ROS101

SLAM & Navigation

Best SLAM algorithm for a low-cost LIDAR setup?

Kwame Asante·· 5 comments
ROS2SLAMLIDAR

I've got an RPLIDAR A1 on a differential-drive base running ROS2 Jazzy. Trying to decide between slam_toolbox and Cartographer for a robot that'll mostly map single-floor indoor spaces under 200 sq meters. Any real-world experience with either on this exact sensor class?

5 comments

  1. Mira Kovač·Verified on 2026-07-09, RPLIDAR A1, diff-drive base, ROS2 Jazzy

    For under 200 sqm indoors, slam_toolbox is the easier starting point — it's what ships in the default Nav2 tutorials and has synchronous mode which is much more forgiving to tune than Cartographer's parameter set.

  2. Andy Chen·✓ Accepted answerVerified on 2026-07-09, RPLIDAR A1, ROS2 Jazzy

    Agreed, and to add the specific reason: slam_toolbox's lifelong mapping mode also lets you keep extending the same map across multiple sessions without a full remap, which matters a lot once you've got a 200 sqm space — you won't want to redo the whole thing every time. Cartographer is better if you specifically need multi-floor/multi-trajectory merging, which it sounds like you don't.

  3. Dev Okafor·Verified on 2026-07-09, RPLIDAR A2, ROS2 Jazzy

    This matches what I saw too. One gotcha: the default max_laser_range in slam_toolbox's params assumes a 12m sensor (A1). If anyone reading this has an A2/A3, bump that parameter or you'll silently lose range.

  4. Dev Okafor·

    Different angle: have you looked at your scan rate vs robot speed? At the A1's ~5.5Hz scan rate, if your base can move faster than about 0.5 m/s you'll see smearing at corners regardless of which SLAM backend you pick.

  5. Kwame Asante·

    Good catch, I capped our max linear velocity at 0.4 m/s for exactly this reason. Worth its own troubleshooting note honestly.