Design And Implementation Of An Open Control Platform For A Humanoid Robot

Development of humanoid robots with human’s appearance characteristics, flexibilityand intelligence which have the ability of communicating with people, adapt to theenvironment, has been a dream of man. Humanoid Robot imitates the physiological featureand part of function of human brain, thus in the view of the complexity of the humanoidrobotics system, many key issues such as architecture design, intelligent computing, compute&store resources, need to be addressed.In this paper, referring to the basis of the reconfigurable modular humanoid robotarchitecture and its hierarchical software model, design and implement of an Open HumanoidRobot Control Platform is proposed. Also, referring to existing LabGrid and HNetCP, aHybrid Intelligent Computing Framework is designed for imitating function of human brain.Firstly, this paper analyzes the necessity of modularization design for humanoid robots,then develops a plugin framework to solve the problem of module integration andmanagement which is brought by modularization design. Secondly, Open HRCP, anexpandable and reconfigurable control system of humanoid robot based on plugin frameworkis designed. Then, in order to make up the deficiency of single intelligent method, a HybridIntelligent Computing Framework is designed under the guidance of HNetCP to slove thehuman’s intelligent problems with imprecision, uncertainty and vagueness. Finally, a VirtualHumanoid Robot–VHR, which can be used to verify the correctness of control algorithms, isdeveloped as an experiment by combining Open HRCP to the3D Real-time SimulationSystem.The experiment results of VHR prove that Open HRCP can realize characteristics suchas transparent deployment of modules, expandability, and reconfigurability. As an openplatform, Open HRCP also provides API and SDK to the developers in order to developmodules with unified standard interface. HICF is also designed on the basis of pluginframework, which makes it able to expand more hybrid intelligent computing methods forhumanoid robots in the future.