A comparative review on mobile robot path planning : classical or meta-heuristic methods?

TLDR: Dijkstra ’s Algorithm (DA) is considered a benchmark solution and Constricted Particle Swarm Optimization (CPSO) is found performing better than other meta-heuristic approaches in unknown environments.

About: This article is published in Annual Reviews in Control.The article was published on 2020-01-01 and is currently open access. It has received 92 citations till now. The article focuses on the topics: Motion planning & Dijkstra's algorithm.

Figures

Figure 15: Pie Chart of which region the robot has been during the SubExperiment 3 (Experiment II)

Figure 8: Pie Chart of which region the robot has been during the Experiment I

Figure 7: Region classification on where the robot has been during the Experiment I

Figure 16: Average performance difference from DA (optimal global path planner) in Sub-Experiments of Experiment II (Execution Time Factor)

Table 6: Summary of the First, Second, and Third Best Performing Algorithm in Each Experiments

Figure 11: Pie Chart of which region the robot has been during the SubExperiment 1 (Experiment II)

Frequently asked questions

Q1. What are the contributions in "A comparative review on mobile robot path planning : classical or meta-heuristic methods?" ?

This study explores the implementation of many meta-heuristic algorithms, e. g. Genetic Algorithm ( GA ), Differential Evolution ( DE ), Particle Swarm Optimization ( PSO ) and Cuckoo Search Algorithm ( CSA ) in multiple motion planning scenarios. The study provides comparison between multiple meta-heuristic approaches against a set of well-known conventional motion planning and navigation techniques such as Dijkstra ’ s Algorithm ( DA ), Probabilistic Road Map ( PRM ), Rapidly Random Tree ( RRT ) and Potential Field ( PF ). Several performance measures such as total travel time, number of collisions, travel distances, energy consumption and displacement errors are considered for assessing feasibility of the motion planning algorithms considered in the study. 

Q2. What are the classical algorithms for motion planning considered in this study?

The classical algorithms for motion planning considered in this study are Potential Field (PF), Dijkstra’s Algorithm (DA), Rapidly-explore Random Tree (RRT) and Probabilistic Road Map (PRM). 

Q3. What is the performing method in this category?

The worst performing method in this performance measure factor is RRT where all sub-experiments have at least 15% performance difference compared to DA. 

Q4. How many iterations did CPSO perform in this category?

CPSO closely followed with an average performance of 453.9 iterations and GA as the third best performing algorithm for this category with an average of 462.0 iterations. 

Q5. How long did DA take to complete the navigation task?

2. Execution Time (s): DA set the benchmark with an average execution time where it clocked only 87.55 seconds to complete the navigation task. 

Q6. What is the average convergence iteration of CPSO?

5. Convergence Iteration: CPSO outperformed other meta-heuristic methods in average convergence iterations performance factor with 370.0 iterations. 

Q7. What is the average execution time of a CPSO algorithm?

2. Execution Time (s): CPSO managed to outperform other algorithms including DA with 252.66 seconds on average across 10 executions. 

Q8. What is the performing variant of a PSO?

Within variants of PSO considered in this study, TVAC is found to be the worst performing variant since it has never achieved less than 5% performance difference from DA. 

Q9. What is the performing method in this experiment?

PF and DE are the second and third best performing methods in this experiment due to their consistent high performances in most categories across all three sub-experiments in this layout.