Project Details
Description
Lower-extremity traumas account for nearly 345,000 hospitalizations registered in North America every year. About one third of these injuries are limb-threatening. Reconstructive surgery may prevent limb loss, but limb salvage patients often suffer chronic pain, muscle weakness, and diminished sensation, leading many to opt for late amputation. The prescription of ankle-foot orthoses (AFOs) combined with rehabilitation programs may promote functional recovery, reduce pain, and decrease late amputation rates. However, customizing the design of an AFO for best comfort, fit, and function – all-important factors to promote patient acceptance – involves substantial skilled labor and multiple design re-adjustments. Moreover, conventional (passive) AFOs partially restrict natural ankle movements, resulting in less efficient walking patterns. The emerging powered AFOs can serve as mobility aids to complement the patient's residual motor ability without inhibiting ankle movements. They can also serve as a clinical tool for gait rehabilitation, to facilitate the patient's transition to unassisted walking. Yet existing powered AFOs for gait rehabilitation do not fully address one of the key aspects of exercise-based therapy: the possibility of individualizing the interventions to the patient's motor abilities.
In response to the FY21 PRORP CTRA Focus Area 5 (Prosthetic and Orthotic Devices) the project proposes to address two major drawbacks of current orthotic technology for patients with reconstructed lower limb, namely the lack of automated procedures to fabricate AFOs that conform to a patient's body: (1) to improve comfort and the lack of control methods to self-tune the level of assistance of a powered AFO to the user's changing motor performance and (2) to promote their active participation in the therapeutic exercises and ultimately enhance rehabilitation outcomes. We will develop a new machine learning (ML) assisted design methodology for powered AFOs using design space exploration and optimality searches based on expert knowledge rules, as well as a new ML-based optimal policy search to enable self-tuning of the AFO's assistive forces. Obtaining patient-tailored orthotic designs will help reduce excessive pressure points in the wearer's skin and relative motions between the human limb and the orthosis due to poor fit, thereby improving comfort and, ultimately, patient acceptance/satisfaction. Self-tuning the AFO assistance to the wearer's motor abilities will discourage users' overreliance on the AFO and instead promote their active engagement in the walking exercises, which is a critical enabler of motor recovery. In Years 3 and 4 of the project, we will test the proposed design and control methods at Kessler Institute for Rehabilitation with a group of individuals who sustained lower leg reconstruction. First, we will assess safety, reliability, and comfort through a single-session study. Then, we will evaluate the clinical feasibility of the robotic intervention (at the exploratory level) by studying pre/post changes in participants' self-selected walking speed and other standardized functional outcomes, following a 6-week rehabilitation program.
The project will result in a new generation of robotic orthoses featuring patient-tailored form and function, which can be fabricated with substantially less labor involvement than traditional powered orthoses (and hence at a lower cost), using additive manufacturing technologies and the new ML-assisted design methodology. This research will also generate algorithmic support for future powered orthoses, to automatically tune the device's response to the patient's motor performance on the fly during overground walking exercises, without the need for specialized sensors in addition to those already embedded in most force-controlled powered orthoses. Taken together, these advances have the potential to impact the outcomes of future robot-assisted gait rehabilitation protocols, not only for limb salvage patients, but also for other populations (e.g., brain injury) for whom exercise-based therapy is critical for functional recovery. In the long run, the project will also pave the way for more accessible patient-tailored powered orthoses capable of administering individualized gait training exercises in the clinic or, potentially, in the comfort of the patient's home. Though the project will focus on the clinical feasibility of patient-tailored powered AFOs for exercise therapy, we expect the ML design methods to be also applicable to passive AFOs (e.g., to enable more efficient customization of current AFOs for limb salvage patients). Furthermore, the project will pave the way for next-generation powered AFO technology that can be used as a personal mobility aid. While the recovery of motor function in limb salvage patients represents a serious health concern in the U.S. and around the world, it is even more pressing for military and Veteran individuals owing to the higher incidence of lower-leg traumas in these subpopulations. The proposed patient-tailored orthoses have the potential to improve the lives of military, Veteran, and civilian personnel who sustained lower-leg reconstruction by enhancing their ability to obtain and maintain employment and participate fully in society and family life, thus promoting increased community integration and improved quality of life.
Status | Active |
---|---|
Effective start/end date | 1/01/21 → … |