雙足和人型機器人在複雜的機器人系統中代表著最先進的技術。相較於輪型機器人,雙足式服務型機器人在人類生存的環境中有著高度的靈活性,能夠替人類提供服務。然而,雙足機器人研究領域的一個重大瓶頸在於考量靜態與動態的平衡。平衡的課題是必要的考量,才能夠順利地完成更進一步的應用。 即便在一個特別設計好的環境中,雙足機器人不可避免地會遭遇到外在環境的擾動。期望雙足機器人在不久的將來,能夠在日常生活的環境中替人類提供服務,因此,受到外力干擾後,如何回復平衡是一個必要探討的課題,這樣才能夠使雙足機器人在日常生活環境中,與人類的互動過程更加安全。 本研究旨在實現外力干擾回復系統,並且將此系統整合至雙足機器人行走的步態軌跡產生器模組中,使得雙足機器人能夠實現穩定的動態行走模式。此外,這個系統是基於雙足機器人的研究,來自於我們對人類的了解此一概念來建構而成。 在本研究中,我們成功地將擬人式的外力干擾回復系統實現在雙足機器人行走的步態軌跡產生器中,使得雙足機器人在行走過程中,面對意外發生地外力仍然能夠維持平衡。本研究包括雙足機器人理論的推導,並透過在國立臺灣大學智慧機器人及自動化國際研究中心實驗室研究發展中的雙足機器人,實際行走並受外力干擾後回復平衡,來驗證此外力干擾回復系統的實務性及可行性。期望本研究能夠在雙足機器人行走的外力干擾回復研究上提供學術及實務應用之貢獻。
The biped and humanoid robots represent the state of the art technology in complex robot systems. Biped walking robots have high flexibilities to serve as service robots in comparison with the wheeled structure robots in human environments. However, one of the most important bottlenecks in the research field of the biped robots is the static balancing and dynamic balancing issues. Further applications cannot be successfully fulfilled without considering the balancing issues. In the hopes that the biped robots assist humans in everyday environments in the near future, it is inevitable for the biped robots to confront unexpected perturbations even though the environments are well structured. As a result, the push-recovery issue is a necessary research that must be investigated to interact safely with humans as well. The research objective of this thesis is to implement the push-recovery system to the walking pattern generation module for the biped walking robots to fulfill dynamic balancing walking. Furthermore, the push-recovery system is constructed based on the idea that the researches on biped and humanoid robots are mainly to implement the behaviors as far as we know about the human beings. In this research, we successfully implement the push-recovery system that mimics human motions to the walking pattern generator for balancing unexpected perturbations during walking process. The research consists of the theoretical derivations of biped robot system. Furthermore, practical possibilities and feasibilities of the push-recovery system are verified by push-recovery experiments on the biped robot developed in our NTU-iCeiRA laboratory. Hope that this research would contribute to the academic and practical applications in the field of biped robot push-recovery researches.