Series elastic actuators (SEAs) stand as pivotal components in the landscape of human-robot interaction, offering indispensable safety features and precise force control capabilities by incorporating a compliant element between the actuator and load. Traditionally, force sensing in SEAs has relied on deformation-based force measurement (DFM) methods utilizing encoders on both masses. However, this approach presents drawbacks in terms of cost, mechanical design, and performance, inspiring the exploration of alternative sensing technologies like low-cost MEMS accelerometers. This chapter introduces a thorough methodology centered around leveraging load acceleration measurements for a variety of common SEA applications within human-robot interaction contexts. These applications include interaction force estimation, parameter identification, and impedance control. An innovative Kalman Filter (KF)-based algorithm is employed for accurate system state estimation, comprehending the interaction force, proving particularly effective in scenarios influenced by gravitational forces. Parameter identification, essential for effective model-based control design, is achieved through a novel disturbance observer-based method utilizing only actuator position and load acceleration measurements. Moreover, the chapter proposes a novel impedance control architecture that surmounts the well-known passivity constraint inherent in conventional methods, where the desired virtual spring value cannot exceed the physical one. This solution is demonstrated to be effective, since it can be applied in combination with the KF introduced previously. Theoretical discussions are provided along with rigorous experimental validations, solidifying the efficacy of these methods in enhancing the performance of SEAs across diverse application domains. By embracing MEMS accelerometers, SEAs uphold their safety features while potentially elevating overall system performance, thereby driving advancements in human-robot interaction technology.
Use of MEMS accelerometers in series elastic actuators for physical interaction
Budau Petrea R. A.
;Oboe R.
2026
Abstract
Series elastic actuators (SEAs) stand as pivotal components in the landscape of human-robot interaction, offering indispensable safety features and precise force control capabilities by incorporating a compliant element between the actuator and load. Traditionally, force sensing in SEAs has relied on deformation-based force measurement (DFM) methods utilizing encoders on both masses. However, this approach presents drawbacks in terms of cost, mechanical design, and performance, inspiring the exploration of alternative sensing technologies like low-cost MEMS accelerometers. This chapter introduces a thorough methodology centered around leveraging load acceleration measurements for a variety of common SEA applications within human-robot interaction contexts. These applications include interaction force estimation, parameter identification, and impedance control. An innovative Kalman Filter (KF)-based algorithm is employed for accurate system state estimation, comprehending the interaction force, proving particularly effective in scenarios influenced by gravitational forces. Parameter identification, essential for effective model-based control design, is achieved through a novel disturbance observer-based method utilizing only actuator position and load acceleration measurements. Moreover, the chapter proposes a novel impedance control architecture that surmounts the well-known passivity constraint inherent in conventional methods, where the desired virtual spring value cannot exceed the physical one. This solution is demonstrated to be effective, since it can be applied in combination with the KF introduced previously. Theoretical discussions are provided along with rigorous experimental validations, solidifying the efficacy of these methods in enhancing the performance of SEAs across diverse application domains. By embracing MEMS accelerometers, SEAs uphold their safety features while potentially elevating overall system performance, thereby driving advancements in human-robot interaction technology.
Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
