Traditional automated guided vehicle(AGV)primarily relies on scheduling systems to manage warehouse locations and execute picking or placing tasks on fixedheight pallets.However,these conventional systems are illsuite...Traditional automated guided vehicle(AGV)primarily relies on scheduling systems to manage warehouse locations and execute picking or placing tasks on fixedheight pallets.However,these conventional systems are illsuited for scenarios involving variable heights,such as vehicle loading and unloading or the complex stacking of soft packages.To address the challenges of AGV endeffector operations in nonfixed height scenarios,this paper proposes an innovative solution leveraging lowcost depth camera sensors.By capturing image and depth data,and integrating deep learning,image processing,and spatial attitude calculation techniques,the method accurately determines the position of the endeffector center point relative to the upper plane of the fork.The approach effectively resolves a key issue in AGV operations within intelligent logistics scenarios that lack fixed heights.The proposed algorithm is deployed on a domestic embedded,lowcost ARM chip controller,and extensive experiments are conducted on a real AGV equipped with multiple stacked vehicles and nonstandard vehicles.The experimental results demonstrate that for diverse vehicles with different heights,the measurement error can be maintained within±10 mm,satisfying the requirements for highprecision measurement.The height measurement method developed in the paper not only enhances the AGV’s adaptability in nonfixed height scenarios but also significantly broadens its application potential across various industries.展开更多
基金Supported by the Key Research and Development Program of Anhui Province(No.201904a05020035)the Postdoctoral Research Initiative of Anhui Province(No.2024B804)the Hefei City Key Technology Research and Development‘Ranking’(No.2023SGJ017).
文摘Traditional automated guided vehicle(AGV)primarily relies on scheduling systems to manage warehouse locations and execute picking or placing tasks on fixedheight pallets.However,these conventional systems are illsuited for scenarios involving variable heights,such as vehicle loading and unloading or the complex stacking of soft packages.To address the challenges of AGV endeffector operations in nonfixed height scenarios,this paper proposes an innovative solution leveraging lowcost depth camera sensors.By capturing image and depth data,and integrating deep learning,image processing,and spatial attitude calculation techniques,the method accurately determines the position of the endeffector center point relative to the upper plane of the fork.The approach effectively resolves a key issue in AGV operations within intelligent logistics scenarios that lack fixed heights.The proposed algorithm is deployed on a domestic embedded,lowcost ARM chip controller,and extensive experiments are conducted on a real AGV equipped with multiple stacked vehicles and nonstandard vehicles.The experimental results demonstrate that for diverse vehicles with different heights,the measurement error can be maintained within±10 mm,satisfying the requirements for highprecision measurement.The height measurement method developed in the paper not only enhances the AGV’s adaptability in nonfixed height scenarios but also significantly broadens its application potential across various industries.
