Think about how many tasks go into gripping a pen as you prepare to write your name. It’s a simple task, right? Well, maybe not. The mundane task involves calculating the geometry of the pen, calculating the friction and contact forces to ensure the pen can be held without slipping, the force distribution to avoid snapping the pen in half, the finger placement, feedback control, and other safety considerations. Suddenly, when you break down everything your brain does when you pick up a pen, it's not so simple.
This seemingly simple task underscores the complexity of human hand dexterity. For years, researchers have been working towards having robotic hands match the dexterity and flexibility of the human one. The ultimate goal is to have robotics hands be able to do the mundane tasks that many of us take for granted every day. From manufacturing to disaster response, the impact of such advanced robotic manipulation capabilities will be profound for society.
Tool Skill Acquisition by a Robotic Hand
One researcher working towards the goal of having robotic hands master the complexity of human hands is Iowa State Computer Science Ph.D. student Ling Tang. Tang works in the Robotics Lab alongside Yan-Bin Jia, a Professor of Computer Science at Iowa State University. Tang is photographed on the left with the robot hand.
“Ling is solidly trained in mathematics and well-balanced in programming and hardware skills,” said Jia. “She is a quick learner and has the courage and ability to pick up a missing background encountered in research, whether the knowledge is from mechanics, control, or optimization. She is highly motivated for problem investigation, which she performs with meticulous attention to detail.”
Picture 1: Tang with the robot.
Tang’s focus lies in enabling robotic hands to comprehend and adapt to the dynamics between tools, manipulated objects, and their environment. Her recent research is one having a robotic arm and hand pair be able to use a screwdriver. Her paper “Robotic Manipulation of Hand Tools: The Case of Screwdriving” was recently accepted to the IEEE International Conference on Robotics and Automation, a premier conference that provides a platform for researchers to present and discuss their latest advancements in the field of robotics and automation. The paper was written alongside Dr. Jia and Dr. Yuechuan Xue, a recently graduated Ph.D. student from the Department of Computer Science at Iowa State University.
Picture 2: Tang working with the robot.
In the paper, she focused on how to make a robot arm and hand screw things in by carefully controlling the forces involved and making sure everything moves smoothly together. This was done through a force control scheme that was created through backward chaining to leverage the dynamics of the screwdriver and arm/hand. They used feedback from what the robot hand was feeling to adjust how much force was being used, which allowed the robot hand to know how hard it should grip the screw and how much force it should apply. Using that information, they were able to calculate the torques that needed to be applied to keep everything tight.
Picture 3: Tang is at the computer, to get the robot ready.
Implications and Future Directions
Tang’s research is a step towards bridging the gap between human and robotic manipulation capabilities. By focusing on the intricacies of tool manipulation and developing methodologies, Tang is pushing the boundaries of what robotic hands can achieve. As her work demonstrates, robotic hands are continuing to inch closer to being able to emulate the finesse of human hands. Researchers such as Tang will play an important role in shaping the future landscape of robotics as they continue to make innovations and advancements in their field.