Structure from WiFi (SfW): RSSI-based Geometric Mapping of Indoor Environments

TL;DR

The paper tackles indoor SLAM using WiFi RSSI to recover 2D geometry without cameras or lidar. It introduces Structure from WiFi (SfW) and the inverse $k$-visibility framework to infer free and occupied spaces from RSSI and robot trajectories. The method includes a dense inverse $k$-visibility variant and a sparse one using RSSI-derived $k$-values, with probabilistic occupancy modeling. Experiments in simulation and on real robots demonstrate the ability to recover substantial portions of indoor free space and detect walls, highlighting potential for privacy-preserving, sensor-light SLAM.

Abstract

With the rising prominence of WiFi in common spaces, efforts have been made in the robotics community to take advantage of this fact by incorporating WiFi signal measurements in indoor SLAM (Simultaneous Localization and Mapping) systems. SLAM is essential in a wide range of applications, especially in the control of autonomous robots. This paper describes recent work in the development of WiFi-based localization and addresses the challenges currently faced in achieving WiFi-based geometric mapping. Inspired by the field of research into k-visibility, this paper presents the concept of inverse k-visibility and proposes a novel algorithm that allows robots to build a map of the free space of an unknown environment, essential for planning, navigation, and avoiding obstacles. Experiments performed in simulated and real-world environments demonstrate the effectiveness of the proposed algorithm.

Figures (8)

• Figure 1: Demonstration of $k$-visibility where $k=0$ (red), $k=1$ (green), $k=2$ (blue) and $k=3$ (yellow) are shown. $k$-visibility refers to the number of times a signal from a reference point (e.g. a router, shown in dark blue) passes through a wall/obstacle when making a straight-line path to a desired location.
• Figure 2: Diagram demonstrating the ray-drawing principle upon which the inverse $k$-visibility algorithm is based. The wall is located at the exact coincidence of two consecutive $k$-value regions along a ray cast from the router point. Note that only the region of interest in the $k=1$ region is shown.
• Figure 3: Ray-drawing for an arbitrary trajectory coordinate $T_i$ which has an associated $k$-value $k_i$. By definition of $k$-visibility, the ray $\overline{RT_i}$ must have a $k_i$ number of walls along the ray.
• Figure 4: Visual demonstration of the geometric rules with $k=0$ and $k=1$ area. Rule 1 was excluded from the legend for clear visualization as it shows the trajectory of the robot.
• Figure 5: Initial wall estimate along a ray (left). Improved wall estimate along a ray after updating the lower and upper endpoints (right).