RIT Team 3D Prints a Prosthetic Hand With Built-In Sensing

A team at Rochester Institute of Technology has built a prosthetic hand prototype that combines four separate research disciplines into one printable device. The result uses multi-material 3D printing, bioprinting, piezoelectric sensors, and mechanical modeling to address problems that usually get solved one at a time.

Most 3D-printed prosthetics today rely on a single material for the whole structure. That keeps costs low but limits durability, grip strength, and how well the device fits an individual's limb. RIT's approach prints rigid skeletal sections in PLA alongside soft silicone joints made from PDMS, giving the hand both structure and flexibility without assembly.

The piezoelectric sensors embedded in the design let the hand sense pressure and joint movement. That is a step beyond simple mechanical gripping. The team hopes future versions can use that data to adjust grip strength automatically or trigger other functions.

Closing the Integration Gap

Research groups tend to specialize. One lab masters bioprinting. Another studies sensors. A third optimizes slicer settings for soft materials. Those advances stay separated unless someone forces them to work together. RIT's contribution was running all four through one coordinated workflow.

Salman Pervaiz, one of the researchers, put it plainly: multi-material printing and smart sensors were both available before this project. The gap was bringing them into the same build.

Other recent work points in similar directions. Johns Hopkins and Florida Atlantic University researchers have demonstrated soft robotic fingers with tactile sensing. Chinese and Japanese teams have 3D-printed self-powered robotic digits that do not need batteries. RIT's prototype is notable for trying to combine durability, sensing, and customization in a single manufacturing process.

What Is Still Missing

The RIT hand is a proof of concept, not a finished product. The sensors have not yet been tested for real-time joint tracking. The viscosity models cover only PDMS before it cures, so a full thermoset model is still needed. The long-term bond between the silicone and the PLA skeleton under repeated stress has not been proven.

The team plans to improve sensor sensitivity, test new material combinations, and bring in machine learning to reduce the amount of physical trial and error. If those steps pan out, the design could move toward commercialization.