Research on Improving TEB Algorithm for Mobile Robot Path Planning

To address the issue of unstable velocity output in the TEB algorithm during multi-point navigation, this paper introduces the TEB-confidence algorithm. This algorithm resolves the problem by incorporating buffer and expansion zones around the target points, which impose velocity constraints as the points are approached. Moreover, the Chebyshev series method is utilized to accurately fit the decay function, erf(x), enhancing the precision of the fitting process. Experimental evaluations were conducted using the Robot Operating System (ROS) with an Ackermann motion model robot. The experiments involved navigating through seven target points using both the TEB-confidence algorithm and the classical TEB algorithm. The TEB-confidence algorithm successfully completed the multi-point navigation task in 22.5 s, whereas the classical TEB algorithm required 25.9 s. This demonstrates a significant 15.11% reduction in execution time when using the TEB-confidence algorithm compared to the classical TEB algorithm. Additionally, during the experiments, the classical TEB algorithm resulted in a collision, leading to a complete stop. The experimental results provide solid evidence that the enhanced TEB-confidence algorithm effectively prevents collisions resulting from velocity instability.