Algorithm enhances gait patterns in robotic knee prosthetics

The research project aimed at improving how robotic knee prosthetics support natural motion in users, including women and men with above-knee amputations. The newly developed algorithm combines two processes that both refine the movement of the prosthetic device and help the user’s body move more naturally, addressing common health issues such as hip discomfort or lower back pain.

Above-knee amputations alter how other parts of the body move, often affecting trunk stability and hip function. Earlier robotic prosthetics primarily focused on replicating the missing joint’s mechanics.
“Algorithms designed to improve the behavior of robotic prosthetics are not new – but this is the first algorithm that also holistically improves the physical behavior of the person interacting with those prosthetics,” says Varun Nalam, co-lead author of the study.
The approach builds on an earlier system using reinforcement learning to quickly tune powered knee prostheses without extensive clinical adjustments. The current work integrates inverse reinforcement learning to incorporate both the prosthetic’s action and the user’s movement, offering a more comprehensive method for supporting gait patterns in daily activities of patients, rehabilitation teams and clinical specialists.

The robotic knee includes sensors that capture movement data, while users’ hip motion was also monitored via external sensors.
“The new algorithm essentially accounts for the movement of both joints – the prosthetic knee and the user’s hip – and adjusts the behavior of the prosthetic knee to help the user exhibit their natural hip movement,” explains Nalam.

To test the system, the team recruited five participants: two individuals with above-knee amputations and three without amputations. Each participant completed walking tasks using software that either relied on the earlier control system or the new combined algorithms.
“The main finding here was that incorporating the new algorithm improved hip range of motion for all five subjects, which demonstrates that it can make a difference for hip health,” Nalam says. Participants also showed gait characteristics associated with more natural walking, such as longer step lengths.

“From a practical standpoint, next steps include working with clinicians to see how it affects user well-being over time,” says Helen Huang. The team also plans to collaborate with companies producing robotic prosthetics to explore integration opportunities. The researchers note that the algorithm may eventually support additional aspects of human movement, such as trunk control or symmetrical gait training.

COMPAMED-tradefair.com; Source: North Carolina State University