Symbiotic Insights Prediction on a Multi-Cooperative Modular Mobile Robot using Machine Learning

This paper presents the use of machine learning in predicting the symbiotic insights such as coefficient and carrying capacity of a given modular mobile robot configuration doing a cooperative load pushing task. The independent parameters were load and the number of modules. The actual experiment results from a DFRobot Maqueen mobile robot pushing a range of loads were used as dataset in the study. Various regression models were run as an ML algorithm to assess symbiotic relationship of a given configuration whether harmful, beneficial or has no effect and to yield the total distance that can be traversed given a set of input parameters. The researchers used python in running the ML model on Google Colab Notebook. Results show that Gradient Boosting performed best in the prediction of distance carrying capacity with 95.87% and Extra Trees for symbiotic coefficient with 93.75% accuracy which took about only 5 seconds training time on a T4 GPU device. The researchers were able to develop a custom trained ML regressor model that can immediately return a symbiotic insight on a cooperative pushing modular mobile robots.