1D Kalman.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "See 19 in https://github.com/rlabbe/Kalman-and-Bayesian-Filters-in-Python/blob/master/04-One-Dimensional-Kalman-Filters.ipynb\n",
    "\n",
    "Another useful ref https://stackoverflow.com/questions/33384112/kalman-filter-one-dimensional-several-approaches"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "from random import Random\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "from collections import namedtuple\n",
    "gaussian = namedtuple('Gaussian', ['mean', 'error']) # var\n",
    "\n",
    "def predict(posterior, process):\n",
    "    \"\"\"Predict from posterior and process model.\"\"\"\n",
    "    posterior_mean, posterior_error = posterior\n",
    "    process_mean, process_error = process\n",
    "    predicted_state = posterior_mean + process_mean\n",
    "    predicted_error = posterior_error + process_error\n",
    "    return gaussian(predicted_state, predicted_error)\n",
    "\n",
    "def update(prior, measurement):\n",
    "    \"\"\"Calc posterior from prior and measurement.\"\"\"\n",
    "    prior_mean, prior_error = prior\n",
    "    measurement_mean, measurement_error = measurement\n",
    "    \n",
    "    residual = measurement_mean - prior_mean\n",
    "    kalman_gain = prior_error / (prior_error + measurement_error)\n",
    "\n",
    "    posterior_mean = prior_mean + kalman_gain*residual\n",
    "    posterior_error = (1 - kalman_gain) * prior_error\n",
    "    return gaussian(posterior_mean, posterior_error)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-0.8908232124563659"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "random = Random()\n",
    "random.gauss(mu=0, sigma=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x107b909b0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "randoms = []\n",
    "for i in range(1000000):\n",
    "    randoms.append(random.normalvariate(mu=0, sigma=1))\n",
    "plt.hist(randoms, bins=300);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The process variance (`process_var`) tells the filter how much the system is changing over time. If you set this number too low the filter will not be able to react to changes that are happening"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def volt(voltage, std):\n",
    "    return voltage + (random.gauss(mu=0, sigma=1) * std)\n",
    "\n",
    "voltage_std = .13  # Need to configure/estimate ORIG\n",
    "process_var = .05**2  # Need to configure/estimate ORIG\n",
    "#process_var = .5**2  # Need to configure/estimate HIGH\n",
    "actual_voltage = 16.3\n",
    "\n",
    "x = gaussian(25., 1000.) # initial state\n",
    "process_model = gaussian(0., process_var)\n",
    "\n",
    "N = 50\n",
    "time = [i for i in range(N)]\n",
    "readings = [volt(actual_voltage, voltage_std) for i in range(N)]\n",
    "ps = []\n",
    "estimates = []"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "for z in readings:\n",
    "    prior = predict(x, process_model)\n",
    "    x = update(prior, gaussian(z, voltage_std**2))\n",
    "\n",
    "    # save for latter plotting\n",
    "    estimates.append(x.mean)\n",
    "    ps.append(x.error)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10795fcf8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,5))\n",
    "plt.plot(time, readings, 'bo', label='readings') # Dummy reading\n",
    "plt.plot(time, estimates, 'ro--', label='estimates')\n",
    "plt.legend()\n",
    "#plt.ylim(16, 17)\n",
    "plt.xlabel('step')\n",
    "plt.ylabel('volts')\n",
    "plt.show()\n",
    "    \n",
    "#plt.plot(ps)\n",
    "#plt.title('Variance')\n",
    "#print('Variance converges to {:.3f}'.format(ps[-1]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Not bad, but note that our process model is a gaussian of mean zero. If there were drift, we would need to include this in the process model."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Now add in a drift\n",
    "Experiment with the process variance `process_var`, vary between 0 - 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def volt(drift, voltage, std):\n",
    "    return drift + voltage + (random.gauss(mu=0, sigma=1) * std)\n",
    "\n",
    "voltage_std = .7  # Need to configure/estimate ORIG\n",
    "process_var = .5**2  # Need to configure/estimate ORIG\n",
    "#process_var = .5**2  # Need to configure/estimate HIGH\n",
    "actual_voltage = 16.3\n",
    "\n",
    "x = gaussian(25., 1000.) # initial state\n",
    "process_model = gaussian(0., process_var)\n",
    "\n",
    "N = 50\n",
    "time = [i for i in range(N)]\n",
    "readings = [volt(i/3, actual_voltage, voltage_std) for i in range(N)]\n",
    "ps = []\n",
    "estimates = []"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "for z in readings:\n",
    "    prior = predict(x, process_model)\n",
    "    x = update(prior, gaussian(z, voltage_std**2))\n",
    "\n",
    "    # save for latter plotting\n",
    "    estimates.append(x.mean)\n",
    "    ps.append(x.error)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x115937f98>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(17,7))\n",
    "plt.plot(time, readings, 'bo', label='readings') # Dummy reading\n",
    "plt.plot(time, estimates, 'ro--', label='estimates')\n",
    "plt.legend()\n",
    "#plt.ylim(16, 17)\n",
    "plt.xlabel('step')\n",
    "plt.ylabel('volts')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This doesn't look too bad\n",
    "\n",
    "## Kalman 1D class\n",
    "Lets create a simple class that we could use in Home-Assistant"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "Class for a 1D Kalman filter. The user will have to configure the filter sensitivity (range 0-1),\n",
    "and the measurement standard deviation - if the process is stationary this is easy to calculate using \n",
    "one of the stats sensors.\n",
    "\"\"\"\n",
    "\n",
    "class kalman:\n",
    "    \"\"\"Class for a 1D Kalman filter.\"\"\"\n",
    "    \n",
    "    def __init__(self, sensitivity):\n",
    "        \"\"\"Initialise the class\"\"\"\n",
    "        self._gaussian = namedtuple('Gaussian', ['mean', 'var'])\n",
    "        self._process_model = self._gaussian(0., (2*sensitivity)**2) # sensitivity should be in range 0-1\n",
    "        self._prior = None\n",
    "        self._state = None\n",
    "        self._measurement_error = None\n",
    "    \n",
    "    def process_measurement(self, measurement):\n",
    "        \"\"\"Process a measurement\"\"\"\n",
    "        if self._state is None and measurement is not None:\n",
    "            self._measurement_error = 0.1 * measurement # Just guess variance to be 10%\n",
    "            self._state = self._gaussian(measurement, self._measurement_error) # initial state and variance\n",
    "        else:\n",
    "            self._prior = self._predict(self._state, self._process_model)\n",
    "            self._state = self._update(self._prior, self._gaussian(measurement, self._measurement_error))\n",
    "\n",
    "    def _predict(self, posterior, process_model):\n",
    "        \"\"\"Predict from posterior and process model.\"\"\"\n",
    "        posterior_mean, posterior_error = posterior\n",
    "        process_mean, process_error = process_model\n",
    "        prior_state = posterior_mean + process_mean\n",
    "        prior_error = posterior_error + process_error\n",
    "        return self._gaussian(prior_state, prior_error)\n",
    "\n",
    "    def _update(self, prior, measurement):\n",
    "        \"\"\"Calculate posterior from prior and measurement.\"\"\"\n",
    "        prior_mean, prior_error = prior\n",
    "        measurement_mean, measurement_error = measurement\n",
    "        residual = measurement_mean - prior_mean\n",
    "        kalman_gain = prior_error / (prior_error + measurement_error)\n",
    "        posterior_mean = prior_mean + kalman_gain*residual\n",
    "        posterior_error = (1 - kalman_gain) * prior_error\n",
    "        return self._gaussian(posterior_mean, posterior_error)\n",
    "\n",
    "    @property\n",
    "    def state(self):\n",
    "        \"\"\"Return the state of the entity.\"\"\"\n",
    "        return self._state.mean\n",
    "    \n",
    "    @property\n",
    "    def variance(self):\n",
    "        \"\"\"Return the state of the entity.\"\"\"\n",
    "        return self._state.var"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "sensitivity = 0.8 # range 0 - 1\n",
    "\n",
    "kalman_1 = kalman(sensitivity)\n",
    "\n",
    "states = []\n",
    "variance = []\n",
    "for z in readings:\n",
    "    kalman_1.process_measurement(z)\n",
    "    # save for latter plotting\n",
    "    states.append(kalman_1.state)\n",
    "    variance.append(kalman_1.variance)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x115c43470>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(17,7))\n",
    "plt.plot(time, readings, 'bo', label='readings') # Dummy reading\n",
    "plt.plot(time, states, 'ro--', label='states')\n",
    "plt.plot(time, variance, 'go--', label='variance')\n",
    "plt.legend()\n",
    "#plt.ylim(16, 17)\n",
    "plt.xlabel('step')\n",
    "plt.ylabel('volts')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.4"
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}