ALATS: Analysis of localization algorithms in terrestrial surveillance bots

Effective surveillance is crucial for optimal security and autonomy. It has led to the development of various surveillance systems, such as smart cameras, bots, and drones, with diverse applications. To achieve an efficient surveillance system for multiple use cases, a proposed solution involves the creation of a terrestrial bot. This Surveillance Bot is a mobile unit designed to monitor the environment, detect objects (including human intruders) using smart cameras, and provide crucial data like temperature and humidity for remote surveillance. The bot employs simultaneous localization and mapping (SLAM) techniques, specifically the bug2 algorithm, to navigate autonomously and avoid obstacles. Object detection and human intruder detection algorithms are implemented, considering factors like accuracy, video feed lag, and algorithm robustness. Experimental results demonstrate the bot’s high accuracy in object detection (mean average precision of 33.7) and its frame rate performance. For continuous video recording and analysis, the bot utilizes the YOLOv7 real-time object identification algorithm, enhancing its surveillance capabilities by enabling real-time tracking of identified objects. The combination of YOLOv7 with the bot’s autonomous movement and mapping abilities creates a powerful and efficient surveillance solution. The proposed work aims to develop a smart surveillance robot suitable for military and civil applications, ensuring comprehensive surveillance.