Design and Implementation of an Intelligent Security Patrol Robot with Nighttime Dynamic Object Detection Functionality

Current intelligent security robots on the market capture images of relevant objects in patrol areas during inspections. However, they do not conduct real-time analysis of security events or detect dynamic object behaviors, making them challenging to use in nighttime environments. Based on this, this paper designs an intelligent security patrol robot with nighttime dynamic object detection capabilities. This robot can perform patrol tasks in nighttime environments while detecting and analyzing the dynamic behavior of pedestrians. In terms of hardware design, a four-wheel differential robot was designed; it uses a 3D LiDAR to obtain point cloud data, an IMU module to gather robot posture information, and an infrared camera to capture nighttime infrared images. The robot also has an embedded processor, providing a computing platform. In software design, based on the four-wheel differential chassis, the kinematic model of the robot was first constructed. Then, the robot's autonomous navigation functions were designed, focusing on localization and mapping, as well as integrated path planning algorithms. For nighttime detection, a dynamic object behavior detection algorithm was proposed. The algorithm was trained using a custom-built dataset, enabling the robot to perform visual security tasks in nighttime environments. Experimental results demonstrate that the robot designed in this paper can detect dynamic behavior objects in real-time during nighttime patrol tasks and maintain efficient security operations in low-light and no-light environments.