiaomi, most Ecovacs) use a spinning laser distance sensor on top of the unit. If that turret is obstructed by pet hair, dust buildup, or a low-clearance surface, the robot cannot build a map and defaults to random-bounce cleaning. A simple visual check — spin the turret by hand; it should rotate freely with no resistance — and a wipe with a dry cloth usually resolves the issue.

Camera-SLAM robots (most Roomba i/j/s-series, Shark AI, some Eufy) navigate by recognising visual landmarks. Bright light conditions, reflective floors, or very dark rooms all disrupt this. They also require a consistent home environment: if you rearrange furniture, some models need a full remap before they clean efficiently again.

Random-bounce models (older Roombas, entry-level Eufy) use infrared cliff sensors and physical bumpers. These are simpler but fail when bumpers stick, cliff sensors see dark floor patterns as drops, or the robot loses track of its charging dock location. For these models, consistent dock placement and clean sensors are the entire maintenance requirement.

Wheel encoder errors are often reported as "Error 2" or "side wheel stuck" across multiple brands. If the robot does not have visible wheel obstructions, the encoder (a small optical sensor inside the wheel assembly) may be dirty or failing. Partial cleaning of accessible wheel components resolves about 60% of these cases without disassembly.