Robotic hands to increase the flexibility of robots

Grasping is an essential function for understanding how humans and machines interact with the environment in the context of task performance. Addressing the complexity of this interaction requires the development and control of increasingly advanced grippers, ranging from specialized industrial grippers offering a low level of flexibility, to prosthetic hands offering an intermediate level of flexibility, all the way to dexterous multi-fingered robotic hands capable of adaptive grasping and internal manipulation as well.

When the handling operations to be automated are simple and repetitive—as is often the case in manufacturing—simple symmetrical grippers are used, ranging up to specific or specialized grippers. This includes grippers controlled in on/off mode, as well as adhesion-based gripping systems utilizing various effects: electromagnetic for ferrous materials, electrostatic for very small objects, suction, etc. Gripper technologies, their performance, and their areas of application are presented in this industrial context.

As the need for flexibility and dexterity grows, the human hand remains the gold standard and serves as the universal gripper due to its ability to adapt to a wide variety of grasping tasks. This has led to the development of multi-fingered hands with varying degrees of actuation and technological complexity, ranging from adaptive grasping capabilities to in-hand manipulation. The development of prosthetic hands is a perfect example of the challenges involved in developing adaptive grippers and clearly illustrates the challenges of replicating human grasping taxonomies.

To address these challenges in replicating the capabilities of the human hand, we present the technologies and control strategies implemented to mimic this dexterity using multi-fingered hands. The design of the hands is discussed with a focus on the degree of actuation. The state of the art in robotic hand development clearly illustrates this key challenge: how to simplify the design while maintaining a high level of performance in terms of adaptive grasping and fine manipulation?

The control strategy for robotic hands must therefore serve a dual purpose: ensuring grip stability and executing a defined trajectory for the object within the hand. This strategy relies on modeling the multi-fingered hand, grip quality, and grip synthesis, all of which must be implementable within the hand controller in a real-time context. For this implementation, the robotic hand control architecture must meet strict specifications regarding motion control, which are addressed and illustrated through examples of robust adaptive grasping and fine manipulation.

To account for environmental variability and the robotic...