BLELongCharacteristic MP34DT05_PDM2(MP34DT05_PDM_Characteristic2_UUID, BLERead | BLEWrite);

Sensor_MP34DT05_Service.addCharacteristic(MP34DT05_PDM2);

void readSamplesNum(){
  if (MP34DT05_PDM2.written()) {
    if (MP34DT05_PDM2.value()) {   // any value other than 0
      samples_num = 1024;
    } else {                              // a 0 value
      samples_num = 256;
    }
  }
}

 await client.start_notify(char_uuid, notification_handler)

client.write_gatt_char(char2_uuid, sampleNo)

import paho.mqtt.client as mqtt
mqtt_client = mqtt.Client()
mqtt_client.connect(broker_address, port)

mqtt_client.on_connect = on_connect
mqtt_client.on_message = on_message
mqtt_client.loop_start()

def on_connect(client, userdata, flags, rc):
    print("Connected with result code "+str(rc))
    client.subscribe(topic2)

def on_message(client, userdata, msg):
    global recv_sample
    recv_sample = json.loads(msg.payload)

dict_data = json.loads(msg.payload)

N = dict_data["count"] + 2

window = np.hanning(N)

data = dict_data["data"]
y = np.array(data[:N])

y = y * window

def setlayer(self):
        self.layers['Affine1'] = Affine(self.params['W1'], self.params['b1'])
        self.layers['Relu1'] = Relu()
        self.layers['Affine2'] = Affine(self.params['W2'], self.params['b2’])

# Deep Learning Calc
network256  = TwoLayerNet(input_size=127, hidden_size=100, output_size=5)
network1024 = TwoLayerNet(input_size=511, hidden_size=100, output_size=5)

# Deep Learning用paramter読込み
with open('params_1024.pkl', 'rb') as f1:
    network1024.params = pickle.load(f1)
network1024.setlayer()

with open('params_256.pkl', 'rb') as f2:
    network256.params = pickle.load(f2)
network256.setlayer()

# Deep Learningによる判定
    if N == 256:
        y = network256.predict(np.tile(Amp[1:int(N/2)], (2,1)))
    else:
        y = network1024.predict(np.tile(Amp[1:int(N/2)], (2,1)))
    y = np.argmax(y, axis=1)
    tp = int(y[0])

 # Sampling point数の変更
    if (tp == 2 or tp == 4):
        next = 1024
    else:
        next = 256

    if next != N:
        publish(str(next))

/*
 * prei-ble-pdm03: 間欠動作対応 サンプル数256とサンプル数1024の切換
 * Arduino LSM9DS1 : send sensordata on BLE
 * Arduino MP34DT05 : Digital Microphone
 */
 
#include 
#include 

#define localNAME "peri_pdm"
#define DeviceNAME "PDM"

// UUID for MP34DT05(Digital MicroPhone)
#define MP34DT05_SERVICE_UUID "9997c274-9b20-4181-808f-ef2de31032a8"
#define MP34DT05_PDM_Characteristic_UUID "9997c274-9b21-4181-808f-ef2de31032a8"
#define MP34DT05_PDM_Characteristic2_UUID "9997c274-9b22-4181-808f-ef2de31032a8"

// send data size
#define MAX_SAMPLES 1024
int max_samples = 256;
int samples_num = 256; 
short sendData[MAX_SAMPLES];

//Variables for MP34DT05
short PDM_MP34DT05 = 0;
short sampleBuffer[256];
short sampleData;
short sampleDataBuffer[256];
unsigned long theTime;
volatile int samplesRead;
int cnt = 0;

// BLE Service
BLEService Sensor_MP34DT05_Service(MP34DT05_SERVICE_UUID);

// BLE Characteristic
BLELongCharacteristic MP34DT05_PDM(MP34DT05_PDM_Characteristic_UUID, BLERead | BLENotify);
BLELongCharacteristic MP34DT05_PDM2(MP34DT05_PDM_Characteristic2_UUID, BLERead | BLEWrite);

void setup() {
  Serial.begin(115200);
  //while (!Serial);
  Serial.println("Started");

  // initialize the built-in LED pin to indicate when a central is connected
  pinMode(LED_BUILTIN, OUTPUT); 
  
  // MP34DT05 begin initialization
  PDM.onReceive(onPDMdata);

  // optionally set the gain, defaults to 20
  // PDM.setGain(30);

  // initialize PDM with: one channel (mono mode) & 16 kHz sample rate
  if (!PDM.begin(1, 16000)) {
    Serial.println("Failed to start PDM!");
    while (1);
  }

  // BLE begin initialization
  if (!BLE.begin()) {
    Serial.println("starting BLE failed!");
    while (1);
  }
  BLE.setLocalName(localNAME);
  BLE.setDeviceName(DeviceNAME);

  // Initialize MP34DT05 service
  set_MP34DT05();

  // start advertising
  BLE.advertise();
  Serial.println("Bluetooth device active, waiting for connections...");
}

void loop() {
  // wait for a BLE central
  BLEDevice central = BLE.central();
 
  // if a central is connected to the peripheral:
  if (central) {
    Serial.print("Connected to central: ");
    // print the central's BT address:
    Serial.println(central.address());
    // turn on the LED to indicate the connection:
    digitalWrite(LED_BUILTIN, HIGH);
 
    // while the central is connected:
    while (central.connected()) {
      readSamplesNum();
      if (cnt == 0){
        max_samples = samples_num;
        Read_PDM_Data();   // reads 1024 data from MP34DT05
        //IncDataGen();      // Incremental data generation for sending test
      }
      RGB_LED();
      notify_PDM();       // SEND_BLE
      counter();
      delay(50);
    }

    // when the central disconnects
    digitalWrite(LED_BUILTIN, LOW);
    Serial.print("Disconnected from central: ");
    Serial.println(central.address());
  } else {
      readMP34DT05_nonBLE();    
  }
}

#include 
#include 

void onPDMdata() {
  // query the number of bytes available
  int bytesAvailable = PDM.available();

  // read into the sample buffer
  PDM.read(sampleBuffer, bytesAvailable);

  // 16-bit, 2 bytes per sample
  samplesRead = bytesAvailable / 2;
}

void set_MP34DT05(){
  // add the service UUID
  BLE.setAdvertisedService(Sensor_MP34DT05_Service);

  // add characteristic
  Sensor_MP34DT05_Service.addCharacteristic(MP34DT05_PDM);
  Sensor_MP34DT05_Service.addCharacteristic(MP34DT05_PDM2);

  // Add service
  BLE.addService(Sensor_MP34DT05_Service);
}

// turn on RGB LED according max_samples
void RGB_LED(){
  if (max_samples == 1024){
    digitalWrite(LEDR,LOW);
    digitalWrite(LEDG,HIGH);
    digitalWrite(LEDB,HIGH);
  } else if (max_samples == 256){
    digitalWrite(LEDR,HIGH);
    digitalWrite(LEDG,HIGH);
    digitalWrite(LEDB,LOW);
  }      
}

// read pdm data and set to the dimensions
void Read_PDM_Data(){
  // Wait for samples to be read
  for (int j = 0; j<(max_samples/256); j++){
    while (samplesRead!=256);
    for (int i = 0; i < samplesRead; i++) {
      sendData[j*256+i] = sampleBuffer[i];
    }
    // Clear the read count
    samplesRead = 0;
  }
}

// Increment data generator for sending test
void IncDataGen(){
  for (int i=0; i<max_samples; i++){
    sendData[i] = i;
  }
}

// counter for send data control
void counter(){
  cnt++;
  if (cnt >= (max_samples/2+1)) cnt = 0;
}

// read 
void notify_PDM(){
  int i = 2*(cnt - 1);
  long sending;
  short sendingL, sendingH;
  if (cnt == 0){
    if (max_samples == 256){
      sending=0x7FFFFFFF;  // 32767      
    } else {
      sending=0x7FF0FFFF;  // 32752
    }
  } else {
    sending=((long)sendData[i+1] << 16) | (sendData[i] & 0xFFFF);
  }
  MP34DT05_PDM.writeValue(sending);
  sendingH = (short)((sending >> 16) & 0xFFFF);
  sendingL = (short)(sending & 0xFFFF);
  Serial.print(cnt);
  Serial.print(": "); 
  Serial.print(sendingH);
  Serial.print(", "); 
  Serial.println(sendingL);
}

// read samples number indicated by PDM
void readSamplesNum(){
  if (MP34DT05_PDM2.written()) {
    if (MP34DT05_PDM2.value()) {   // any value other than 0
      samples_num = 1024;
    } else {                              // a 0 value
      samples_num = 256;
    }
  }
}

void readMP34DT05_nonBLE() {
  // print samples to the serial monitor or plotter
  for (int i = 0; i < samplesRead; i++) {
    sampleData = sampleBuffer[i];
    LED_Control();
  }
}

void LED_Control(){
  if (abs(sampleData)>=0 && abs(sampleData) < 50){
    digitalWrite(LEDR,HIGH);
    digitalWrite(LEDG,HIGH);
    digitalWrite(LEDB,HIGH);
  } else if (abs(sampleData) < 100){
    digitalWrite(LEDR,HIGH);
    digitalWrite(LEDG,HIGH);
    digitalWrite(LEDB,LOW);        
  } else if (abs(sampleData) < 150){
    digitalWrite(LEDR,HIGH);
    digitalWrite(LEDG,LOW);
    digitalWrite(LEDB,HIGH);        
  } else if (abs(sampleData) < 200){
    digitalWrite(LEDR,HIGH);
    digitalWrite(LEDG,LOW);
    digitalWrite(LEDB,LOW);        
  } else if (abs(sampleData) < 250){
    digitalWrite(LEDR,LOW);
    digitalWrite(LEDG,HIGH);
    digitalWrite(LEDB,HIGH);        
  } else if (abs(sampleData) < 300){
    digitalWrite(LEDR,LOW);
    digitalWrite(LEDG,HIGH);
    digitalWrite(LEDB,LOW);        
  } else if (abs(sampleData) < 350){
    digitalWrite(LEDR,LOW);
    digitalWrite(LEDG,LOW);
    digitalWrite(LEDB,HIGH);
  } else {
    digitalWrite(LEDR,LOW);
    digitalWrite(LEDG,LOW);
    digitalWrite(LEDB,LOW);
  }
}

# -*- coding: utf-8 -*-
#####
# Successively receive micro-phone data of int type
# Rev.0.03: 2023/11/16 無限loopにする。
# Rev.0.04: 2023/11/20 short X 2 = Longの転送に対応
# Rev.0.05: 2023/11/25 send sampling number
# Rev.0.06: 2023/12/20 FFT結果の受信とSampling数の変換指示追加
####
import sys
import signal
import asyncio
from bleak import BleakClient

import paho.mqtt.client as mqtt
import time
import json
import struct

# setting for BLE
ADDRESS = (
#    "93:B3:5F:16:CD:A7" # Arduino Nano 33 BLE sense Flat
    "33:07:90:D9:BB:5D" # Arduino Nano 33 BLE sense MinoMushi
)

# UUID for MP34DT05
CHARACTERISTIC_UUID = "9997c274-9b21-4181-808f-ef2de31032a8" # read & notify
CHARACTERISTIC2_UUID = "9997c274-9b22-4181-808f-ef2de31032a8" # read & write

# MQTT ブローカーの設定
broker_address = "localhost"
port = 1883
topic = "pdm_pic"
topic2 = "fft_condition"

# MQTT クライアントを作成
mqtt_client = mqtt.Client()

data = [0.0,0.0,0.0]
sdata = {    # 辞書型データ
    'sensor': 'MP34DT05',
    'func': 'none',
    'time': 'time',
    'data': data
}

def publish(sdata):
    # Publish on MQTT
    json_data = json.dumps(sdata)
    mqtt_client.publish(topic, json_data)

def bytes_to_long(data):
    Data = [0]*2
    for i in range(0, len(data), 4):
        # 4バイトずつ読み込み、それをlong型に変換
        long_data = struct.unpack('I', data[i:i+4])[0]
    # 上位と下位の16bitに分ける(符号付き)
    Data[1] = struct.unpack('h', struct.pack('H', (long_data >> 16) & 0xFFFF))[0]
    Data[0] = struct.unpack('h', struct.pack('H', long_data & 0xFFFF))[0]
    return Data

def bytes_to_int(data):
    return int.from_bytes(data, byteorder='little')

def bytes_to_double(data):
    return struct.unpack('d', data)[0]

def bytes_to_signed_int(data):
    return int.from_bytes(data, byteorder='little', signed=True)

def notification_handler(sender, data: bytearray):
    """Simple notification handler which sends MP34DT05 PDM data"""
    data1 = bytes_to_long(data)
    sdata['sensor'] = 'MP34DT05'
    sdata['func'] = 'PDM'
    sdata['time'] = time.time()
    sdata['data'] = data1
    #print(data1)
    publish(sdata)

sampleNo = bytearray([0])
sampleNoOld = bytearray([0])
recv_sample = 256

async def setSampleNo():
    global sampleNo
    if (recv_sample == 256):
        sampleNo = bytearray([0])
    elif (recv_sample == 1024):
        sampleNo = bytearray([1])
    else:
        sampleNo = bytearray([0])
    #print(f'{cycleNo}, {sampleNo}, {recv_sample}')

def on_connect(client, userdata, flags, rc):
    print("Connected with result code "+str(rc))
    client.subscribe(topic2)

def on_message(client, userdata, msg):
    global recv_sample
    recv_sample = json.loads(msg.payload)
    print(f'{msg.payload},{recv_sample}')
    
async def main(address):
    global sampleNoOld
    char_uuid = CHARACTERISTIC_UUID
    char2_uuid = CHARACTERISTIC2_UUID

    # MQTT ブローカーに接続
    mqtt_client.connect(broker_address, port)
    print(f"MQTT connected: {broker_address}")
    mqtt_client.on_connect = on_connect
    mqtt_client.on_message = on_message
    mqtt_client.loop_start()
    
    print(address, char_uuid)
    async with BleakClient(address) as client:
        print(f"Connected: {client.is_connected}")
        await client.start_notify(char_uuid, notification_handler)
        try:
            while True:
                await setSampleNo()
                if sampleNo != sampleNoOld:
                    await client.write_gatt_char(char2_uuid, sampleNo)
                    print(f'sampleNo={sampleNo}')
                    sampleNoOld = sampleNo
                await asyncio.sleep(1.0)
        except KeyboardInterrupt:
            print("Stop!! Key board Interrupt!!");
            #await client.stop_notify(char_uuid)
        finally:
            await client.stop_notify(char_uuid)

    print("Disconnected")
    # MQTT クライアントを切断
    mqtt_client.disconnect()
    print(f"MQTT disconnected: {broker_address}")
        
if __name__ == "__main__":
    asyncio.run(
        main(
            sys.argv[1] if len(sys.argv) > 1 else ADDRESS,
        )
    )

[{"id":"287721ebd0af0b63","type":"tab","label":"PDM_PIC2","disabled":false,"info":"","env":[]},{"id":"00823d1a50950b0a","type":"debug","z":"287721ebd0af0b63","name":"debug 81","active":true,"tosidebar":true,"console":false,"tostatus":false,"complete":"false","statusVal":"","statusType":"auto","x":420,"y":80,"wires":[]},{"id":"abe28cfc0dda4281","type":"exec","z":"287721ebd0af0b63","command":"python3 /home/pi/source/python/ble/ble_pdh06.py","addpay":"payload","append":"","useSpawn":"false","timer":"","winHide":false,"oldrc":false,"name":"ble_pdh06","x":270,"y":80,"wires":[[],["00823d1a50950b0a"],["00823d1a50950b0a"]]},{"id":"ce137ed9b7a8867b","type":"inject","z":"287721ebd0af0b63","name":"受信開始","props":[{"p":"filename","v":"/home/pi/Documents/pdm_imu/receive_","vt":"str"},{"p":"payload"}],"repeat":"","crontab":"","once":false,"onceDelay":0.1,"topic":"","payload":"33:07:90:D9:BB:5D","payloadType":"str","x":100,"y":100,"wires":[["885ca16d7a763aed","abe28cfc0dda4281","382c3786e2e8efe7","52491d06734e366b"]]},{"id":"0cd751654fc88190","type":"inject","z":"287721ebd0af0b63","name":"受信終了","props":[{"p":"kill","v":"","vt":"str"},{"p":"topic","vt":"str"}],"repeat":"","crontab":"","once":false,"onceDelay":0.1,"topic":"","x":100,"y":140,"wires":[["abe28cfc0dda4281"]]},{"id":"885ca16d7a763aed","type":"moment","z":"287721ebd0af0b63","name":"日時","topic":"","input":"","inputType":"date","inTz":"Asia/Tokyo","adjAmount":0,"adjType":"days","adjDir":"add","format":"YYYYMMDD_HHmmss","locale":"ja-JP","output":"datetime","outputType":"msg","outTz":"Asia/Tokyo","x":250,"y":140,"wires":[["41ec1a3b54d21e73"]]},{"id":"382c3786e2e8efe7","type":"function","z":"287721ebd0af0b63","name":"初期化","func":"flow.set('count', 0);\nflow.set('cycle', 0);\nflow.set('count_max', 256);","outputs":1,"noerr":0,"initialize":"","finalize":"","libs":[],"x":250,"y":180,"wires":[[]]},{"id":"52491d06734e366b","type":"function","z":"287721ebd0af0b63","name":"配列作成","func":"// 配列の要素数\nflow.set(\"length\", 1024);\n// 配列作成\n//if (!context.flow.array) {\ncontext.flow.array = new Array(flow.get(\"length\"));\n//}\n// 配列の初期化:データを0で埋める\ncontext.flow.array.fill(0);","outputs":1,"noerr":0,"initialize":"","finalize":"","libs":[],"x":260,"y":220,"wires":[[]]},{"id":"41ec1a3b54d21e73","type":"change","z":"287721ebd0af0b63","name":"filename","rules":[{"t":"set","p":"filename","pt":"flow","to":"filename&datetime&ext","tot":"jsonata"}],"action":"","property":"","from":"","to":"","reg":false,"x":380,"y":140,"wires":[[]]},{"id":"d802477288e1c9d3","type":"inject","z":"287721ebd0af0b63","name":"初期化","props":[{"p":"payload"},{"p":"topic","vt":"str"}],"repeat":"","crontab":"","once":true,"onceDelay":0.1,"topic":"","payload":"","payloadType":"date","x":100,"y":200,"wires":[["382c3786e2e8efe7","52491d06734e366b"]]},{"id":"4550065dbe0175df","type":"inject","z":"287721ebd0af0b63","name":"FFT終了","props":[{"p":"kill","v":"","vt":"str"},{"p":"topic","vt":"str"}],"repeat":"","crontab":"","once":false,"onceDelay":0.1,"topic":"","x":660,"y":120,"wires":[["cc57b01a40dff466"]]},{"id":"cc57b01a40dff466","type":"exec","z":"287721ebd0af0b63","command":"python3 ","addpay":"payload","append":"","useSpawn":"false","timer":"","winHide":false,"oldrc":false,"name":"fft_calc04","x":800,"y":80,"wires":[[],["eea006e5a7f7e122"],["eea006e5a7f7e122"]]},{"id":"1edbfdb4622b3e10","type":"inject","z":"287721ebd0af0b63","name":"FFT開始","props":[{"p":"payload"},{"p":"topic","vt":"str"}],"repeat":"","crontab":"","once":false,"onceDelay":0.1,"topic":"","payload":"/home/pi/source/python/deep_learning/fft/fft_calc04.py","payloadType":"str","x":660,"y":80,"wires":[["cc57b01a40dff466"]]},{"id":"eea006e5a7f7e122","type":"debug","z":"287721ebd0af0b63","name":"debug 82","active":true,"tosidebar":true,"console":false,"tostatus":false,"complete":"payload","targetType":"msg","statusVal":"","statusType":"auto","x":940,"y":80,"wires":[]},{"id":"77cb07192336961c","type":"moment","z":"287721ebd0af0b63","name":"日時","topic":"","input":"","inputType":"date","inTz":"Asia/Tokyo","adjAmount":0,"adjType":"days","adjDir":"add","format":"YYYY-MM-DD_HH_mm_ss","locale":"ja-JP","output":"payload.datetime","outputType":"msg","outTz":"Asia/Tokyo","x":350,"y":360,"wires":[["00b398a331e6550e","27a33a153566adeb"]]},{"id":"4195344a881735ae","type":"change","z":"287721ebd0af0b63","name":"削除","rules":[{"t":"delete","p":"payload.sensor","pt":"msg"},{"t":"delete","p":"payload.func","pt":"msg"},{"t":"delete","p":"payload.time","pt":"msg"}],"action":"","property":"","from":"","to":"","reg":false,"x":230,"y":360,"wires":[["77cb07192336961c"]]},{"id":"00b398a331e6550e","type":"function","z":"287721ebd0af0b63","name":"集約(ble_pdh04)","func":"var cycle = flow.get('cycle');\nvar count = flow.get('count');\nvar count_max = flow.get('count_max');\nif (msg.payload.data[1] >= (2 ** 15 - 16) || count > 1024){\n count = 0;\n context.flow.array.fill(0);\n if (msg.payload.data[1] >= 2**15-1){\n count_max = 256;\n } else if (msg.payload.data[1] >= 2**15-16){\n count_max = 1024;\n }\n node.warn(\"first data detexted(\"+String(msg.payload.data[1]+\")!!\"));\n flow.set('count_max',count_max);\n} else {\n context.flow.array[count] = msg.payload.data[0];\n context.flow.array[count+1] = msg.payload.data[1];\n msg.payload.count = count;\n msg.payload.cycle = cycle;\n msg.payload.full = false;\n count = count+2;\n}\nflow.set('count', count);\nif (count == count_max){\n msg.payload.data = context.flow.array;\n msg.payload.full = true;\n cycle++;\n flow.set('cycle', cycle);\n return msg;\n}\n","outputs":1,"noerr":0,"initialize":"","finalize":"","libs":[],"x":510,"y":360,"wires":[["0fc3768d2ca70b8c","6c7733f802402d1e"]]},{"id":"27a33a153566adeb","type":"debug","z":"287721ebd0af0b63","name":"debug 83","active":false,"tosidebar":true,"console":false,"tostatus":false,"complete":"false","statusVal":"","statusType":"auto","x":480,"y":400,"wires":[]},{"id":"947bc2a4a89f7186","type":"mqtt in","z":"287721ebd0af0b63","name":"","topic":"pdm_pic","qos":"2","datatype":"auto-detect","broker":"496bc63a53e5165b","nl":false,"rap":true,"rh":0,"inputs":0,"x":100,"y":360,"wires":[["4195344a881735ae"]]},{"id":"0fc3768d2ca70b8c","type":"change","z":"287721ebd0af0b63","name":"filename","rules":[{"t":"set","p":"filename","pt":"msg","to":"filename","tot":"flow"},{"t":"set","p":"filename","pt":"msg","to":"filename&\".csv\"","tot":"jsonata"}],"action":"","property":"","from":"","to":"","reg":false,"x":680,"y":360,"wires":[["46f796c3c04083b5"]]},{"id":"6c7733f802402d1e","type":"mqtt out","z":"287721ebd0af0b63","name":"","topic":"fft_data","qos":"","retain":"","respTopic":"","contentType":"","userProps":"","correl":"","expiry":"","broker":"496bc63a53e5165b","x":680,"y":400,"wires":[]},{"id":"46f796c3c04083b5","type":"csv","z":"287721ebd0af0b63","name":"","sep":",","hdrin":"","hdrout":"none","multi":"one","ret":"\\n","temp":"cycle,count,datetime,data","skip":"0","strings":true,"include_empty_strings":"","include_null_values":"","x":810,"y":360,"wires":[["50e9a5dbf4aab002"]]},{"id":"50e9a5dbf4aab002","type":"file","z":"287721ebd0af0b63","name":"","filename":"filename","filenameType":"msg","appendNewline":false,"createDir":true,"overwriteFile":"false","encoding":"none","x":940,"y":360,"wires":[[]]},{"id":"cb78e1cd8f1677a3","type":"comment","z":"287721ebd0af0b63","name":"送信データ集約","info":"","x":120,"y":320,"wires":[]},{"id":"3905c9ed879c6422","type":"mqtt in","z":"287721ebd0af0b63","name":"","topic":"fft_data","qos":"2","datatype":"auto-detect","broker":"496bc63a53e5165b","nl":false,"rap":true,"rh":0,"inputs":0,"x":90,"y":440,"wires":[["9cf1ca24ac162b8d"]]},{"id":"9cf1ca24ac162b8d","type":"debug","z":"287721ebd0af0b63","name":"debug 84","active":true,"tosidebar":true,"console":false,"tostatus":false,"complete":"false","statusVal":"","statusType":"auto","x":220,"y":440,"wires":[]},{"id":"3cc3a9ad104246b2","type":"watch","z":"287721ebd0af0b63","name":"png監視","files":"/home/pi/Documents/pdm_imu/pictures","recursive":"","x":100,"y":540,"wires":[["43975452a45c4645"]]},{"id":"43975452a45c4645","type":"function","z":"287721ebd0af0b63","name":"png file","func":"if (context.get(\"fileName\") === undefined) {\n context.set(\"fileName\", \"newfile.png\")\n}\nif (context.fileName == msg.payload){\n context.fileName = msg.payload;\n return msg;\n} else {\n context.fileName = msg.payload;\n}\n","outputs":1,"noerr":0,"initialize":"","finalize":"","libs":[],"x":240,"y":540,"wires":[["3e896ea2754d9266","85fce9f8600c3316"]]},{"id":"3e896ea2754d9266","type":"image viewer","z":"287721ebd0af0b63","name":"","width":160,"data":"payload","dataType":"msg","active":true,"x":370,"y":540,"wires":[[]]},{"id":"85fce9f8600c3316","type":"debug","z":"287721ebd0af0b63","name":"debug 85","active":false,"tosidebar":true,"console":false,"tostatus":false,"complete":"false","statusVal":"","statusType":"auto","x":380,"y":500,"wires":[]},{"id":"496bc63a53e5165b","type":"mqtt-broker","name":"","broker":"localhost","port":"1883","clientid":"","autoConnect":true,"usetls":false,"protocolVersion":"4","keepalive":"60","cleansession":true,"birthTopic":"","birthQos":"0","birthPayload":"","birthMsg":{},"closeTopic":"","closeQos":"0","closePayload":"","closeMsg":{},"willTopic":"","willQos":"0","willPayload":"","willMsg":{},"userProps":"","sessionExpiry":""}]

'''
fft_calc04.py : execute fft
  rev0.3: change sampling number according msg.payload.count
  rev0.4: fftの結果をDeep Learningで判断できるように修正
'''
import paho.mqtt.client as mqtt
import numpy as np
import matplotlib.pyplot as plt
import json
import time

# Deep Learning用
import numpy as np
import pickle
from two_layer_net import TwoLayerNet

# MQTT ブローカーの設定
broker_address = "localhost"
port = 1883
topic = "fft_data"
topic2 = "fft_condition"

# FFT Configulation
fs = 16000       # sampling frequency
N  = 1024       # number of samples
dt = 1/fs       # sampling period

# Window function
func_window = 'hanning'

# MQTT クライアントを作成
mqtt_client = mqtt.Client()

# Deep Learning Calc
network256  = TwoLayerNet(input_size=127, hidden_size=100, output_size=5)
network1024 = TwoLayerNet(input_size=511, hidden_size=100, output_size=5)

sdata = {    # 辞書型データ
    'current_SMP': 1024, 
    'next_SMP' : 1024
}

def publish(sdata):
    # Publish on MQTT
    #json_data = json.dumps(sdata)
    mqtt_client.publish(topic2, sdata)

def on_connect(client, userdata, flags, rc):
    print("Connected with result code "+str(rc))
    client.subscribe(topic)

def on_message(client, userdata, msg):
    dict_data = json.loads(msg.payload)
    N = dict_data["count"] + 2
    window = np.hanning(N)
    data = dict_data["data"]
    y = np.array(data[:N])
    #y = np.array(dict_data["data"])
    y = y * window
    if dict_data["full"] == 1:
        print(y)
    else:
        print(dict_data["full"])

    # FFTを実行
    y_fft = np.fft.fft(y)               # 離散フーリエ変換
    freq = np.fft.fftfreq(N, d=dt)      # 周波数を割り当てる
    Amp = abs(y_fft/(N/2))              # 音の大きさ（振幅の大きさ）

    # 窓補正
    acf=1/(sum(window)/N)
    Amp = acf*Amp
    
    output_FN = "/home/pi/Documents/pdm_imu/pictures/"+dict_data["datetime"]+".png"

    ### 音波のスペクトル ###
    plt.plot(freq[1:int(N/2)], Amp[1:int(N/2)]) # A-f グラフのプロット
    if N == 256:
        plt.xlabel("frequency [Hz](sp=256)")
    else:
        plt.xlabel("frequency [Hz](sp=1024)")
    plt.ylabel("amplitude [V/rtHz]")
    plt.xscale("log")                           # 横軸を対数軸にセット
    plt.yscale("log")                           # 縦軸を対数軸にセット
    plt.savefig(output_FN, dpi=300)
    plt.clf()

    # Deep Learningによる判定
    if N == 256:
        y = network256.predict(np.tile(Amp[1:int(N/2)], (2,1)))
    else:
        y = network1024.predict(np.tile(Amp[1:int(N/2)], (2,1)))
    y = np.argmax(y, axis=1)
    tp = int(y[0])

    # Sampling point数の変更
    if (tp == 2 or tp == 4):
        next = 1024
    else:
        next = 256

    if next != N:
        sdata['current_SMP'] = N
        sdata['next_SMP'] = next
        publish(str(next))

    print(f"fft type: {tp}, current_SMP: {N}, next_SMP: {next}")

    # NumPy配列をPythonリストに変換
    freq_list = freq[1:int(N/2)].tolist()
    amp_list = Amp[1:int(N/2)].tolist()

    # PythonリストをJSON形式に変換
    out_data = {"Num":N, "type":tp, "datatime":dict_data["datetime"], "freq": freq_list, "amp": amp_list}
    data_json = json.dumps(out_data)
    #print(data_json)

    outFile = "/home/pi/Documents/pdm_imu/fftResult/"+dict_data["datetime"]+".json"
    # JSONデータをファイルに出力
    with open(outFile, "w") as f:
        f.write(data_json)
    
def main():
    # Deep Learning用paramter読込み
    with open('params_1024.pkl', 'rb') as f1:
        network1024.params = pickle.load(f1)
    network1024.setlayer()

    with open('params_256.pkl', 'rb') as f2:
        network256.params = pickle.load(f2)
    network256.setlayer()

    # MQTT ブローカーに接続
    mqtt_client.connect(broker_address, port)
    print(f"MQTT connected: {broker_address}")
    mqtt_client.on_connect = on_connect
    mqtt_client.on_message = on_message

    try:
        mqtt_client.loop_forever()
    except KeyboardInterrupt:
        # MQTT クライアントを切断
        mqtt_client.disconnect()
        print(f"MQTT disconnected: {broker_address}")

if __name__ == "__main__":
    main()

# coding: utf-8
import sys, os
sys.path.append(os.pardir)  # 親ディレクトリのファイルをインポートするための設定
import numpy as np
from common.layers import *
from common.gradient import numerical_gradient
from collections import OrderedDict


class TwoLayerNet:

    def __init__(self, input_size, hidden_size, output_size, weight_init_std = 0.01):
        # 重みの初期化
        self.params = {}
        self.params['W1'] = weight_init_std * np.random.randn(input_size, hidden_size)
        self.params['b1'] = np.zeros(hidden_size)
        self.params['W2'] = weight_init_std * np.random.randn(hidden_size, output_size) 
        self.params['b2'] = np.zeros(output_size)

        # レイヤの生成
        self.layers = OrderedDict()
        self.layers['Affine1'] = Affine(self.params['W1'], self.params['b1'])
        self.layers['Relu1'] = Relu()
        self.layers['Affine2'] = Affine(self.params['W2'], self.params['b2'])

        self.lastLayer = SoftmaxWithLoss()

    def setlayer(self):
        self.layers['Affine1'] = Affine(self.params['W1'], self.params['b1'])
        self.layers['Relu1'] = Relu()
        self.layers['Affine2'] = Affine(self.params['W2'], self.params['b2'])
        
    def predict(self, x):
        for layer in self.layers.values():
            x = layer.forward(x)
        return x
        
    # x:入力データ, t:教師データ
    def loss(self, x, t):
        y = self.predict(x)
        return self.lastLayer.forward(y, t)
    
    def accuracy(self, x, t):
        y = self.predict(x)
        y = np.argmax(y, axis=1)
        if t.ndim != 1 : t = np.argmax(t, axis=1)
        
        accuracy = np.sum(y == t) / float(x.shape[0])
        return accuracy
        
    # x:入力データ, t:教師データ
    def numerical_gradient(self, x, t):
        loss_W = lambda W: self.loss(x, t)
        
        grads = {}
        grads['W1'] = numerical_gradient(loss_W, self.params['W1'])
        grads['b1'] = numerical_gradient(loss_W, self.params['b1'])
        grads['W2'] = numerical_gradient(loss_W, self.params['W2'])
        grads['b2'] = numerical_gradient(loss_W, self.params['b2'])
        
        return grads
        
    def gradient(self, x, t):
        # forward
        self.loss(x, t)

        # backward
        dout = 1
        dout = self.lastLayer.backward(dout)
        
        layers = list(self.layers.values())
        layers.reverse()
        for layer in layers:
            dout = layer.backward(dout)

        # 設定
        grads = {}
        grads['W1'], grads['b1'] = self.layers['Affine1'].dW, self.layers['Affine1'].db
        grads['W2'], grads['b2'] = self.layers['Affine2'].dW, self.layers['Affine2'].db

        return grads