Comparison of ROS Local Planners with Differential Drive Heavy Robotic System

A Comparison Study between Traditional and Deep-Reinforcement-Learning-Based Algorithms for Indoor Autonomous Navigation in Dynamic Scenarios

“…In this context, we selected DWA and TEB as the conventional local-planning methods of choice. This choice stemmed from the fact that DWA has been widely used as a benchmark in prior studies [ 52 , 53 ], serving as a reference point for evaluating and understanding the improvements and challenges offered by novel approaches. Conversely, TEB, also featured in previous comparison studies [ 52 , 53 ], has emerged as a significant enhancer of traditional-navigation algorithms, particularly in dynamic and non-stationary environments.…”

“…This choice stemmed from the fact that DWA has been widely used as a benchmark in prior studies [ 52 , 53 ], serving as a reference point for evaluating and understanding the improvements and challenges offered by novel approaches. Conversely, TEB, also featured in previous comparison studies [ 52 , 53 ], has emerged as a significant enhancer of traditional-navigation algorithms, particularly in dynamic and non-stationary environments. Furthermore, TEB demonstrates superior computational efficiency compared to DWA [ 54 ], making it a suitable candidate for comparison to the most robust RL approaches available.…”

Benchmarking Off-the-Shelf Human-Aware Robot Navigation Solutions

“…When it comes to evaluating robotic navigation methods, there are several performance metrics. For example, Cybulski et al [42] and Naotunna and Wongratanaphisan [43] have evaluated different traditional navigation methods for ROS. They used metrics such as time, robot path length, minimum distance between the robot and its target to be able to quantify the effectiveness of a robotic navigation method.…”

“…As used by [42] and [43], among the classic metrics, we can cite the length of the path, the time required, the average speed and acceleration, whether the robot has reached its goal, the final distance between the robot and the goal, and finally the number of collisions. In addition to these very common metrics, we can find metrics like path efficiency and path irregularity.…”

Solving Optimal Path Planning Problem of an Intelligent Mobile Robot in Dynamic Environment Using Bidirectional Rapidly-exploring Random Tree Star-Dynamic Window Approach (BRRT*-DWA) with Adaptive Monte Carlo Localization (AMCL)

“…Fox et al [28] proposed the Dynamic Window Approach to collision avoidance. Other authors proposed DWA for different applications [13][14][15]. Ferrer Sánchez J [16], and Looi et al [10], compare the local planners (controllers) Time Elastic Band (TEB), Model Predictive Control (MPC), and Dynamic Window Approach (DWA), and give the insight and applicability of each method which led to the choosing of DWA to our system.…”