This paper presents a novel approach for navigation in complex and unknown environments. At present, existing local path planners whose control output is the mapping of current sensor data have been widely studied. However, these methods cannot really solve the problem of being trapped by obstacles. We analyzed the reasons and made improvements, and finally our approach can avoid being trapped in complex environments. The proposed method is based on 2D LIDAR. A central part of the approach is finding out gaps in the environment by analyzing sensor data. Then, we choose one of the gaps we find as the sub-goal. Linear and angular velocities are provided by the approach considering nonholonomic mobile robots. The method does not rely on global planners and environment maps. Therefore, it has the advantages of low computational complexity and fast response, which is of great significance to robots with low computing power; it will also help to reduce the manufacturing cost of robots. In addition, simulations and real tests were performed using the Turtlebot2 robotic platform. Successful results are achieved in both simulations and experimental tests.
Keywords: 2D LIDAR; Turtlebot2; complex environments; local path planners.