Mapless Navigation Based on 2D LIDAR in Complex Unknown Environments

Area of Science:

• Robotics and Autonomous Systems
• Artificial Intelligence
• Sensor Data Processing

Background:

• Existing local path planners often fail in complex environments, leading to robots becoming trapped by obstacles.
• Current methods rely on mapping current sensor data for control, which has limitations in dynamic or unknown settings.

Discussion:

• The novel approach analyzes 2D LIDAR data to identify environmental gaps, utilizing these as sub-goals for navigation.
• It generates linear and angular velocities suitable for nonholonomic mobile robots.
• The method operates without global planners or environment maps, reducing computational load.

Key Insights:

• The proposed navigation strategy effectively prevents robots from getting stuck in complex, unmapped environments.
• By focusing on gap detection, the system offers a robust alternative to traditional path planning.

Outlook:

• This technique's low computational complexity and fast response are ideal for robots with limited processing power, potentially lowering manufacturing costs.
• Successful simulations and real-world tests on the Turtlebot2 platform validate the approach's effectiveness.