summaryrefslogtreecommitdiff
path: root/wasp/ppg.py
blob: 324392d4c184feae1673ffc793e20e12bd7e98c9 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
# SPDX-License-Identifier: LGPL-3.0-or-later
# Copyright (C) 2020 Daniel Thompson

"""Photoplethysmogram (PPG) Signal Processing
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Algorithms and signal processing primatives that can be used to convert
raw PPG signals into something useful.
"""

import array
import micropython

@micropython.viper
def _compare(d1, d2, count: int, shift: int) -> int:
    """Compare two sequences of (signed) bytes and quantify how dissimilar
    they are.
    """
    p1 = ptr8(d1)
    p2 = ptr8(d2)

    e = 0
    for i in range(count):
        s1 = int(p1[i])
        if s1 > 127:
            s1 -= 256

        s2 = int(p2[i])
        if s2 > 127:
            s2 -= 256

        d = s1 - s2
        e += d*d
    return e

class Biquad():
    """Direct Form II Biquad Filter"""

    def __init__(self, b0, b1, b2, a1, a2):
        self._coeff = (b0, b1, b2, a1, a2)
        self._v1 = 0
        self._v2 = 0

    def step(self, x):
        c = self._coeff
        v1 = self._v1
        v2 = self._v2

        v = x - (c[3] * v1) - (c[4] * v2)
        y = (c[0] * v) + (c[1] * v1) + (c[2] * v2)

        self._v2 = v1
        self._v1 = v
        return y

class PTAGC():
    """Peak Tracking Automatic Gain Control

    In order for the correlation checks to work correctly we must
    aggressively reject spikes caused by fast DC steps. Setting a
    threshold based on the median is very effective at killing
    spikes but needs an extra 1k for sample storage which isn't
    really plausible for a microcontroller.
    """
    def __init__(self, start, decay, threshold):
        self._peak = start
        self._decay = decay
        self._boost = 1 / decay
        self._threshold = threshold

    def step(self, spl):
        # peak tracking
        peak = self._peak
        if abs(spl) > peak:
            peak *= self._boost
        else:
            peak *= self._decay
        self._peak = peak

        # rejection filter (clipper)
        threshold = self._threshold
        if spl > (peak * threshold) or spl < (peak * -threshold):
            return 0

        # booster
        spl = 100 * spl / (2 * peak)

        return spl

class PPG():
    """
    """

    def __init__(self, spl):
        self._offset = spl
        self.data = array.array('b')

        self._hpf = Biquad(0.87033078, -1.74066156, 0.87033078,
                                       -1.72377617, 0.75754694)
        self._agc = PTAGC(20, 0.971, 2)
        self._lpf = Biquad(0.11595249, 0.23190498, 0.11595249,
                                      -0.72168143, 0.18549138)

    def preprocess(self, spl):
        """Preprocess a PPG sample.

        Must be called at 24Hz for accurate heart rate calculations.
        """
        spl -= self._offset
        spl = self._hpf.step(spl)
        spl = self._agc.step(spl)
        spl = self._lpf.step(spl)
        spl = int(spl)

        self.data.append(spl)
        return spl

    def _get_heart_rate(self):
        def compare(d, shift):
            return _compare(d[shift:], d[:-shift], len(d)-shift, shift)

        def trough(d, mn, mx):
            z2 = compare(d, mn-2)
            z1 = compare(d, mn-1)
            for i in range(mn, mx+1):
                z = compare(d, i)
                if z2 > z1 and z1 < z:
                    return i
                z2 = z1
                z1 = z

            return -1

        data = memoryview(self.data)

        # Search initially from ~210 to 30 bpm
        t0 = trough(data, 7, 48)
        if t0 < 0:
            return None

        # Check the second cycle ...
        t1 = t0 * 2
        t1 = trough(data, t1 - 5, t1 + 5)
        if t1 < 0:
            return None

        # ... and the third
        t2 = (t1 * 3) // 2
        t2 = trough(data, t2 - 5, t2 + 4)
        if t2 < 0:
            return None

        # If we can find a fourth cycle then use that for the extra
        # precision otherwise report whatever we've found
        t3 = (t2 * 4) // 3
        t3 = trough(data, t3 - 4, t3 + 4)
        if t3 < 0:
            return (60 * 24 * 3) // t2
        return (60 * 24 * 4) // t3

    def get_heart_rate(self):
        if len(self.data) < 200:
            return None

        hr = self._get_heart_rate()

        # Clear out the accumulated data
        self.data = array.array('b')

        return hr