Estimation of Tire Friction for Tip-Over Analysis Based on Genetic Algorithm and Long Short-Term Memory

초록

The forklifts used in various industries have a dynamic characteristic similar to common vehicles. For example, a forklift can tip over while driving in a harmful environment. The possibility of tip-over is increased when the load applied to the fork is larger. Therefore, the longitudinal motion of the forklift must be obtained mathematically to predict the tip-over of a forklift analytically and quickly. This study used a forklift model currently in development to build the longitudinal and analytical model of the forklift. The mathematical model was generated based on the vehicle dynamics and Pacejka's tire model. A limitation is that the friction coefficient defined by the Magic formula must be obtained from experiments, which is time-consuming. Therefore, the friction coefficient was estimated using a genetic algorithm and long short-term memory. The mathematical model with the methods is compared to the multibody dynamics simulation model. The mathematical model with the methods represented the simulation model well. Experiments were also conducted to validate the mathematical model. Finally, the mathematical model was modified to predict the tip-over motion of the forklift and the results were discussed.

키워드