Have a personal or library account? Click to login
Computationally Efficient Dynamic Window Approach Based on Pattern Search Optimization Cover

Computationally Efficient Dynamic Window Approach Based on Pattern Search Optimization

Open Access
|Dec 2025

Figures & Tables

Fig. 1.

Visualisation of the DWA procedure for a single linear velocity
Visualisation of the DWA procedure for a single linear velocity

Fig. 2.

Example trajectories of the AGV obtained for the original DWA and the proposed PSDWA
Example trajectories of the AGV obtained for the original DWA and the proposed PSDWA

Comparison of quality indicators obtained for DWA and the proposed PSDWA with different numer of examinations

Quality indicatorDWAPSDWA
Maximum number of examinations (itermax)
51530
Mean path length [m]6.237.036.196.18
Mean smoothness [rad]1.0631.2531.0591.056
Mean goal-reaching time [s]12.6014.5712.5812.56
Speed-up1.005.311.930.92

Comparison of quality indicators obtained for DWA and the proposed PSDWA examined in a thousand randomly generated environments with ten obstacles

Quality indicatorDWAPSDWA
Mean path length [m]3.9723.996
Mean smoothness [rad]0.9891.009
Mean goal-reaching time [s]8.1148.162
Mean computation time of the algorithm [ms]22.9711.60
Speed-up1.001.98

Parameters of the examined local path planning algorithms

ParameterSymbolValue
Sampling periodT_s0.01 s
Horizon of predictionHprediction1.0 s
Maximum linear accelerationamax1.0 m/s2
Maximum linear velocityvmax0.5 m/s
Maximum angular accelerationεmax1.0 · π rad/s2
Maximum angular velocityωmax0.5 · π rad/s
Obstacle distance reactiondobstaclemind_{obstacle}^{min}2.0
DWA: number of samples for linear velocityNv3
DWA: number of samples for angular velocityNω10
PSDWA: maximum number of examinationsitermax15
PSDWA: Initial step sizeΔInitial14[ vAGVmaxvAGVminωAGVmaxωAGVmin ]{1 \over 4}\left[ {\matrix{ {v_{AGV}^{max} - v_{AGV}^{min}} \cr {\omega _{AGV}^{max} - \omega _{AGV}^{min}} \cr } } \right]

Comparison of computation time of algorithms for various obstacle density in the environment

Number of obstaclesMean computation time [ms]Speedup
DWAPSDWA
24.97 ± 1.492.68 ±0.811.85
613.26 ±3.776.76 ± 1.931.97
1022.97 ±4.6711.60 ± 3.591.98
1542.92 ± 5.7821.40 ±4.492.01
3065.67 ± 7.1233.07 ±5.131.99
DOI: https://doi.org/10.2478/ama-2025-0074 | Journal eISSN: 2300-5319 | Journal ISSN: 1898-4088
Language: English
Page range: 659 - 664
Submitted on: Apr 27, 2025
Accepted on: Oct 5, 2025
Published on: Dec 19, 2025
Published by: Bialystok University of Technology
In partnership with: Paradigm Publishing Services
Publication frequency: 4 issues per year

© 2025 Rafal SZCZEPAŃSKI, published by Bialystok University of Technology
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.