The drive torque for each motor can be performed independently for the distributed drive electric wheel loader (DDEWL). On the shovelling condition, the torque distribution can be optimised according to the tire load. Then the parasitic power and wheel slippage can be reduced, and the tractive force and efficiency are improved. In this study, the singular value decomposition unscented Kalman filter is adopted to estimate the shovelling load. Based on the estimated shovelling load, the vertical tire force is calculated. The tire load rate is used to build the optimisation objective. The drive antiskid is set as the boundary condition. And a modified particle swarm optimisation is applied as the optimisation algorithm. A wheel loader is utilised for the testing data acquisition. Based on the acquired testing data, the comparative study about the optimisation and un-control strategies is operated. The results show that the slippage occurs less in the optimisation strategy, and the tractive force and efficiency are higher than the un-control strategy. In this study, the feasibility of the shovel load estimator is verified. The advantages of the optimisation strategy of distributed drive electric wheel loaders in shovelling process are analysed.

Bucket fill factor is of paramount importance in measuring the productivity of construction vehicles, which is the percentage of materials loaded in the bucket within one scooping. Additionally, the locational information of the bucket is also indispensable for scooping trajectory planning. Some research has been conducted to measure it via state‐of‐the‐art computer vision approaches, but their robustness against various environment conditions is not considered. The aim of this study is to fill this gap and six distinctive environment settings are included. Images captured by a stereo camera are used to generate point clouds before being structured into 3D maps. This novel preprocessing pipeline for deep learning is originally proposed and its feasibility has been validated through this study. Moreover, multitask learning is employed to exploit the positive relationship among two tasks: fill factor prediction and bucket detection. Therefore, after preprocessing, 3D maps are forwarded to a faster region with convolutional neural network incorporated with an improved residual neural network. The value of fill factor is acquired via a classification and probabilistic–based approach, which is novel, achieving an inspiring result (overall volume estimation accuracy: 95.23% and detection precision: 92.62%) at the same time.

Articulated wheel loaders that travel on unstructured roads experience severe vibration and poor stability. Introducing suspended axles on wheel loaders, which are traditionally constructed without wheel suspension, is desirable for ride comfort. This study mainly focuses on the parameter optimization of the hydropneumatic suspension to obtain the minimum root mean square of vertical accelerations under different driving conditions, thereby improving the ride comfort of the wheel loader. The multibody model of the wheel loader with hydropneumatic suspension was developed by RecurDyn in co-simulation with MATLAB/Simulink. The vertical acceleration root mean square at the seat position was analyzed when the wheel loader was traveling on class C, D, and E roads with different travel speeds. The surrogate model of the vertical acceleration root mean square with respect to the suspension parameters was established based on kriging method. The established surrogate model was then optimized using particle swarm optimization algorithm. The optimization results of the hydropneumatic suspension parameters of the wheel loader under different road excitations and driving speeds were obtained. Simulation and optimization results show that a well-designed hydropneumatic suspension system can significantly improve wheel loader performance in reducing the vertical acceleration at the seat position compared to a suspension system without optimization.