Learning how to grasp based on neural network retraining

Autor(es)

Pedro, Leonardo M. ^[1] ; Belini, Valdinei L. ^[1] ; Caurin, Glauco A. P. ^[2]

Afiliação do(s) autor(es)

^[1] Univ Fed Sao Carlos, Ctr Exact Sci & Technol, Sao Paulo - Brazil

^[2] Univ Sao Paulo, Engn Sch Sao Carlos, Sao Paulo - Brazil

• Tipo de documento: Artigo Científico
• Fonte: ADVANCED ROBOTICS; v. 27, n. 10, p. 785-797, JUL 1 2013.
• Citações Web of Science: 5

Resumo

Humans have an incredible capacity to manipulate objects using dextrous hands. A large number of studies indicate that robot learning by demonstration is a promising strategy to improve robotic manipulation and grasping performance. Concerning this subject we can ask: How does a robot learn how to grasp? This work presents a method that allows a robot to learn new grasps. The method is based on neural network retraining. With this approach we aim to enable a robot to learn new grasps through a supervisor. The proposed method can be applied for 2D and 3D cases. Extensive object databases were generated to evaluate the method performance in both 2D and 3D cases. A total of 8100 abstract shapes were generated for 2D cases and 11700 abstract shapes for 3D cases. Simulation results with a computational supervisor show that a robotic system can learn new grasps and improve its performance through the proposed HRH (Hopfield-RBF-Hopfield) grasp learning approach. (AU)

• Processo FAPESP: 00/11321-2 - Habilidades senso-motoras aplicadas ao desenvolvimento de mãos artificiais robotizadas
• Beneficiário: Glauco Augusto de Paula Caurin
• Modalidade de apoio: Auxílio à Pesquisa - Jovens Pesquisadores
• Processo FAPESP: 06/06597-5 - Kanguera: olho local - mão distante: WebLab sobre operação de mão mecatrônica baseada em visão
• Beneficiário: Adilson Gonzaga
• Modalidade de apoio: Auxílio à Pesquisa - Tecnologia da Informação no Desenvolvimento da Internet Avançada - TIDIA