Koh Hosoda Pengarang

Development of an upper-limb neuroprosthesis to voluntarily control elbow and h a n d Yosuke Ogiria, Yusuke Yamanoia, Wataru Nishinoa, Ryu Katoa, Takehiko Takagib and Hiroshi Yokoic aGraduate School of Engineering, Yokohama National University, Yokohama, Japan; bDepartment of Orthopaedic Surgery, Surgical Science, Tokai University School of Medicine, Kanagawa, Japan; cGraduate School of Informatics and Engineering, The University of Electro-Communications, Tokyo, Japan ABSTR AC T This work reports on the research and development of a lightweight neuroprosthesis that can control impaired motion using voluntary biological signals. The total weight of the developed neuroprosthesis is 900 g, which is the weight of 40% of the defective limb. Further, it is lighter than commercially available models. For a transhumeral amputee who had targeted muscle reinnervation (TMR) surgery, we attempted pattern classification using an artificial neural network (ANN) of a surface electromyogram (s-EMG) extracted from an innervated muscle. The result shows that classification of only five motions was possible using an s-EMG extracted from four dry electrodes. However, seven motion classifications were possible using eight wet-gel electrodes. The transhumeral amputee who had TMR surgery could thus successfully perform pick-and-place tasks using the neuroprosthesis. KEY WORDS Electromyography; above-elbow prosthesis; An improved approach to the inverse dynamic analysis of parallel manipulators by a given virtual screw Shuai Fan and Shouwen Fan S chool of Mechatronics Engineering, Universit y of Elec tronic S cience and Technology of China, Chengdu, Sichuan, People’s Republic of China ABSTR AC T This paper provides an improved approach to the inverse dynamic analysis of parallel manipulators (PMs) based on the screw theory and Jourdain’s principle of virtual power. First, velocity and acceleration mappings from the Cartesian coordinate system to the screw system are established. Next, by introducing a novel concept of virtual screw that is formulated by a combination of virtual angular velocity and virtual linear velocity, four theorems are defined and proven to build the dynamic equations of PMs. Owing to the existing expression of acceleration screws and the introduced virtual screw, the proposed approach not only has the advantages of intuitive physical concepts and universal form but also avoids the difficult derivatives of time and the determination of generalized velocities, which is employed by conventional methods and is determined difficultly for some hybrid PMs. Finally, taking a 1PU + 3UPS PM as an instance, the inverse dynamic analysis and numerical examples are presented to demonstrate the feasibility of the proposed approach. KEY WORDS I nverse dynamics; parallel manipulato r ; screw theor y ; vir tual screw ; vir tual power principle targeted muscle reinnervation; pattern recognition