MIT Has Enabled A Soft Robotic Arm To Understand Its Configuration
MIT has announced that for the first time, its team of scientists has enabled a soft robotic arm to understand its configuration in a 3D space. It has done this by leveraging motion and position data from the robotic arm's own "sensorized skin." Soft robots are made from compliant materials that are like those found in living organisms.Soft robots are an area of intense study because they are safer, more adaptable, resilient, and bioinspired alternatives to traditional robots made of rigid materials. A big challenge in making those soft robots is that the deformable design of soft robots allow them to move in virtually any direction at any moment. Scientists say that flexibility makes it hard to train planning and control modules to drive automation.
The team has developed a system of soft sensors that cover the robot's body to provide "proprioception," meaning awareness of motion and position of its body. The data gathered by the sensors runs through a deep-learning model that sifts through the noise and captures clear signals to estimate the robot's 3D configuration. A validation test was conducted on a robotic arm that looked like the trunk of an elephant and could predict its own position and autonomously swing around and extend.
The sensors the team used are made using off-the-shelf materials. One of the goals in the future for the researchers is to help make artificial limbs that can more dexterously handle and manipulate objects in the environment. A long-time goal for soft robotics was to integrate body sensors into the robots fully. One breakthrough for the team came when a researcher found sheets of conductive materials used for electromagnetic interference shielding that can be purchased anywhere in rolls.
The material has piezoresistive properties, and the scientists realized they could make effective soft sensors with the material. As the sensor is deformed, its electrical resistance is converted to a specific output. The voltage is used as a signal to correlate to the movement. Eventually, the team turned to a variation of origami called kirigami to laser-cut rectangular strips of conductive silicone sheets into various patterns to make them more flexible. Those were then added to the body of previously designed soft robots.