Comparative Analyses of ROS Local Planners for Quadrupedal Locomotion: A Study in Real and Simulated Environments

Advancements in hardware, computation, and algorithms are driving a revolution in the field of mobile robots. This study compares the ROS-based local planner navigation algorithms Dynamic Window Approach (DWA), Timed Elastic Band (TEB), and Trajectory Rollout for application in legged locomotion. The tests were performed in simulation and on a real quadruped robot using a small maze-type environment. The experiments measured the total time and distance achieved for each local planner. The results reveal a comparable performance among navigation algorithms in real and simulated environments. Specifically, the Timed Elastic Band (TEB) algorithm demonstrates superior performance, exhibiting an average traversal duration of 31.3 s in simulated environments and 30.8 s in real-world conditions, coupled with an average traversal distance of 14.5 m (simulated) and 13.7 m (real-world).