Differential Barometric Altimetry for Submeter Vertical Localization and Floor Recognition Indoors

TL;DR

This work tackles indoor multi-floor localization by leveraging differential barometric sensing with a networked base station, implemented as a ROS-compatible module that publishes base-referenced altitude alongside drift-free mobile readings. Relative altitude is computed by comparing heights derived from pressure and temperature measurements from both base and mobile barometers, enabling precise floor identification through nearest-floor matching. In experiments conducted in a four-story stairwell and an enclosed elevator shaft, the barometer pipeline achieved a sub-meter RMSE of 0.29 m and perfect 100% floor recognition, outperforming visual SLAM, LiDAR odometry, and Wi‑Fi fingerprinting under the same conditions. The results demonstrate a practical, low-cost solution for robust vertical awareness in real-world robotic deployments, with the implementation openly released as open-source software for reproducibility and integration.

Abstract

Accurate altitude estimation and reliable floor recognition are critical for mobile robot localization and navigation within complex multi-storey environments. In this paper, we present a robust, low-cost vertical estimation framework leveraging differential barometric sensing integrated within a fully ROS-compliant software package. Our system simultaneously publishes real-time altitude data from both a stationary base station and a mobile sensor, enabling precise and drift-free vertical localization. Empirical evaluations conducted in challenging scenarios -- such as fully enclosed stairwells and elevators, demonstrate that our proposed barometric pipeline achieves sub-meter vertical accuracy (RMSE: 0.29 m) and perfect (100%) floor-level identification. In contrast, our results confirm that standalone height estimates, obtained solely from visual- or LiDAR-based SLAM odometry, are insufficient for reliable vertical localization. The proposed ROS-compatible barometric module thus provides a practical and cost-effective solution for robust vertical awareness in real-world robotic deployments. The implementation of our method is released as open source at https://github.com/witsir/differential-barometric.

Figures (5)

• Figure 1: The custom-designed SLAM device used in our experiments.
• Figure 2: The data pipeline of the differential barometer ROS package.
• Figure 3: Schematic of the experimental environment. The data collection trajectory covered a full ascent through the multi-story stairwell and a subsequent descent in the elevator.
• Figure 4: 3D point cloud with trajectory generated by GLIM.
• Figure 5: Altitude estimation traces in stair (CP1–CP7) and elevator (CP7–CP11) scenarios.