EMG-Based Control of a Multi-Joint Robot for Operating a Glovebox
Toshio Tsuji (Hiroshima University, Japan), Taro Shibanoki (Ibaraki University, Japan) and Keisuke Shima (Yokohama National University, Japan)
Copyright: © 2015
This chapter describes a control method for a multi-joint robotic manipulator using Electromyogram (EMG) signals for operating a glovebox. The system uses a Probabilistic Neural Network (PNN) to estimate the user's intended motion from EMG patterns, and generates a control command for the glovebox and robotic manipulator corresponding to the estimated motions. The user can therefore control the manipulator as well as various functions of the glovebox system through his/her EMG signals while performing some manual operations through gloves. With this system, the authors produce intuitive control of the glovebox with the robotic manipulator. The authors confirm the effectiveness of the proposed system with an experiment using the developed prototype.
Research and development activity in industry and many fields such as chemistry has become more technical and complex in recent years, and it is important to secure the safety of researchers and provide them with an optimal environment suited to the purpose of the studies and/or experiments. The glovebox (or glove box) is a piece of equipment that is widely used in various research institutions to overcome these problems.
In this chapter, we outline a novel glovebox operation system with a multi-joint robotic manipulator controlled by an electromyogram (EMG) signal. The proposed system estimates the user’s intended motions from extracted EMG patterns for each motion using a probabilistic neural network (PNN) called a log-linearized Gaussian mixture network (LLGMN). The user can then control the environment in the glovebox in addition to the room environment based on the estimated motion, and also simultaneously perform tasks with the robotic manipulator in the box based on an EMG signal. We designed a set of experiments to demonstrate the effectiveness of the developed prototype.
Key Terms in this Chapter
Environmental Control System (ECS): A system that can control devices such as domestic appliances or electrical beds.
Probability Density Function (PDF): A statistical measure that defines a probability distribution for a random variable. Teleoperation: real-time operation of a system from a remote place. Distance varies from tens of centimeters to millions of kilometers.
Glovebox: (Also known as glove box) A sealed container with protective gloves set into the side of the compartment (transparent panel) used for handling hazardous substances preventing contamination of the product, the environment, and the operator.
Layer-Based Selection: An operation method that groups operation command in each layer. The desired operation can be carried out by repeatedly selecting and executing commands. It is often used in systems like ATMs.
Electromyogram (EMG): A record of electrical activity in the muscle. An electrical potential is also known as an EMG signal, which arises from muscle activity and contains information about parameters such as intended motion and force.
Probabilistic Neural Network (PNN): an artificial neural network that can estimate the probability density function of a given set of data. It can compute nonlinear decision boundaries and is used for classification problems.