Utilizing EEG Signal Data and Motion to Aid in Prosthetic Hand Motion

Abstract

Current prosthetic arm technologies are often difficult to use intuitively by amputees, require invasive surgical procedures, and can be extremely costly with prices ranging from $20,000 to $80,000. To address these challenges, the engineering goal of this project aims to design a smart, low-cost, mind-controlled transhumeral prosthesis by integrating the brain-interfacing capabilities of electroencephalography (EEG), the economical means of 3D printing technology, and gesture-detecting attributes of an accelerometer-gyroscope. Single-channel brain signals are transmitted through Bluetooth to be interpreted by a novel EEG decoding algorithm and head gestures from the inertial measurement unit actuate movement within the arm. Force sensitive resistors were employed to regulate force control in real-time to optimize grasp type. An LCD screen is integrated within the arm’s design to display the type of touch it is exerting on an object. The arm itself was printed with an original design utilizing PLA plastic filament making it extremely durable and lightweight. After thoroughly testing the prosthesis, the novel EEG decoding system boasts an accuracy of 94.7% and a user cognition to machine delay of 1.64 + - 0.37 seconds. The inertial measurement unit system recognizes user gestures with a delay of .136 seconds. With a bill of materials approximately $375 USD and ability to be moved in 4 degrees of motion, this novel upper limb prosthesis serves as a promising alternative to existing units on the market. The algorithms developed within this project have a wide range of use within other brain control interface projects.