Hardware-in-the-Loop Optimization of an Interaction Controller for Improved Coupled Dynamics

Autor(es)

Lahr, Gustavo J. G. ; Garcia, Henrique B. ; Silva, Thiago H. S. ; Caurin, Glauco A. P. ; Fleury, AD ; Rade, DA ; Kurka, PRG

• Tipo de documento: Artigo Científico
• Fonte: PROCEEDINGS OF DINAME 2017; v. N/A, p. 13-pg., 2019-01-01.

Resumo

This paper presents the implementation of an optimization method to find, without knowledge of the environment characteristics, the best interaction controller parameters to revamp the coupled dynamics. The objective is to improve various industrial robot applications that involves mechanical contact. An enhanced contact is accomplished by lowering the following metrics: rise time, total variation and steady state error. Hence, the impedance controller was the interaction control technique chosen to be optimized. Contact is established between a Kuka KR16 robot TCP and an aluminum platform, where the force data was acquired by a 6-axis force-torque sensor located in the robot's end-effector. Using a hardware-in-the-loop optimization approach, the force feedback is processed by a NSGA-II algorithm. Each individual of the GA represents a specific impedance controller and as the generations passes, these values get more suitable for lowering the metrics. Results show convergence in 5 generations. (AU)

• Processo FAPESP: 15/07484-9 - Síntese de movimentos em tempo real para tarefas dinâmicas de montagem de precisão utilizando robôs
• Beneficiário: Gustavo Jose Giardini Lahr
• Modalidade de apoio: Bolsas no Brasil - Mestrado
• Processo FAPESP: 15/24343-0 - Estudo comparativo entre métodos de aprendizagem por reforço com aplicação na produção de mobiliário aeronáutico
• Beneficiário: João Victor Rocha Soares
• Modalidade de apoio: Bolsas no Brasil - Iniciação Científica
• Processo FAPESP: 13/07276-1 - CEPOF - Centro de Pesquisa em Óptica e Fotônica
• Beneficiário: Vanderlei Salvador Bagnato
• Modalidade de apoio: Auxílio à Pesquisa - Centros de Pesquisa, Inovação e Difusão - CEPIDs