Research on the Prediction of Insertion Resistance of Wheel Loader Based on PSO-LSTM

Insertion resistance is the caused by a pile to wheel loader when latter inserts into pile. It significant clarify insertion avoid slippage, increase additional energy consumption, and protect during process. To address problem that current methods cannot accurately obtain magnitude variation trend, we propose composite model based on particle swarm optimization (PSO) algorithm long short-term memory (LSTM) neural network. Firstly, Pearson correlation coefficient method used test parameters related resistance. Following this, hyperparameters in LSTM are optimized PSO. Finally, different proportions of training sets set PSO-LSTM compared with LSTM. The experimental data selected from gravel sample groups sand consisting depths 600 mm, 800 1000 mm. results show has higher prediction accuracy, better robustness, stability, generalization ability In PSO-LSTM, proportion 80%, average relative errors 2.28%, 1.57%, 1.53% for group 1.14%, 0.71%, 0.60% group.