MIT Active Joint Brace Research











Our patent pending technology is based on determining desired force via non-invasive measurements of surface electromyogram (sEMG). These signals are filtered and processed to produce an estimate of what the patient is "trying" to do (i.e. - produce enough force to lift their arm), even if they cannot do it themselves. The force estimate is scaled by an adjustable amount, and is used to control a force feedback system that applies force to a given joint. Since the device only applies an amount of force proportional to what the user exerts, this device will not cause disuse-atrophy.

Surface EMG based control is a conventional technology. Its most popular application has been in the field of powered prosthetics, such as the Utah Arm. There are many different EMG processing algorithms, ranging from simple averaging schemes, to more elaborate filters involving AI techniques and advanced signal processing. Almost all application of sEMG based control is for prosthesis control in otherwise healthy individuals.

Recently, people have begun to use the EMG signal in FES (Functional Electrical Stimulation). In that case, the EMG signal is detected, and the muscle is subsequently stimulated by an external electrical device to cause contraction and movement. FES is commonly used to provide stimulation to muscle following a neurological injury. The stimulation prevents the muscle from wasting away entirely due to lack of use. Other, implantable, FES systems provides the user control over paralyzed parts of their body using the sEMG signals from parts they have control over. The drawback to FES is that it is very invasive, injecting current into the muscles through the skin (or requiring implants). When the stimulators are external, the process can cause significant skin irritation and is not suitable for extended use

In the case of the active joint brace, the functionality is different than both of the existing technologies. The powered prostheses treat a class of individuals on whom this device wouldn't work (missing limbs). With respect to FES, the technology we propose is completely noninvasive. There is no current or probe that breaks the surface of the skin. In addition, FES stimulates whatever muscle still exists and only provides the amount of strength that can be drawn from the muscles, contrary to our proposed design.

Over the past year we have developed a prototype system that senses desired joint torque from surface EMG, and applies a proportional amount of assistance about the elbow joint. We are currently evaluating functionality on spinal cord injury patients that have suffered from high level injuries and have almost no voluntary arm control.

   © 2004 MIT Active Joint Brace