Bayesian optimization for robust robotic grasping using a sensorized compliant hand
This work addresses the challenge of flexible and efficient robotic grasping for applications like industrial automation and assistive technologies, representing an incremental improvement over trial-and-error methods.
The paper tackled the problem of enabling robots to grasp unknown objects robustly in real-world environments by using Bayesian optimization to adapt grasping strategies, achieving successful grasps despite noise and uncertainty.
One of the first tasks we learn as children is to grasp objects based on our tactile perception. Incorporating such skill in robots will enable multiple applications, such as increasing flexibility in industrial processes or providing assistance to people with physical disabilities. However, the difficulty lies in adapting the grasping strategies to a large variety of tasks and objects, which can often be unknown. The brute-force solution is to learn new grasps by trial and error, which is inefficient and ineffective. In contrast, Bayesian optimization applies active learning by adding information to the approximation of an optimal grasp. This paper proposes the use of Bayesian optimization techniques to safely perform robotic grasping. We analyze different grasp metrics to provide realistic grasp optimization in a real system including tactile sensors. An experimental evaluation in the robotic system shows the usefulness of the method for performing unknown object grasping even in the presence of noise and uncertainty inherent to a real-world environment.