基于少位姿传感器的挖掘机自动称重与位姿协同估计方法OA

Automatic weighing of excavator's material is essential for improving truck loading efficiency and enabling intelligent operation management.However,existing weighing methods often rely on multiple external pose sensors to obtain complete joint information of the working mechanism,which leads to high system costs,reduced reliability and difficulties in retrofitting.To overcome these limitations,we propose a data-driven"soft sensor"technique that uses only standard hydraulic cylinder pressure sensors and boom inclination sensors already standard on excavators to achieve real-time collaborative estimation of material weight and key poses during operation.By extracting dynamic features from specific operational phases,the method simplifies complex pose variations into slowly changing parameters.It employs a constrained nonlinear least squares fitting strategy to accurately estimate both material weight and stick position/orientation.These estimates are then used as known inputs to inversely derive the complete bucket pose.The results from simulations and field tests confirm that estimation errors for material weight and key pose parameters are within 3%.This approach provides an efficient and practical solution for automatic weighing systems and establishes a precise pose reference for enabling automated functions such as trajectory tracking,demonstrating strong potential for industrial applications.