The Design Evolution of a Lower Extremity Exoskeleton Device for Leg Muscle Rehabilitation

Publication Date


Document Type

Conference Proceeding

Publication Title

ASME International Mechanical Engineering Congress and Exposition, Proceedings (IMECE)






The proposed design of the robotic-assisted pneumatic knee-brace is activated by using surface electromyographic (EMG) sensors to activate motion for rehabilitation exercises. These muscles are actuators powered by a pneumatic system that expands or contracts depending on the compressed air entering the system. An Arduino-based controller performs the pneumatic control. The first design consisted of two fluidic muscles mounted on a human leg where several EMG sensors were placed on the upper part of the leg randomly to activate the controllers. This device is currently further modified for patients who have undergone total kneed replacement surgery and need leg rehabilitation exercises to recover. In this design, multiple fluidic muscles are incorporated to minimize the reaction time of the device. The brace is designed to mount on a knee of a sitting person and consists of four bars (two bars are mounted on the thighs and the remaining two on the calves). EMG sensors are placed on six muscles, including Vastus Medialis, Vastus Lateralis, Rectus Femoris, Semitendinosus, Biceps Femoris, and Semimembranosus. The new knee brace is fully characterized by mounting it on various sitting persons' legs. In the current experiment, a healthy person's legs are targeted. The leg motion is being investigated as a function of air pressure inside the fluid muscles and the location of EMG sensors on the legs for various age groups. The current presentation will include the systematic design improvement in our knee-brace device and its applications. The software that processed the EMG sensors was modified to reject the input noise to give an accurate and fast response. The details of the software and its processing will be discussed in greater detail. The device characterization results will be summarized for practical applications.

Funding Sponsor

RIKEN Brain Science Institute


assistive joint, rehabilitation, robotics, voice control


Mechanical Engineering