In recent years, researchers have been actively pursuing research into developing robots that can be useful in many fields of industry （e.g., service, medical, and aging care）. Such robots must be safe and flexible ...In recent years, researchers have been actively pursuing research into developing robots that can be useful in many fields of industry （e.g., service, medical, and aging care）. Such robots must be safe and flexible so that they can coexist with people. Pneumatic actuators are useful for achieving this goal because they are lightweight units with natural compliance. Our research focuses on joint angle control for a pneumatically driven musculoskeletal model. In such a model, we use a one-degree-of-freedom joint model and a five-fingered robot hand as test beds. These models are driven by low pressure-driven pneumatic actuators, and mimic the mechanism of the human hand and musculoskeletal structure, which has an antagonistic muscle pair for each joint. We demonstrated a biologically inspired control method using the parameters antagonistic muscle ratio and antagonistic muscle activity. The concept of the method is based on coordination of an antagonistic muscle pair using these parameters. We have investigated the validity of the proposed method both theoretically and experimentally, developed a feedback control system, and conducted joint angle control by implementing the test beds.展开更多
It is well known that the transient behaviors of the traditional adaptive control may be very poor in general, and that the adaptive control designed based on switching between multiple models is an intuitively appeal...It is well known that the transient behaviors of the traditional adaptive control may be very poor in general, and that the adaptive control designed based on switching between multiple models is an intuitively appealing and practically feasible approach to improve the transient performances. In this paper, we shall prove that for a typical class of linear systems disturbed by random noises, the multiple model based least-squares(LS) adaptive switching control is stable and convergent, and has the same convergence rate as that established for the standard least-squares-based self-tunning regulators. Moreover, the mixed case combining adaptive models with fixed models is also considered.展开更多
文摘In recent years, researchers have been actively pursuing research into developing robots that can be useful in many fields of industry （e.g., service, medical, and aging care）. Such robots must be safe and flexible so that they can coexist with people. Pneumatic actuators are useful for achieving this goal because they are lightweight units with natural compliance. Our research focuses on joint angle control for a pneumatically driven musculoskeletal model. In such a model, we use a one-degree-of-freedom joint model and a five-fingered robot hand as test beds. These models are driven by low pressure-driven pneumatic actuators, and mimic the mechanism of the human hand and musculoskeletal structure, which has an antagonistic muscle pair for each joint. We demonstrated a biologically inspired control method using the parameters antagonistic muscle ratio and antagonistic muscle activity. The concept of the method is based on coordination of an antagonistic muscle pair using these parameters. We have investigated the validity of the proposed method both theoretically and experimentally, developed a feedback control system, and conducted joint angle control by implementing the test beds.
基金This research is supportedby the National Natural Science Foundation of China.
文摘It is well known that the transient behaviors of the traditional adaptive control may be very poor in general, and that the adaptive control designed based on switching between multiple models is an intuitively appealing and practically feasible approach to improve the transient performances. In this paper, we shall prove that for a typical class of linear systems disturbed by random noises, the multiple model based least-squares(LS) adaptive switching control is stable and convergent, and has the same convergence rate as that established for the standard least-squares-based self-tunning regulators. Moreover, the mixed case combining adaptive models with fixed models is also considered.
