RepMotion

A Swift library for detecting repetitive motion patterns using Apple Watch accelerometer and gyroscope data. Count exercise reps, measure workout intensity, and auto-calibrate detection parameters – all in real time.

Features

• Real-time rep detection with adaptive thresholding and peak analysis
• Auto-calibration that optimizes detection parameters from labeled recordings
• Power estimation with rolling average, rolling max, and absolute modes
• Distance estimation based on rep power output
• Diagnostic stream for debugging and visualizing the detection pipeline
• Fixture replay for testing without hardware

Quick Start

Add RepMotion to your Package.swift:

dependencies: [
    .package(url: "https://github.com/thefern/RepMotion.git", from: "0.1.0")
]

Add the libraries you need:

.target(
    name: "YourApp",
    dependencies: [
        "RepMotionCore",
        "RepMotionDetection",
        "RepMotionCapture"
    ]
)

Minimal Example

import Combine
import RepMotionCore
import RepMotionDetection
import RepMotionCapture

let capture = MotionCaptureService(sampleRate: 50)
let detector = RepDetectionService()
var cancellables = Set<AnyCancellable>()
var repCount = 0

capture.samplePublisher
    .sink { detector.processBatch($0) }
    .store(in: &cancellables)

detector.repPublisher
    .sink { event in
        repCount += 1
        print("Rep \(repCount) — confidence: \(event.confidence)")
    }
    .store(in: &cancellables)

capture.start()

That’s it. Ten lines to count reps from live sensor data.

Platform Support

Platform	Minimum Version
iOS	17.0
watchOS	10.0
macOS	14.0

Modules

RepMotion ships as three independent libraries so you can import only what you need.

Module	Purpose	Dependencies
RepMotionCore	Data types: MotionSample, RepEvent, RepCalibration, etc.	None
RepMotionDetection	Detection engine, calibration, power and distance estimation	RepMotionCore
RepMotionCapture	Live sensor capture via CoreMotion and fixture replay	RepMotionCore

Interactive Demo

Try the Signal Explorer – adjust detection parameters with sliders and watch the pipeline process a simulated exercise session in real time. See how alpha, threshold, refractory period, axis, and polarity affect rep detection.

Documentation

• Getting Started – installation, setup, first integration
• Architecture – how the detection pipeline works
• Interactive Signal Explorer – live parameter tuning visualization
• Basic Detection – using RepDetectionService
• Calibration – auto-tuning with CalibrationEngine
• Power and Distance – workout intensity metrics
• Testing – using FixtureMotionProvider for tests
• Diagnostics – debugging detection with the diagnostic stream
• API Reference – all public types and methods

License

MIT. See LICENSE.