KalmanKit

A high-precision Kalman Filter for smoothing GPS coordinates (latitude, longitude, altitude) and reducing noise from CLLocationManager updates in iOS applications.

Overview

Kalman Filters are recursive estimators ideal for noisy sensor data—especially location data from mobile devices. This implementation models position and velocity in a six-dimensional state space and adjusts predictions over time to improve location tracking fidelity.

Features

• ✅ Supports 6D state modeling: position and velocity for latitude, longitude, and altitude
• ✅ Tunable sensor and acceleration noise parameters
• ✅ Swift-native and ready for integration with CLLocationManager
• ✅ Designed for real-time, recursive updates

Requirements

• iOS 14.0+ / macOS 11+ / tvOS 14+ / watchOS 7+ / macCatalyst 14+
• Swift 6.1+

Installation

Use Swift Package Manager:

// swift-tools-version:6.1

import PackageDescription

let package = Package(
    name: "YourPackageName",
    dependencies: [
        .package(url: "https://github.com/Stof83/KalmanKit.git", from: "1.0.0")
    ],
    targets: [
        .target(
                name: "YourTargetName",
                dependencies: [
                    .product(name: "KalmanKit", package: "KalmanKit")
                ]
        )
    ]
)

---

Documentation

init Parameters:

• initialLocation: The initial CLLocation object representing the starting location of the filter. This is used to set the initial state of the filter (e.g., position and velocity).

Public Methods:

process(_ currentLocation: CLLocation) -> CLLocation

• Parameters: A CLLocation object representing the latest GPS location.
• Returns: A corrected CLLocation object with reduced noise.
• Description: This method processes a new location update and returns a location with reduced noise using the Kalman filter algorithm.

setSensorNoiseCovariance(_ newValue: Double)

• Parameters: A new value (Double) to adjust the sensor noise covariance.
• Description: This setter function allows modification of the sensor noise covariance. Higher values result in smoother paths but may lag behind sharp movements.

reset(_ newStartLocation: CLLocation)

• Parameters: A CLLocation object representing the new location to reset the filter with.
• Description: Resets the Kalman filter with a new initial location. Useful for restarting tracking when signal is lost or tracking resumes after a pause.

---

Usage

1. Initialization

let filter = KalmanFilter(location)

Pass an initial CLLocation object to bootstrap the filter.

2. Processing New Location Updates

let correctedLocation = filter.process(newLocation)

Pass each new GPS reading to the filter. The return value is a noise-reduced CLLocation.

3. Resetting the Filter

filter.reset(newStartlocation)

Useful for resetting state if tracking resumes after a pause or signal loss.

Configuration

Sensor Noise

filter.setSensorNoiseCovariance(50.0)

Higher values smooth the output more but may lag behind sharp movements.

Example

import CoreLocation
import KalmanKit

class LocationHandler: NSObject, CLLocationManagerDelegate {
    private var kalmanFilter: KalmanFilter?
    private var shouldResetFilter = false
    private let locationManager = CLLocationManager()

    override init() {
        super.init()
        locationManager.delegate = self
        locationManager.requestWhenInUseAuthorization()
        locationManager.startUpdatingLocation()
    }

    func locationManager(_ manager: CLLocationManager, didUpdateLocations locations: [CLLocation]) {
        guard let currentLocation = locations.first else { return }

        if kalmanFilter == nil {
            kalmanFilter = KalmanFilter(currentLocation)
            return
        }

        if shouldResetFilter {
            kalmanFilter?.reset(currentLocation)
            shouldResetFilter = false
            return
        }

        if let smoothedLocation = kalmanFilter?.process(currentLocation) {
            print("Smoothed Coordinate: \(smoothedLocation.coordinate)")
        }
    }
}

License

This SDK is provided under the MIT License. See LICENSE for details. Feel free to use, modify, and distribute it as per the terms of the license.

---

Contributing

Contributions are welcome! Please open issues and submit pull requests for features, fixes, or improvements.

---