Real-time and Cost-efficient Indoor Localization and Mapping Solution for Emergency Response Applications

Abstract. In GNSS-denied environments such as indoor and underground locations, real-time mapping and navigation are paramount, especially for emergency response scenarios like search and rescue or building evacuations. First responders rely on accurate maps of entrances and exits to prevent disorientation. This research addresses the challenges posed by the absence of GNSS signals and the need for high-performance data processing by developing a cost-effective multi-sensor system for real-time indoor mapping. The system incorporates consumer-grade technologies, including 2D LiDAR, RGB-D cameras, and MEMS-based IMU. It utilizes the Google Cartographer SLAM engine to generate real-time 2D raster maps and stream position and orientation data at 200 poses per second, ensuring continuous poses feed while minimizing latency. The resulting maps are continuously being optimized using a loop-closing algorithm to reduce drift and maintain the trajectory integrity. The system was validated in both real-time and post-processing configurations, the proposed system demonstrates promising consistency and repeatability according to the verification results furnished in this research paper. The data collection and processing times with this system are significantly reduced compared to static LiDAR systems, without the need for high-end computing resources. This research underscores the benefits of using inexpensive sensors and components, providing an efficient solution for indoor mapping and navigation in emergency scenarios.