Although roboticists have developed increasingly sophisticated systems over the past decades, most current robots still lack the dexterity and sensory capabilities of human hands. As a result, their effectiveness in real-world tasks—such as household chores, disaster debris removal, and assembly or maintenance work is often limited, especially in high-temperature environments like steel mills and foundries, where extreme heat can impair performance and reduce the precision needed for safe and accurate operation.
Researchers at the University of Southern California have developed a new robotic hand called MOTIF (Multimodal Observation with Thermal, Inertial, and Force sensors), designed to enhance the object manipulation skills of humanoid robots. Detailed in a paper published on the arXiv preprint server, the MOTIF hand integrates multiple types of sensors, including tactile sensors, a depth sensor, a thermal camera, inertial measurement unit (IMU) sensors, and a visual sensor.
“Our work stemmed from the need to push robotic manipulation beyond conventional visual and tactile sensing,” explained Daniel Seita, Hanyang Zhou, Wenhao Liu, and Haozhe Lou in an interview with Tech Xplore. “Many current multi-fingered robotic hands lack the integrated sensory systems needed for tasks that require thermal perception and responsive touch feedback.”
Designing a Human-Inspired Robotic Hand for Versatile and Precise Object Manipulation
The primary goal of Seita and his team was to design a flexible, multimodal robotic hand inspired by the human hand—capable of handling objects with both care and accuracy. By incorporating richer sensory input, their robotic hand can manage more delicate and complex object interactions across diverse settings, from domestic environments to industrial workplaces.
“At the outset, our main idea was to explore how to design systems that better reflect the way humans perceive the world in daily life. Our inspiration came from observing human behavior,” said Zhou and Liu.
“In the past, we worked with several off-the-shelf dexterous robotic hands. We found that most existing hardware prioritizes mechanical transmission, and the commercially available models didn’t meet the sensory needs required for many of our research goals. That experience led us to pursue a customizable, DIY hardware platform that would give us the flexibility to design and refine systems according to our specific vision.”
According to the research team, “The MOTIF hand is a sophisticated robotic platform that integrates multiple sensing technologies thermal, depth, RGB visual, inertial (IMU), and tactile into a single, dexterous unit. It builds upon the popular LEAP hand design, offering expanded sensing capabilities. Its standout features include a high level of environmental awareness, safe interaction with hot objects, and the ability to distinguish between items that look similar but have different physical characteristics.”
Evaluating Temperature-Aware Grasping Through Controlled Laboratory Experiments
To assess the capabilities of the MOTIF hand, the research team conducted two separate experiments in a controlled lab environment. The first focused on the hand’s ability to grasp objects while taking surface temperature into account, helping it avoid contact with dangerously hot areas.
The second experiment tested the hand’s capacity to distinguish between objects of identical shape but differing weights. For this task, the hand used simple fingertip flicking motions to estimate each object’s mass.
“The key achievement of our work is the successful integration of multiple sensor types into a single dexterous robotic hand, which greatly enhances its manipulation abilities,” said Lou. “Our results demonstrate that thermal sensing paired with 3D reconstruction supports safer grasping, while inertial sensing enables accurate mass classification.”
Promising Early Results and Future Plans for Real-World Deployment
In early trials, the MOTIF hand performed impressively grasping various items safely and estimating their weight with notable precision. Looking ahead, the researchers plan to refine the hand further and test it in broader scenarios to explore its potential in real-world applications.
Seita and his team believe the hand holds strong promise for use in domestic tasks, commercial kitchens, and industrial settings. With its ability to handle hot objects safely and perform fine manipulation, it could surpass current robotic systems in applications like cooking, welding, or fastening components tasks that demand both force sensitivity and adaptation to varying object temperatures.
“In the future, we plan to further improve the MOTIF hand by incorporating high-resolution fingertip sensors, such as the Digit 360, to achieve even more detailed tactile feedback,” Liu said. “We also intend to enhance our multimodal sensing algorithms and investigate new applications in complex real-world scenarios, including advanced in-hand manipulation and settings that demand intricate thermal and force-based interaction.”
“This work introduces a platform system,” Zhou added. “We presented it last week at the ISER 2025 conference in Santa Fe, where we had the opportunity to exchange ideas with many outstanding researchers. Our hope is that this project will encourage the community to explore additional sensing modalities in their own research.”
Read the original article on: Tech Xplore
Read more: New Study Shows AI Struggles With Overconfidence And UncertaintImage
