Research on Navigation and Stability Optimization Methods for Snowplow Robots Based on Multi-Sensor Fusion

The snowplow robot plays an essential role in maintaining roads during winter, ensuring that essential routes remain passable. However, it faces a series of challenges in snowy environments, such as slippage due to low friction, visual sensor interference caused by snowflakes during extreme weather, and reduced stability from uneven snow depths. To overcome these obstacles, we first develop a SLAM system that integrates LiDAR, IMU, and depth cameras, with multi-sensor data fusion carried out using an Extended Kalman Filter (EKF). Next, we enhance the <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$A^{*}$</tex> global path planning algorithm by incorporating dynamic weights based on snow depth into the cost function, enabling the robot to prioritize paths with less snow or areas that have already been cleared. Experimental results demonstrate that combining multi-sensor fusion with dynamic weight adjustment significantly improves both SLAM accuracy and the stability of path planning, effectively reducing slippage and enhancing the robot's performance in challenging snowy conditions.