Triumph in Dice Manipulation: A Success Story from the 1st Doctoral Summer School on Robotics and Intelligent Machines

In the realm of robotics innovation, a remarkable success story unfolded at the recent 1st Doctoral Summer School on Robotics and Intelligent Machines. Doctoral students showcased their prowess in tackling a unique challenge that revolved around the manipulation of a dice to display a specific number. The event, held from August 30 to September 5 at the scenic Scuola Internazionale di Alta Formazione (SIAF) in Volterra, Pisa, Italy, witnessed a fusion of cutting-edge techniques and interdisciplinary methodologies employed by these brilliant minds.

The Challenge

The primary objective for the participating students was to manipulate a dice, turning it to reveal a chosen number on one of its faces. The challenge unfolded in a controlled environment where teams had access to advanced tools such as the YuMi manipulator from ABB (Switzerland), a fixed OAK-D Pro camera from Luxonis (USA), and the integration of force sensors onto gripper fingers.

The Multidisciplinary Approach

Undaunted by the complexity of the task, the students embarked on a multidisciplinary journey, pooling their expertise in finger design, strategic control, vision-based perception, and sensor integration. Their collective goal was to craft a robotic finger compatible with a robotic arm, addressing specific constraints such as finger distance, sensor accommodation, and ease of installation. They then utilized this engineered finger to adeptly manipulate a dice through a robotic arm equipped with a vision sensing setup.

The Strategic Solution

The heart of their success lay in a well-thought-out control strategy. Leveraging the inherent properties of a dice—where the sum of opposite faces always equates to seven—the students devised a strategy built on three core primitives: picking and placing, flipping the dice by 90°, and performing z-axis rotations. This strategy, became the backbone of their approach, ensuring precision and efficiency in achieving the goal.

Vision and Integration

The vision task unfolded in three key steps: calibration, detection, and identification. By integrating sensors onto gripper fingers, the students enhanced their manipulation capabilities. Real-time feedback and control were achieved through UDP communication with external systems, enabling advanced operations such as slip detection.

Triumphant Integration

In their pursuit of mastery over dice manipulation, the students successfully integrated finger design, vision, and sensor components. This cohesive approach not only met the challenge but propelled their team to victory. Consistently and accurately manipulating the dice to display the target number within the designated time frame, they minimized iterations and identifications, standing out as the triumphant team at the 1st Doctoral Summer School on Robotics and Intelligent Machines.