TWC-SLAM: Multi-Agent Cooperative SLAM with Text Semantics and WiFi Features Integration for Similar Indoor Environments

TL;DR

TWC-SLAM tackles the challenge of cooperative SLAM in indoor environments with repetitive architecture by fusing text semantics and WiFi fingerprints to improve cross-agent location identification and loop closure. The approach combines a FAST-LIO2-based front-end, a text/WiFi multimodal matching module, and a global mapping component that aligns sub-maps via ICP and least-squares optimization, producing globally consistent trajectories and maps. Key contributions include the integration of text semantics with WiFi features for robust same-location recognition, a multimodal loop closure strategy, and the demonstration of significant gains over point-cloud and single-modality baselines on complex indoor datasets. The results indicate substantial practical impact for multi-robot navigation in repetitive indoor spaces, enabling more reliable exploration, inspection, and mapping tasks in real-world scenarios.

Abstract

Multi-agent cooperative SLAM often encounters challenges in similar indoor environments characterized by repetitive structures, such as corridors and rooms. These challenges can lead to significant inaccuracies in shared location identification when employing point cloud-based techniques. To mitigate these issues, we introduce TWC-SLAM, a multi-agent cooperative SLAM framework that integrates text semantics and WiFi signal features to enhance location identification and loop closure detection. TWC-SLAM comprises a single-agent front-end odometry module based on FAST-LIO2, a location identification and loop closure detection module that leverages text semantics and WiFi features, and a global mapping module. The agents are equipped with sensors capable of capturing textual information and detecting WiFi signals. By correlating these data sources, TWC-SLAM establishes a common location, facilitating point cloud alignment across different agents' maps. Furthermore, the system employs loop closure detection and optimization modules to achieve global optimization and cohesive mapping. We evaluated our approach using an indoor dataset featuring similar corridors, rooms, and text signs. The results demonstrate that TWC-SLAM significantly improves the performance of cooperative SLAM systems in complex environments with repetitive architectural features.

Figures (6)

• Figure 1: Schematic diagram of text semantic matching and WiFi feature matching. Reliance on text semantic matching without complementary WiFi feature matching may induce localization inaccuracies. Specifically, two representative error patterns emerge: 1) In Agent 1's odometry trajectory, Text 2 could be erroneously associated with Text 1 as loop closure candidates; 2) Cross-agent confusion occurs when Text 2 captured by Agent 3's sensors is mismapped as Text 1. These erroneous associations propagate through the system, ultimately degrading global map consistency.
• Figure 2: The overview of our proposed TWC-SLAM system. The system comprises four main components: (1) Multi-Agent Input, (2) Text Semantic Matching, (3) WiFi Feature Matching, and (4) Global Mapping. Data is collected from diverse platforms, including handheld devices, wheeled robots, and legged robots, to enhance environmental adaptability.
• Figure 3: WiFi and text conditions in the experimental scenario. Figure (a) is an example of some WiFi information, while Figures (b) and (c) demonstrate identical text semantics positioned at distinct locations.
• Figure 4: Comparison of trajectory results for Scene #01. The blue odometry in (a) has completely deviated, and the odometry in (b) and (c) have also deviated.
• Figure 5: Comparison of trajectory results for Scene #02. The red odometry in (a) and (c) have completely deviated, and the odometry in (b) has deviated.