| Summary: | Security and surveillance play a crucial role in maintaining the safety and integrity of the building surroundings. Conventional human security patrolling is subject to human error and limited by factors such as fatigue and inattentiveness. To tackle this challenge, a robot patrolling with 2D visual algorithm is proposed for collecting data from the IMU, LiDAR, and Raspberry Pi Camera to safely maneuver within the targeted environment. A hybrid waypoint and Rapidly-Exploring Random Tree (RRT) navigation algorithm is developed using the ROS platform. The Histogram of Oriented Gradients (HOG) descriptor is integrated into the security patrolling robot to detect the presence of humans. Simulation and real-world testing have been conducted to evaluate the effectiveness and reliability of the developed system. As a result, the navigation system achieved 100% and 90% success rate in simulations and real-world environment tests respectively with the ability to detect humans with 95% accuracy in low visibility (nighttime) environments. The tree iteration value of 2000 was the best value to achieve the consistency of RRT navigation. The robot achieved maximum velocity of 0.26 m/s in obstacle free environment while the velocities to navigate with obstacles range between 0.1 m/s to 0.15 m/s. © 2024 IEEE.
|
|---|