Abstract

The central question of my research is how we can create robots that are capable of adapting such that they can co-inhabit our world. This means designing systems that are capable of functioning in unstructured environments that are continuously changing with unlimited combinations of shapes, sizes, appearance, and positions of objects; adapting to changes and learning from experience and humans, and importantly do so from small amounts of data. In specific, my work focuses on grasping and manipulation, fundamental aspects to enable a robot to interact with the environment, humans and other agents, along with dexterity (e.g. to use objects/tools successfully) and high-level reasoning (e.g. to decide about which object/tool to use). Despite decades of research, robust autonomous grasping and manipulation at a level approaching human skills remains an elusive goal. One main difficulty lies in dealing with the inevitable uncertainties in how a robot perceives the world, based on sensory measurements that can be noisy and incomplete, which poses challenges to avoid failures. In my research I study how to enable a robot to interact with natural objects and learn about object properties and relations between tasks and sensory streams. I develop tools that allow a robot to use multiple streams of sensory data in a complementary fashion. In this talk I will address how a robot can use sensory data, e.g. visual and tactile, in various steps involved to achieve robot grasping and manipulation tasks.

Abstract

The classic approach to grasping and manipulation with rigid robotic hands generally favored object-centric analytical solutions, which - although very elegant and theoretically sound - has not yet produced the desired outcomes in the practice. This talk aims at discussing a different path, which entails moving the focus from the object to the robotic hand. This shift of perspective opens up several challenges which require a full integration of models, materials, machine learning, and bio-inspiration to be tackled effectively. Combining these ingredients, control intelligence can be embedded directly in the hand mechanics. As a result, soft end-effectors can achieve high-level grasping and manipulation performance when operated by humans. However, such a level of dexterity is still unmatched in autonomous grasp execution. Indeed, classic approaches cannot be applied to this kind of hands, which - by their own nature - do not allow fingertips placement with the required precision and relative independence. On the contrary, data driven approaches could be the key to learn from humans how to manage soft hands, towards higher levels of autonomous grasping capabilities.

Abstract

Automation in food agricultural industries is not as developed as in the automotive and manufacturing industries. Several challenges hinder the rapid introduction of robots into the food and agricultural industries, for example, the required processing speed, the cost benefit of using robot systems, lack of handling strategies and robotic end-effectors designed to cope with the large variety and variability of food and agricultural products, and lack of sufficient understanding of the product properties as an “engineering” material for handling tasks. In this talk, I will introduce my work on robotic systems, particularly on robotic grippers, for handling various food and agricultural products, and the modeling and measurements of physical properties of food materials.

Abstract

The human hand is the pinnacle of dexterity – it has the ability to powerfully grasp a wide range of object sizes and shapes as well as delicately manipulate objects held within the fingertips. Current robotic and prosthetic systems, however, have only a fraction of that manual dexterity. In this talk, I will discuss how this gap can be addressed in three main ways: examining the mechanics and design of effective hands, studying biological hand function as inspiration and performance benchmarking, and developing novel control approaches that accommodate task uncertainty. I will place a special focus on examining the mechanics of the open- and closed-loop hand/object/environment system during robot manipulation and how contact and kinematic constraints dominate performance limits. Using this understanding, I will show how prioritizing passive mechanics in robot hand design, including incorporating adaptive underactuated transmissions and carefully tuned compliance, can maximize open-loop performance while minimizing complexity.