Paper Contents
Abstract
Guided robots are the result of joint exploration and research by multiple disciplines, playing an important role in the process of realizing intelligent public services. However, current guided robots still face problems such as inaccurate dynamic environment guidance, unstable interaction systems, and difficulty balancing hardware costs and algorithm efficiency. To address these issues, we have built a simulation system on the ROS platform and verified it with physical objects to address these problems.The system adopts multi-sensor data fusion and optimized path planning algorithms, focusing on overcoming the challenges of dynamic obstacle avoidance and environmental modeling. This solution adopts a system framework of "ROS2+Gazebo+Raviz" combination, which consists of four functional layers: perception, decision-making, execution, and interaction. The perception layer uses the Gazebo physics engine to simulate the laser radar, depth camera, and inertial sensor, in order to realistically reproduce the physical environment such as friction and lighting; The decision-making layer integrates the improved SLAM algorithm with the navigation system to implement two-level path planning. The global planning adopts a combination of semantic map feature extraction and improved shortest path algorithm, while the local planning uses dynamic window method to achieve rapid obstacle avoidance, ensuring that the system response time is controlled within 0.5 seconds; The execution layer adopts a differential drive model and a proportional integral controller to simulate the actual motion performance; The interaction layer integrates visual monitoring and voice touch screen operation functions.
Copyright
Copyright © 2025 Jiarui Cao, Wenyi Li. This is an open access article distributed under the Creative Commons Attribution License.
