Shadow Robot and the Fraunhofer Institute for Factory Operation and Automation IFF have started a six-month project on tactile robotic manipulation in manufacturing, announced on October 20, 2026. The work pairs dexterous robotic hands with tactile sensing, teleoperation and advanced robot control. Its target is assembly and handling tasks that factories still give to people because machines cannot do them reliably.
What the project covers
The partners will examine tasks where success depends on precise interaction with individual components. These are jobs that are hard to automate today and are often performed by hand. The project will test whether dexterous hands can offer more flexibility and more physical interaction than conventional industrial grippers.
Tactile sensing is the central topic. The partners want to learn how extra sensing can give a clearer picture of contact and physical interaction during complex assembly, and how it can complement data from other sensors while a robot manipulates an object.
The project is also meant to demonstrate what is achievable when dexterous hardware, tactile sensing and advanced software control are combined. It will identify the engineering approaches, the challenges and the costs that would have to be addressed before such systems could be taken further. The announcement gives no cost figures and does not name the tasks or components involved.
Who contributes what
Shadow Robot is based in London and Fraunhofer IFF in Magdeburg, Germany. Both are member organisations of the euRobotics aisbl. Shadow Robot is supplying dexterous robotic hardware, system integration and teleoperation capabilities. Fraunhofer IFF is contributing expertise in cognitive robotics, challenging manipulation tasks and industrial robotics consultancy. Its Department of Robotic Systems works on dexterous and tactile-driven manipulation and real-time robot control.
Dr. José Saenz, Group Manager at Fraunhofer IFF, said Shadow Robot already had a complete setup for teleoperating a five-finger hand with tactile sensing. That let his team skip integration work and start on the manipulation problems directly, he said.
Rich Walker, Director of Shadow Robot and an elected Director of euRobotics aisbl, said he has known the Fraunhofer IFF team for over a decade through euRobotics activities. He described the project as a chance to work with them directly on real manufacturing challenges.
The gap in automated production
Fraunhofer IFF frames the work around problems that industrial customers have not found solutions for. Its approach to automated manufacture draws on the German tradition of high-quality handicraft. The institute says that even in highly automated production lines, 10 to 20 percent of tasks are still performed manually because the technology to automate them is missing.
That figure is the institute's own estimate and the announcement does not say how it was measured. It does define the scale of the opportunity the project is aimed at: a minority of tasks, but ones that keep human labour in otherwise automated lines.
Hands on cobot arms, not humanoids
The project does not involve humanoid robots. The dexterous hands will be mounted on conventional collaborative industrial robot arms in manufacturing environments. The partners will also test how much dexterity different tasks require, including whether five-fingered hands are necessary for particular applications.
The project is tied to the euRobotics Vienna Statement, which stresses the role of humans in future production systems. In that view, robots are meant to complement human capabilities rather than simply replace them. Teleoperation fits that framing, since it lets a person control a dexterous system directly.
Analysis: why it matters and what is unproven
The project is notable for its restraint. Much attention in robotics currently goes to full humanoids. This effort instead takes a familiar factory platform, the collaborative arm, and changes only the end effector and its sensing. That is a narrower and more testable proposition, and it matches the way many plants would actually adopt new hardware.
The question about five fingers is the most useful part of the plan. A five-finger hand is costly and mechanically complex compared with a conventional gripper. If the partners find that some tasks need fewer fingers, the result would help buyers judge where dexterity pays for itself. If they find the opposite, it would show where simpler grippers hit a ceiling. Either outcome is informative, and the announcement commits to examining it.
The comparison with existing tools is also clear from the source. Conventional industrial grippers are the baseline, and the project asks whether dexterous hands give more flexibility and physical interaction than they do. The announcement names no competing hand or rival programme, so a broader comparison cannot be drawn from it.
Several things remain unproven. The announcement reports no results, no success metrics and no target tasks. It does not state how tactile data will be used in control, or how performance will be compared with manual work. The cost analysis is promised, not delivered. The 10 to 20 percent figure for manual tasks is a claim about the whole of automated production, and it is unclear how much of that share dexterous hands with touch sensing could actually take over.
Six months is also a short span. It is enough to build a demonstration and map the obstacles, which is what the partners say they will do. It is not enough to show that such a system holds up in a production line. Readers should treat the outcome as a feasibility study of tactile robotic manipulation, and judge it by whether it produces concrete cost and engineering findings that others can use.
