过滤流数据减少噪声，卡尔曼滤波c#

本文关键字：卡尔曼滤波 噪声 数据 过滤 | 更新日期: 2023-09-27 18:05:29

我从一个惯性传感器流数据到c#应用程序。数据有点嘈杂，所以我需要添加一个过滤器来平滑它。当给定数组时，我有一个卡尔曼滤波器实现，它工作得很好，但我不知道如何在恒定数据流上使用它。

我有:

double sensorData; //the noisy value, constantly updating from another class.

过滤器:

public static double[] noisySine = new double[20] { 40, 41, 38, 40, 45, 42, 43, 44, 40, 38, 44, 45, 40, 39, 37, 41, 42, 70, 44, 42 };
    public static double[] clean = new double[20];
      public static void KalmanFilter(double[] noisy)  
            {                  
                double A = double.Parse("1"); //factor of real value to previous real value
                // double B = 0; //factor of real value to real control signal
                double H = double.Parse("1"); 
                double P = double.Parse("0.1");
                double Q = double.Parse("0.125");  //Process noise. 
                double R = double.Parse("1"); //assumed environment noise.
                double K;
                double z;
                double x;
                //assign to first measured value
                x = noisy[0];
                for (int i = 0; i < noisy.Length; i++)  
                {
                    //get current measured value
                    z = noisy[i];
                    //time update - prediction
                    x = A * x;
                    P = A * P * A + Q;
                    //measurement update - correction
                    K = P * H / (H * P * H + R);
                    x = x + K * (z - H * x);
                    P = (1 - K * H) * P;
                    //estimated value
                    clean[i] = x;
                    Console.WriteLine(noisy[i] + " " + clean[i]);
                }
            }

我怎么能流一个双精度，而不是一个数组，并返回一个(过滤)双精度?

谢谢。

过滤流数据减少噪声，卡尔曼滤波c#

创建这个类:

public class KalmanFilter
{
    private double A, H, Q, R, P, x;
    public KalmanFilter(double A, double H, double Q, double R, double initial_P, double initial_x)
    {
        this.A = A;
        this.H = H;
        this.Q = Q;
        this.R = R;
        this.P = initial_P;
        this.x = initial_x;
    }
    public double Output(double input)
    {
        // time update - prediction
        x = A * x;
        P = A * P * A + Q;
        // measurement update - correction
        double K = P * H / (H * P * H + R);
        x = x + K * (input - H * x);
        P = (1 - K * H) * P;
        return x;
    }
}

并使用类:

KalmanFilter filter = new KalmanFilter(1, 1, 0.125, 1, 0.1, noisySine[0]);
for (int i = 0; i < noisy.Length; i++) clean[i] = filter.Output(noisySine[i]);

试试下面的代码

using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.IO;
namespace ConsoleApplication1
{
    class Program
    {
        static void Main(string[] args)
        {
            double[] input = {1.1,2.2,3.3,4.4};
            byte[] bArray = input.Select(x => BitConverter.GetBytes(x)).SelectMany(y => y).ToArray();
            MemoryStream inStream = new MemoryStream(bArray);
            long length = inStream.Length;
            byte[] outArray = new byte[length];
            inStream.Read(outArray, 0, (int)length);
            List<double> output = new List<double>();
            for (int i = 0; i < bArray.Length; i += 8)
            {
                output.Add(BitConverter.ToDouble(outArray,i));
            }
        }
    }
}

这是你如何修改你的代码来传入一个双精度类型，并返回一个过滤后的双精度类型。

  public static void KalmanTest()
      {
          double[] noisySine = new double[20] { 40, 41, 38, 40, 45, 42, 43, 44, 40, 38, 44, 45, 40, 39, 37, 41, 42, 70, 44, 42 };
          for (int i = 0; i < noisySine.Length; i++)  
          {
                Console.WriteLine(noisySine[i] + " " + KalmanFilter(noisySine[i]));
          }
      }

  // assign default values
  // for a new mwasurement, reset this values
  public static double P = double.Parse("1");  // MUST be greater than 0
  public static double clean = double.Parse("0"); // any value
  public static double KalmanFilter(double noisy)  
        {                  
            double A = double.Parse("1"); //factor of real value to previous real value
            // double B = 0; //factor of real value to real control signal
            double H = double.Parse("1"); 
            double Q = double.Parse("0.125");  //Process noise. 
            double R = double.Parse("1"); //assumed environment noise.
            double K;
            double z;
            double x;
                //get current measured value
                z = noisy;
                //time update - prediction
                x = A * clean;
                P = A * P * A + Q;
                //measurement update - correction
                K = P * H / (H * P * H + R);
                x = x + K * (z - H * x);
                P = (1 - K * H) * P;
                //estimated value
                clean = x;
                return clean;
        }

注意:有一个bug。当此代码迭代时，P很快变成接近R/100000的值，并且这种行为与噪声无关，因为在P计算中没有引用噪声或稳定读数。干净的代码看起来像一个低通滤波器:

  // assign default values
  public static double clean = double.Parse("0"); // any value
  public static double KalmanFilter(double noisy)  
        {                  
            double K = double.Parse("0.125");  // noise 0 < K < 1
            clean = clean + K * (noisy - clean);
            return clean;
        }
相关文章：