• adaptive synergies
• hand
• robotic
• robotic hand
• stiffness
• synergies
• variable stiffness actuator
• VSA

This thesis deals with the design possibilities concerning the next generation of advanced Robots. Aim of the work is to study, analyse and realise artificial systems that are essentially simple, performing and robust and can live and coexist with humans. The main design guideline followed in doing so is the Soft Robotics Approach, that implies the design of systems with intrinsic mechanical compliance in their architecture. The first part of the thesis addresses design of new soft robotics actuators, or robotic muscles. At the beginning are provided information about what a robotic muscle is and what is needed to realise it. A possible classification of these systems is analysed and some criteria useful for their comparison are explained. After, a set of functional specifications and parameters is identified and defined, to characterise a specific subset of this kind of actuators, called Variable Stiffness Actuators. The selected parameters converge in a data-sheet that easily defines performance and abilities of the robotic system. A complete strategy for the design and realisation of this kind of system is provided, which takes into account their me- chanical morphology and architecture. As consequence of this, some new actuators are developed, validated and employed in the execution of complex experimental tasks. In particular the actuator VSA-Cube and its add-on, a Variable Damper, are developed as the main com- ponents of a robotics low-cost platform, called VSA-CubeBot, that
v
can be used as an exploratory platform for multi degrees of freedom experiments. Experimental validations and mathematical models of the system employed in multi degrees of freedom tasks (bimanual as- sembly and drawing on an uneven surface), are reported.
The second part of the thesis is about the design of multi fingered hands for robots. In this part of the work the Pisa-IIT SoftHand is introduced. It is a novel robot hand prototype designed with the purpose of being as easily usable, robust and simple as an industrial gripper, while exhibiting a level of grasping versatility and an aspect comparable to that of the human hand. In the thesis the main theo- retical tool used to enable such simplification, i.e. the neuroscience– based notion of soft synergies, are briefly reviewed. The approach proposed rests on ideas coming from underactuated hand design. A synthesis method to realize a desired set of soft synergies through the principled design of adaptive underactuated mechanisms, which is called the method of adaptive synergies, is discussed. This ap- proach leads to the design of hands accommodating in principle an arbitrary number of soft synergies, as demonstrated in grasping and manipulation simulations and experiments with a prototype. As a particular instance of application of the method of adaptive syner- gies, the Pisa–IIT SoftHand is then described in detail. The design and implementation of the prototype hand are shown and its effec- tiveness demonstrated through grasping experiments. Finally, control of the Pisa/IIT Hand is considered. Few different control strategies are adopted, including an experimental setup with the use of surface Electromyographic signals.