Robots Atlas>ROBOTS ATLAS
Neurotechnologia

EMG

1890ActivePublished: 27 July 2026Updated: 27 July 2026Published
Key innovation
Direct measurement of muscle electrical activity, allowing motor intentions to be read and devices (prosthetics, robots) to be controlled with the myoelectric signal.
Category
Neurotechnologia
Abstraction level
Building block
Operation level
DataApplicationSystem
Use cases
Myoelectric prosthesis controlGesture recognition (human-machine interface)Teleoperation of robots and manipulatorsExoskeleton controlRehabilitation and neuromuscular diagnosticsLearning robots from demonstration

How it works

1) Electrodes (surface or needle) record the potential differences generated by recruited motor units. 2) The weak signal is amplified and sampled (analog-to-digital conversion). 3) The signal is filtered and cleaned of artifacts (crosstalk, motion, mains). 4) Features are extracted (e.g. RMS, MAV, spectral features) and then classified (e.g. by a machine-learning model) to recognize a gesture or derive a control command.

Problem solved

Enables reading motor intentions directly from the muscles, even before a full movement occurs, providing a natural, touch-free control signal for prosthetics, exoskeletons and robots.

Components

ElectrodesSignal acquisition

Sensors recording muscle potentials: surface (sEMG) on the skin or needle inside the muscle.

Surface electrodes (sEMG)Non-invasive, for gestures and control.
Needle electrodesInvasive, for clinical diagnostics.

Official

Amplifier and ADCAmplification and digitization

Amplifies and digitizes the weak myoelectric signal.

Official

Signal processingSignal analysis

Filtering, artifact removal and feature extraction (RMS, MAV, spectrum) from the EMG signal.

Official

Implementation

Reference implementations
OpenBCI (biosensing, EMG)
OpenBCI
Implementation pitfalls
Crosstalk and electrode shiftHigh

Signal from neighboring muscles (crosstalk) and electrode shift reduce gesture-recognition accuracy.

Fix:More channels, careful placement, calibration and shift-robust models.
Inter-individual variability and muscle fatigueMedium

EMG patterns depend on the person, posture and fatigue, requiring calibration and adaptation.

Fix:Per-user calibration, transfer learning, adaptive models.

Evolution

1890
Marey coins the term "electromyography"
Inflection point

Étienne-Jules Marey makes early recordings of muscle electrical activity and introduces the term electromyography.

1929
Foundations of modern EMG (Adrian & Bronk)

Adrian and Bronk develop motor-unit recording and the concentric electrode, enabling clinical EMG.

2014
Wearable sEMG armbands for gesture control

Surface EMG becomes popular in wearable armbands (e.g. Myo) for touch-free control and human-machine interfaces.

Hyperparameters (configurable axes)

Electrode typeHigh

Surface (sEMG) vs needle.

surface (sEMG)For gestures/control.
needleFor diagnostics.
Number of channelsHigh

Number of electrodes / measurement sites.

8EMG armband.
Sampling rateMedium

sEMG typically up to ~1-2 kHz (band ~20-450 Hz).

1000 HzTypical.

Computational complexity

Time complexity: O(kanaly × probki) na okno. Space complexity: O(kanaly × bufor).

Compute bottleneck

Signal quality, crosstalk and artifacts

Muscle crosstalk, electrode shift, sweating and inter-individual variability make reliable gesture classification difficult.

Depends on
Rozmieszczenie/kontakt elektrodZmiennosc osobnicza

Execution paradigm

Primary mode
Dense

Continuous, parallel sampling of all channels.

Activation pattern
All paths active
Routing mechanism

Continuous signal acquisition, no routing.

Parallelism

Parallelism level
Fully parallel

Channels and features can be processed independently/in parallel.

Scope
Inference
Constraints
!Acquisition and control must run in real time.

Hardware requirements

Primary

EMG acquisition and processing are lightweight; they run on CPU/embedded devices. The key is dedicated analog hardware (electrodes + amplifier).