Metode Improved Gaussian Sampling pada Algoritma Rapidly Exploring Random Tree*
In this paper a method called improved gaussian sampling design and implementation using the Rapidly Exploring Random Tree* (RRT*) algorithm. The design of algotihm using Laboratory Virtual Instrumentation Engineering Workbench (LabVIEW Improved gaussian sampling is the development of the gaussian sampling method by adding the number of sampling. The number of sampling used in this method using 10 sampels. To measure the performance of the proposed sampling method, we compare the performance of the improved gaussian sampling, gaussian sampling and random sampling methods. While the results of gaussian sampling test are obtained: clutter for 10 with a time of 40.09; narrow is 13.53 with a time of 40.03 and a trap is 10.95 with a time of 40.12. The results of random sampling test are: clutter with a length of 10.86 with a time of 0.03; narrow is 14.82 with a time of 0.25 and a trap is 11.71 with a time of 0.21. Based on the results of the improved Gaussian sampling test, average path cost and computation time were obtained: clutter 8.89 with a time of 40.05; narrow passages are 12.58 with a time of 40.03 and obstacle traps are 9.56 with a time of 40.04. The results of the tests carried out showed that the RRT* algorithm using the improved Gaussian sampling method was more optimal than using other methods. The results of the comparison of measurements based on the sampling value obtained an average path cost value of 10.12 with the number of sampling only 1 and the shortest path cost value of 8.9 with a sampling number of 10. Based on these measurements, the more the number of samplings given, the path value the resulting cost is more optimal.