Periodic Reporting for period 2 - FlowBot (Smart fluidic circuits for autonomous soft robots)
Período documentado: 2022-07-01 hasta 2023-12-31
Soft robots have the potential to be more robust, adaptable, and safer for human interaction than traditional rigid robots. State-of-the-art developments push these soft robotic systems towards applications such as rehabilitation and diagnostic devices, exoskeletons for gait assistance, and grippers that can handle delicate objects. However, despite these exciting developments, there are two major challenges: (1) existing electronic control, intelligence and power of soft robots is too bulky to be embedded, and (2) the power efficiency of soft robots is extremely low. As a result, most existing soft robots still rely on tethers to deliver the required power and control.
To increase the application perspective of soft robots, in this project we aim to ‘cut the tethers’, with the ultimate goal to realise fully autonomous and smart soft robots that can interact with humans and complex environments. To do so, we are working towards replacing electronics with embedded fluidic circuits in a three step approach. First, we aim to achieve intelligent behaviour in soft robots by embedding smart fluidic circuits, that have integrated fluidic sensors to allow for feedback with the environment. Second, we are working on closing the fluidic circuit to be able to reuse elastic energy that is currently lost after each actuation cycle, to dramatically increase the efficiency of soft robots. Finally, we are embedding a reliable pressure source based on chemical fuel regulated by soft fluidic circuits.
Taken together, these new concepts will enable us to effectively power and control the behaviour of soft robots, and we are developing new experimental standards, computational approaches, and generalise design principles to achieve this. As a results, this research will pave the way for the first fully autonomous soft robots that are capable of operating for longer periods of time.
To increase the application perspective of soft robots, in this project we aim to ‘cut the tethers’, with the ultimate goal to realise fully autonomous and smart soft robots that can interact with humans and complex environments. To do so, we are working towards replacing electronics with embedded fluidic circuits in a three step approach. First, we aim to achieve intelligent behaviour in soft robots by embedding smart fluidic circuits, that have integrated fluidic sensors to allow for feedback with the environment. Second, we are working on closing the fluidic circuit to be able to reuse elastic energy that is currently lost after each actuation cycle, to dramatically increase the efficiency of soft robots. Finally, we are embedding a reliable pressure source based on chemical fuel regulated by soft fluidic circuits.
Taken together, these new concepts will enable us to effectively power and control the behaviour of soft robots, and we are developing new experimental standards, computational approaches, and generalise design principles to achieve this. As a results, this research will pave the way for the first fully autonomous soft robots that are capable of operating for longer periods of time.
The project aims to develop new concepts to effectively power and control the behaviour of soft robots. To achieve this, we have developed new experimental standards, computational approaches, and generalise design principles.
We have implemented a strategy on the sequential activation of soft actuators, based on fluidic circuits that consists of multiple hysteretic valves. These hysteretic valves have been developed in previous work, and in this project we have coupled used a combination of elastic capacitances and fluidic resistances to enable sequencing. In order to influence the sequences, we have developed kinking tubes that can be in a normally open or closed state, and therefore act as sensors. These kinking tubes are combined with tuneable capacitances to enable both short-term and long-term memory. At the moment, these sensorised fluidic circuits are implemented in a soft robot to demonstrate the underlying principles, which is the first step towards achieving a manual soft interface. Here, we have been able to achieve soft robots that can be manually steered through this programmable kinking tube interface, which allows us to place the robot in four states (slow straight movement, fast straight movement, steering left and steering right).
During this project on sequential actuation using the hysteretic valves, we found that most valves exhibit two stable states, an oscillating state and a regulating state. With the aim to understand the coexistence of two different deformation modes that occur at the same input conditions, and with that be better capable of controlling the behavior of combined system, we studied the mechanics and dynamics of this elastomeric dome-shaped fluid valve. Our results show that we need to account for mechanical and fluidic hysteresis to sufficiently describe the system. Furthermore, we show the applicability of our model in experiments by designing a modified valve with different fluidic behavior, to suppress the coexistence of one of the modes, making the valve more suitable for soft robotic applications.
Furthermore, we are also studying the possibility to use soft actuators directly as a sensor, by characterizing the changes in Pressure-Volume relation of an actuator to an external stimulus such as a mechanical load. In this project, we demonstrate that the compliance of soft fluidic actuators can be directly used to sense the environment, such that no design modifications are needed to implement sensing in soft (fluidic) robots and devices. We demonstrate the simplicity, versatility, applicability and robustness of our sensing strategy by implementing and retrofitting it in several soft actuators and soft grippers, enabling sensing of size, shape, surface roughness and stiffness.
We have demonstrated the possibility to achieve cyclic activation of actuators placed in closed fluidic circuits that contain hysteretic valves. We have started the implementation of numerical models based on the fluidic-electronic analogy, to determine how the fluidic circuits and specific design variables affect the efficiency. Experimentally, we have constructed a testing platform that allows us to measure the efficiency of soft extension and contraction actuators, which will be used as input for the model to better understand how to design systems and actuators with optimal efficiency.
Moreover, we have developed a 3D printing platform for silicones that enables us to print soft reaction chambers for leverage a chemical reaction between citric acid and sodium bicarbonate to generate gas pressure, and couple this chemical reaction to a fluidic setup that contains a hysteretic valve to control the maximum and minimum pressure observed during cyclic loading of soft actuators. We have implemented numerical models that allows us the predict the response of these fluidic systems powered by this chemical reaction. The same printing platform is now used to print soft robotic components such as actuators, connectors, sensors and fluidic circuits, that can be assembled to build a modular 3D printed fully soft robot that will operate to some extend autonomously.
We have implemented a strategy on the sequential activation of soft actuators, based on fluidic circuits that consists of multiple hysteretic valves. These hysteretic valves have been developed in previous work, and in this project we have coupled used a combination of elastic capacitances and fluidic resistances to enable sequencing. In order to influence the sequences, we have developed kinking tubes that can be in a normally open or closed state, and therefore act as sensors. These kinking tubes are combined with tuneable capacitances to enable both short-term and long-term memory. At the moment, these sensorised fluidic circuits are implemented in a soft robot to demonstrate the underlying principles, which is the first step towards achieving a manual soft interface. Here, we have been able to achieve soft robots that can be manually steered through this programmable kinking tube interface, which allows us to place the robot in four states (slow straight movement, fast straight movement, steering left and steering right).
During this project on sequential actuation using the hysteretic valves, we found that most valves exhibit two stable states, an oscillating state and a regulating state. With the aim to understand the coexistence of two different deformation modes that occur at the same input conditions, and with that be better capable of controlling the behavior of combined system, we studied the mechanics and dynamics of this elastomeric dome-shaped fluid valve. Our results show that we need to account for mechanical and fluidic hysteresis to sufficiently describe the system. Furthermore, we show the applicability of our model in experiments by designing a modified valve with different fluidic behavior, to suppress the coexistence of one of the modes, making the valve more suitable for soft robotic applications.
Furthermore, we are also studying the possibility to use soft actuators directly as a sensor, by characterizing the changes in Pressure-Volume relation of an actuator to an external stimulus such as a mechanical load. In this project, we demonstrate that the compliance of soft fluidic actuators can be directly used to sense the environment, such that no design modifications are needed to implement sensing in soft (fluidic) robots and devices. We demonstrate the simplicity, versatility, applicability and robustness of our sensing strategy by implementing and retrofitting it in several soft actuators and soft grippers, enabling sensing of size, shape, surface roughness and stiffness.
We have demonstrated the possibility to achieve cyclic activation of actuators placed in closed fluidic circuits that contain hysteretic valves. We have started the implementation of numerical models based on the fluidic-electronic analogy, to determine how the fluidic circuits and specific design variables affect the efficiency. Experimentally, we have constructed a testing platform that allows us to measure the efficiency of soft extension and contraction actuators, which will be used as input for the model to better understand how to design systems and actuators with optimal efficiency.
Moreover, we have developed a 3D printing platform for silicones that enables us to print soft reaction chambers for leverage a chemical reaction between citric acid and sodium bicarbonate to generate gas pressure, and couple this chemical reaction to a fluidic setup that contains a hysteretic valve to control the maximum and minimum pressure observed during cyclic loading of soft actuators. We have implemented numerical models that allows us the predict the response of these fluidic systems powered by this chemical reaction. The same printing platform is now used to print soft robotic components such as actuators, connectors, sensors and fluidic circuits, that can be assembled to build a modular 3D printed fully soft robot that will operate to some extend autonomously.
These achievements are all original results, and present important steps towards the further development of soft autonomous robots that harness the power of fluidic circuits. Besides the scientific impact, I have experienced the potential of valorizing our research, and creating also impact in society. It is difficult to predict beforehand industrial and societal interest in the research we have performed. While we initially did not plan for significant interaction with such partners, we found that there is considerable interest from industry and society that we are currently trying to leverage to create broad impact. The development of a soft fluidic toolkit for educational purposes for example has gained considerable traction, and we are therefore continuing this development. Moreover, interest from industry in the soft sensing approaches that we developed also warrants more effort into exploring this direction.