For decades, visions of the future have featured robots as household assistants. However, despite years of development in robotics technology, the dream of home robots remains unrealized, except for some basic tasks such as floor vacuuming. Toyota Research Institute stated that it has developed a soft silicone gripper toward this goal.

Toyota Research Institute (TRI) said that robotic assistance can not only help clean houses and offices but also assist the elderly, people with disabilities, or those facing age-related challenges, enabling them to live more independently, with dignity and happiness.

As part of its efforts in home robotics, TRI is committed to improving robot manipulation. Such grippers must be able to maintain a stable grasp during unintentional contact, enable precise placement, and ensure safe interaction. They must also be low-cost to become part of affordable and commercially viable home robots. TRI stated that its engineers designed a manipulator with all these features, called the "soft silicone gripper."

The soft silicone gripper is based on past developments.

The soft bubble gripper is based on the past work of the TRI manipulation team. Previously, TRI robots used traditional two-finger grippers (guided by external cameras) to grasp objects, sort them, and even place items correctly in dishwashers. The grippers relied on these cameras to perform their tasks. In other words, they had no tactile sense.

Now, TRI states that it has found a way to improve robot perception and manipulation using a soft gripper that can both passively hold objects better and actively sense how much force is applied. The soft gripper can also precisely measure lateral forces, which indicate when an object is slipping out of the gripper's grasp.

The gripper development team was led by Alex Alspach, who leveraged his background in soft robotics development to conceive the new gripper design, and Naveen Kuppuswamy, who developed the algorithm for its use. Together with the haptics team at TRI, they developed a technology that utilizes the robustness and compliance of inflated elastic bubbles to facilitate sensing via an internal camera. The camera provides a new perspective from inside the grasp, showing what is happening, including forces that are usually invisible.

Alspach and Kuppuswamy iterated on their design, building upon work at other research organizations. Early designs used a robotic arm with a single, large, round, soft bubble on the end-effector, which then evolved into two arms (also known as "big fingers") to perform dexterous, haptic-driven tasks, such as blind object sorting and threading a nut onto a large bolt.

TRI stated that its researchers now use a single arm with two smaller fingers, each utilizing a soft bubble. This soft silicone combines all the advantages of compliant grasping with real-time, real-world haptics. The soft silicone senses shape and force, identifying the object being grasped and the forces between the object and the fingers.

Furthermore, the team designed the soft silicone gripper using inexpensive materials, thereby removing potential barriers to the adoption of home robots.

Future goals for Toyota Research Institute TRI exploration

The long-term goal of the TRI robotics technology research team is to break through the traditional robot controller paradigm and adapt to changing environments.

Improving reflexes. The team stated that it aims to replicate real-time corrective behavior in robots. It is working toward distributed robot sensing, where information is processed and reacted to "locally," without needing to refer to a centralized "brain" for every decision. This enables robots to execute actions faster with less lookahead.

"This is more like human or animal behavior," Alspach explained. "In less than a second, we react and catch something falling on the table or slipping out of our hands, with almost no thought."

Expanding softness. Currently, only the manipulator is soft. But what if more surfaces could be made from this material?

This could bring significant advantages to human-robot interaction, including working in homes with children, pets, or individuals who may have physical impairments.

Alspach said, "Our goal is to build robots that can interact effortlessly with the world around them." "Softness makes robots more suitable for home applications. Going further, we can begin to combine lightweight materials, air, smaller electronics, and optimized designs to help us reduce weight. Being both soft and lightweight, with the ability to sense and react, contributes to safety."

"This is actually a very tricky problem," Kuppuswamy added. "In a sense, it is very difficult because you are using cutting-edge materials.... This is not just about traditional methods and materials like machining or welding."

Complicating this fact further is that any inflatable device introduces factors of continuous dynamic deformation, which disrupts usual equations. "Mathematically, soft is hard," said Kuppuswamy.

Publication and recognition for more practical robots

For more information on the research regarding the TRI soft silicone gripper r, see the published paper "Soft Silicone Grippers for Robust and Perceptive Manipulation." The team's research also received recognition at the ICRA 2020 annual IEEE International Conference on Robotics and Automation, and another publication on the soft gripper project received the Best Paper Award in Robotics and Automation in 2019.

TRI welcomes scientists working on robot manipulation to build upon their own insights and findings. To inspire such collaboration and further research on soft bubble grippers, they hosted the VisuoTactile 2020, a Visuo-Tactile Technology Thought Leaders Workshop held at the Robotics: Science and Systems (RSS) conference.

The company stated that the foundation of its soft silicone gripper is its manipulation research, aimed at making human-assist robots more reliable and robust. Researchers said that even without sensing capabilities, elastic materials or low stiffness provide superior gripping compared to standard soft grippers. It can adapt to various shapes and achieve a stable grip. The team said that this gripper combines strength and dexterity, enabling it to handle objects requiring delicate touch and steady handling.

By sensing geometry, camera images, and shear forces, robots can now perceive the objects they are holding by estimating object pose and sensing forces on the surface. TRI stated that because the soft silicone gripper uses relatively inexpensive cameras internally, it brings applications such as domestic robots closer to practical reality.