Elevating Gait Rehabilitation: Design of a Torque-Controlled Hip Flexion and Extension Exoskeleton

Cerebral Palsy (CP) occurs is about 1.7/1000 births. In Flanders, about 110 children are diagnosed with CP annually. About 46% of these children need walking aids or cannot walk at all. About 30% of them may benefit from exoskeleton treatment, which can provide intensive and reproducible training.

These types of robotic revalidation devices are not commercially available. Therefore, the cross-border INTERREG 2 Seas project, acronym ‘MOTION’, was launched. KU Leuven is co-applicant in the project and this doctorate will contribute to the ‘MOTION’ project. A powered, bilateral exoskeleton for hip flexion and extension will be developed, built and validated. Mechanical and control aspects will be addressed, as well as safety and clinical application.

Actuators are the most critical components of a powered exoskeleton. We will use actuators based on electric motors, because they provide the possibility to develop a fully portable device, by using Lithium-ion or Lithium-polymer batteries. The actuators will be of the series elastic type (SEA). SEA’s provide compliance, which is beneficial in exoskeleton applications. It acts as shock absorption and increases wearer comfort.

For actuator sizing, we will use numeric gait data of the target population. Load torque will be accounted for, as well as acceleration torque, for the human gait is a highly dynamic and cyclic motion.

CP children are capable of voluntary motion. Therefore, position control of the exoskeleton in not advisable. The low level control of the actuators will be torque-based. Using SEA’s, we can accomplish this by using the deflection of the series spring, without the need for force or torque sensors. High level control modes will be discussed with clinical experts and later implemented. They should be maximally adjustable tot the needs of the patient.

The hip exoskeleton will need real time information about gait events and gait phases as an input for its higher level controller. For this real time detection, we will rely on the work of MOBILAB, where researchers are implementing an algorithm based on IMU’s, attached to the shank.

A custom test bench will be used for developing the low level torque control and for performance benchmarking. The test bench consists of a torque measurement and a more powerful actuator, which can impose a movement profile on the SEA, similar to the real-world circumstances.

The physical interface between the patient and the exoskeleton will consist of a series of easily interchangeable orthotic shells. They must provide a tight fit and a stiff connection, without causing injuries or discomfort.

The safety of the device is of high priority. Software is known to be susceptible to programming bugs. Therefore, we will implement two low level safety mechanisms: mechanical end stops, which limit the powered hip flexion/extension motion to the anatomical limits, and an emergency stop, which acts upon the power stage of the actuators, and guaranties a safe torque off (STO).

During the project, the device will be tested on healthy subjects and patients. The results of these test will be used to fine-tune the device and its control.