智慧生活科技專業社群: [OTTO GO] 離線版語音控制，運用長短音

arduino:
#include <Arduino.h>
#include <WiFi.h>
#include <WebServer.h>
#include <SPI.h>
#include <Adafruit_GFX.h>
#include <Adafruit_ST7789.h>
#include <ESP32Servo.h>
#include "driver/i2s.h"
#include "freertos/semphr.h"

// ===================== Pin map (OTTO GO) =====================
// TFT (SPI)
#define TFT_MOSI 19
#define TFT_SCLK 18
#define TFT_CS    5
#define TFT_DC   16
#define TFT_RST  23

// Ultrasonic
#define US_TRIG 27
#define US_ECHO 39

// Servos
#define SERVO1 21
#define SERVO2 22
#define SERVO3 4
#define SERVO4 14

// Speaker (I2S -> AMP -> SPK white socket)
#define SPK_I2S_LRCK 17
#define SPK_I2S_BCLK 2
#define SPK_I2S_DOUT 32

// Microphone (I2S MEMS)
#define MIC_I2S_SCK 13
#define MIC_I2S_SD  12
#define MIC_I2S_WS  15

// ===================== WiFi AP =====================
const char* AP_SSID = "OTTO_GO_TEST";
const char* AP_PASS = "12345678";

// ===================== Objects =====================
WebServer server(80);
Adafruit_ST7789 tft = Adafruit_ST7789(TFT_CS, TFT_DC, TFT_RST);
Servo s1, s2, s3, s4;

// ===================== Helpers =====================
static inline uint16_t RGB565(uint8_t r, uint8_t g, uint8_t b) {
  return ((r & 0xF8) << 8) | ((g & 0xFC) << 3) | (b >> 3);
}

// ===================== Ultrasonic =====================
long readDistanceCM() {
  digitalWrite(US_TRIG, LOW);
  delayMicroseconds(2);
  digitalWrite(US_TRIG, HIGH);
  delayMicroseconds(10);
  digitalWrite(US_TRIG, LOW);

  unsigned long duration = pulseIn(US_ECHO, HIGH, 30000UL);
  if (duration == 0) return -1;
  return (long)(duration / 58UL);
}
long median3(long a, long b, long c) {
  if (a > b) { long t=a; a=b; b=t; }
  if (b > c) { long t=b; b=c; c=t; }
  if (a > b) { long t=a; a=b; b=t; }
  return b;
}

// ===================== TFT =====================
void tftHello() {
  tft.fillScreen(ST77XX_BLACK);
  tft.setTextColor(ST77XX_WHITE);
  tft.setTextSize(3);
  tft.setCursor(20, 40);
  tft.print("Hello");
}

void drawTigerMouth(bool open) {
  int cx = tft.width()/2;
  int cy = tft.height()/2;

  uint16_t furLight = RGB565(235, 228, 210);
  uint16_t shadow   = RGB565(160, 150, 135);
  uint16_t lineDark = RGB565(40, 40, 40);
  uint16_t dotDark  = RGB565(70, 70, 70);

  tft.fillScreen(ST77XX_BLACK);

  tft.fillCircle(cx - 55, cy + 10, 55, furLight);
  tft.fillCircle(cx + 55, cy + 10, 55, furLight);
  tft.fillRoundRect(cx - 90, cy - 25, 180, 110, 28, furLight);

  tft.fillRoundRect(cx - 85, cy + 60, 170, 22, 12, shadow);
  tft.fillRoundRect(cx - 40, cy + 55, 80, 10, 5, RGB565(120,110,100));

  int dotY1 = cy + 6, dotY2 = cy + 26;
  for (int k=0;k<3;k++){
    tft.fillCircle(cx - 55 - k*12, dotY1 + k*2, 2, dotDark);
    tft.fillCircle(cx - 55 - k*12, dotY2 + k*2, 2, dotDark);
    tft.fillCircle(cx + 55 + k*12, dotY1 + k*2, 2, dotDark);
    tft.fillCircle(cx + 55 + k*12, dotY2 + k*2, 2, dotDark);
  }

  tft.drawLine(cx, cy - 5, cx, open ? (cy + 10) : (cy + 20), lineDark);
  tft.drawLine(cx-1, cy - 5, cx-1, open ? (cy + 10) : (cy + 20), RGB565(80,80,80));

  int mouthY = open ? (cy + 16) : (cy + 18);
  int N = open ? 36 : 34;
  int div = open ? 85 : 95;
  for (int i=0;i<N;i++){
    int y = mouthY + (i*i)/div;
    tft.drawPixel(cx-i, y, lineDark);
    tft.drawPixel(cx-i, y+1, lineDark);
    tft.drawPixel(cx+i, y, lineDark);
    tft.drawPixel(cx+i, y+1, lineDark);
  }

  tft.fillCircle(cx - (open?38:36), mouthY + (open?14:12), 3, lineDark);
  tft.fillCircle(cx + (open?38:36), mouthY + (open?14:12), 3, lineDark);

  if (!open) {
    tft.fillRoundRect(cx - 10, mouthY + 20, 20, 4, 2, RGB565(90,80,70));
  } else {
    uint16_t mouthIn = RGB565(15,15,15);
    uint16_t tongue  = RGB565(210,120,120);
    int w=90, h=45;
    int x=cx-w/2, y0=cy+28;
    tft.fillRoundRect(x, y0, w, h, 18, mouthIn);
    tft.fillCircle(cx, y0+h, 22, mouthIn);
    tft.fillRoundRect(cx-22, y0+18, 44, 20, 10, tongue);
    tft.fillCircle(cx, y0+36, 14, tongue);
  }

  tft.setTextSize(2);
  tft.setTextColor(ST77XX_YELLOW, ST77XX_BLACK);
  tft.setCursor(10, 10);
  tft.print(open ? "MOUTH: OPEN" : "MOUTH: CLOSE");
}

// ===================== I2S0 mode switching (穩定版) =====================
const i2s_port_t I2S_PORT = I2S_NUM_0;

enum I2SMode { MODE_NONE=0, MODE_SPK=1, MODE_MIC=2 };
static volatile I2SMode curMode = MODE_NONE;

static SemaphoreHandle_t i2sMutex = nullptr;

// 狀態顯示
bool spk_ok = false;
bool mic_ok = false;

// 最近一次麥克風能量（給網頁/調參用）
volatile float lastMicEnergy = 0;

// 離線指令狀態顯示
String lastCmd = "NONE";
bool mouthOpenState = false;

static void lockI2S()   { if (i2sMutex) xSemaphoreTake(i2sMutex, portMAX_DELAY); }
static void unlockI2S() { if (i2sMutex) xSemaphoreGive(i2sMutex); }

static void i2sStopAndUninstallSafe() {
  if (curMode == MODE_NONE) return;
  i2s_stop(I2S_PORT);
  i2s_driver_uninstall(I2S_PORT);
  curMode = MODE_NONE;
}

bool i2sInitSpeaker() {
  i2sStopAndUninstallSafe();

  i2s_config_t cfg = {
    .mode = (i2s_mode_t)(I2S_MODE_MASTER | I2S_MODE_TX),
    .sample_rate = 22050,
    .bits_per_sample = I2S_BITS_PER_SAMPLE_16BIT,
    .channel_format = I2S_CHANNEL_FMT_RIGHT_LEFT,
    .communication_format = I2S_COMM_FORMAT_I2S_MSB,
    .intr_alloc_flags = ESP_INTR_FLAG_LEVEL1,
    .dma_buf_count = 4,
    .dma_buf_len   = 128,
    .use_apll = false,
    .tx_desc_auto_clear = true,
    .fixed_mclk = 0
  };

  i2s_pin_config_t pin_cfg = {
    .bck_io_num = SPK_I2S_BCLK,
    .ws_io_num  = SPK_I2S_LRCK,
    .data_out_num = SPK_I2S_DOUT,
    .data_in_num  = I2S_PIN_NO_CHANGE
  };

  esp_err_t e1 = i2s_driver_install(I2S_PORT, &cfg, 0, NULL);
  if (e1 != ESP_OK) return false;

  esp_err_t e2 = i2s_set_pin(I2S_PORT, &pin_cfg);
  if (e2 != ESP_OK) return false;

  i2s_zero_dma_buffer(I2S_PORT);
  i2s_start(I2S_PORT);
  curMode = MODE_SPK;
  return true;
}

bool i2sInitMic() {
  i2sStopAndUninstallSafe();

  i2s_config_t cfg = {
    .mode = (i2s_mode_t)(I2S_MODE_MASTER | I2S_MODE_RX),
    .sample_rate = 16000,
    .bits_per_sample = I2S_BITS_PER_SAMPLE_32BIT,
    .channel_format = I2S_CHANNEL_FMT_ONLY_LEFT,
    .communication_format = I2S_COMM_FORMAT_I2S_MSB,
    .intr_alloc_flags = ESP_INTR_FLAG_LEVEL1,
    .dma_buf_count = 4,
    .dma_buf_len   = 128,
    .use_apll = false,
    .tx_desc_auto_clear = false,
    .fixed_mclk = 0
  };

  i2s_pin_config_t pin_cfg = {
    .bck_io_num = MIC_I2S_SCK,
    .ws_io_num  = MIC_I2S_WS,
    .data_out_num = I2S_PIN_NO_CHANGE,
    .data_in_num  = MIC_I2S_SD
  };

  esp_err_t e1 = i2s_driver_install(I2S_PORT, &cfg, 0, NULL);
  if (e1 != ESP_OK) return false;

  esp_err_t e2 = i2s_set_pin(I2S_PORT, &pin_cfg);
  if (e2 != ESP_OK) return false;

  i2s_zero_dma_buffer(I2S_PORT);
  i2s_start(I2S_PORT);
  curMode = MODE_MIC;
  return true;
}

bool ensureSpeakerModeLocked() {
  if (curMode == MODE_SPK) return true;
  spk_ok = i2sInitSpeaker();
  return spk_ok;
}

bool ensureMicModeLocked() {
  if (curMode == MODE_MIC) return true;
  mic_ok = i2sInitMic();
  return mic_ok;
}

// ===================== Speaker sound =====================
static uint32_t rng = 1;
static inline int16_t noise16() {
  rng = rng * 1664525UL + 1013904223UL;
  return (int16_t)(rng >> 16);
}

void playBeep1kHz_200ms_locked() {
  if (!ensureSpeakerModeLocked()) return;

  const int sampleRate = 22050;
  const float freq = 1000.0f;
  const float durSec = 0.2f;
  const int total = (int)(sampleRate * durSec);

  const int N = 256;
  int16_t stereo[N * 2];
  float phase = 0.0f;

  for (int i=0;i<total;){
    int n = (total-i > N) ? N : (total-i);
    for (int k=0;k<n;k++){
      phase += 2.0f * 3.1415926f * freq / (float)sampleRate;
      if (phase > 2.0f*3.1415926f) phase -= 2.0f*3.1415926f;

      int16_t v = (int16_t)(sinf(phase) * 12000);
      stereo[2*k+0] = v;
      stereo[2*k+1] = v;
    }
    size_t written=0;
    i2s_write(I2S_PORT, (const char*)stereo, n*2*sizeof(int16_t), &written, portMAX_DELAY);
    i += n;
  }
}

void playRoarOnce_locked() {
  if (!ensureSpeakerModeLocked()) return;

  const int sampleRate = 22050;
  const float durSec = 0.6f;
  const int total = (int)(sampleRate * durSec);

  const int N = 256;
  int16_t stereo[N * 2];
  float phase = 0.0f;

  for (int i=0;i<total;){
    int n = (total-i > N) ? N : (total-i);

    for (int k=0;k<n;k++){
      float t = (float)(i+k)/(float)total;
      float f = 120.0f - 50.0f*t;
      phase += 2.0f * 3.1415926f * f / (float)sampleRate;
      if (phase > 2.0f*3.1415926f) phase -= 2.0f*3.1415926f;

      float env = (t < 0.08f) ? (t/0.08f) : (1.0f - (t-0.08f)/0.92f);
      if (env < 0) env = 0;

      float s  = sinf(phase)*0.65f + sinf(phase*0.5f)*0.35f;
      float nz = (noise16()/32768.0f)*0.35f;
      float out = (s + nz) * env;

      float gain = 0.85f;
      int32_t v = (int32_t)(out * gain * 30000.0f);
      if (v > 32767) v = 32767;
      if (v < -32768) v = -32768;

      int16_t vv = (int16_t)v;
      stereo[2*k+0] = vv;
      stereo[2*k+1] = vv;
    }

    size_t written=0;
    i2s_write(I2S_PORT, (const char*)stereo, n*2*sizeof(int16_t), &written, portMAX_DELAY);
    i += n;
  }
}

// ===================== Microphone energy =====================
static const int MIC_FRAME = 128;
static int32_t micBuf[MIC_FRAME];

float readMicEnergyOnce_locked() {
  if (!ensureMicModeLocked()) return 0;

  size_t bytesRead = 0;
  i2s_read(I2S_PORT, (void*)micBuf, sizeof(micBuf), &bytesRead, 20 / portTICK_PERIOD_MS);
  int n = bytesRead / 4;
  if (n <= 0) return 0;

  uint64_t sumAbs = 0;
  for (int i=0;i<n;i++){
    int16_t s16 = (int16_t)(micBuf[i] >> 16);
    sumAbs += (uint16_t)abs(s16);
  }
  return (float)sumAbs / (float)n;
}

// ===================== Offline voice control (方案B) =====================
// 你只需要調這兩個門檻：
// - 安靜時網頁音量可能是 0~200
// - 講話/拍手可能是 1500~8000
// 建議：VOICE_ON_TH 設在安靜值 5~10 倍；VOICE_OFF_TH 比 ON 略小
static float VOICE_ON_TH  = 300; // 有聲開始門檻（可調）
static float VOICE_OFF_TH = 100;  // 有聲結束門檻（可調）

static bool inVoice = false;
static unsigned long voiceStartMs = 0;
static unsigned long lastVoiceSampleMs = 0;

// 指令分類（可調）
static const int OPEN_MIN_MS  = 320; // 長音 -> Open
static const int CLOSE_MIN_MS = 60; // 短音 -> Close
static const int CLOSE_MAX_MS = 200;

// 冷卻時間：避免一次聲音被切成兩段重複觸發
static unsigned long cooldownUntilMs = 0;
static const int COOLDOWN_MS = 500;

void doOpenAction() {
  if (!mouthOpenState) {
    mouthOpenState = true;
    lastCmd = "OPEN";
    drawTigerMouth(true);

    // 張嘴時吼叫一次
    lockI2S();
    playRoarOnce_locked();
    unlockI2S();
  } else {
    lastCmd = "OPEN(IGN)"; // 已經張開就忽略
  }
}

void doCloseAction() {
  if (mouthOpenState) {
    mouthOpenState = false;
    lastCmd = "CLOSE";
    drawTigerMouth(false);
  } else {
    lastCmd = "CLOSE(IGN)";
  }
}

void voiceControlTick() {
  unsigned long now = millis();
  if (now < cooldownUntilMs) return;

  // 每 60ms 抽樣一次
  if (now - lastVoiceSampleMs < 60) return;
  lastVoiceSampleMs = now;

  float e = 0;

  // 讀麥克風能量（短暫切到 MIC，再切回 SPK）
  lockI2S();
  e = readMicEnergyOnce_locked();
  ensureSpeakerModeLocked();
  unlockI2S();

  lastMicEnergy = e;

  if (!inVoice) {
    if (e > VOICE_ON_TH) {
      inVoice = true;
      voiceStartMs = now;
    }
  } else {
    if (e < VOICE_OFF_TH) {
      inVoice = false;
      int dur = (int)(now - voiceStartMs);

      if (dur >= OPEN_MIN_MS) {
        doOpenAction();
        cooldownUntilMs = now + COOLDOWN_MS;
      } else if (dur >= CLOSE_MIN_MS && dur <= CLOSE_MAX_MS) {
        doCloseAction();
        cooldownUntilMs = now + COOLDOWN_MS;
      } else {
        lastCmd = "NOISE";
      }
    }
  }
}

// ===================== Web UI =====================
String htmlPage() {
  String h;
  h += "<!doctype html><html><head><meta charset='utf-8'/>";
  h += "<meta name='viewport' content='width=device-width,initial-scale=1'/>";
  h += "<title>OTTO GO Test</title>";
  h += "<style>body{font-family:Arial;margin:16px}button{padding:10px 14px;margin:6px;border-radius:10px;border:0;background:#2d7;color:#fff;font-size:16px}";
  h += ".row{margin:10px 0}.card{padding:12px;border:1px solid #ddd;border-radius:12px;margin:10px 0}";
  h += "input[type=range]{width:260px}</style></head><body>";
  h += "<h2>OTTO GO 功能測試（192.168.4.1）</h2>";

  h += "<div class='card'><h3>離線語音控制（方案B：聲音長短）</h3>";
  h += "<div>音量：<span id='mic'>--</span>　狀態：<span id='cmd'>--</span></div>";
  h += "<div>門檻 ON/OFF：<span id='th'>--</span></div>";
  h += "<div>規則：長音(&gt;450ms)=OPEN、短音(120~320ms)=CLOSE</div>";
  h += "</div>";

  h += "<div class='card'><h3>Speaker（SPK 白色座）</h3>";
  h += "<button onclick=\"fetch('/spk?m=beep')\">Beep</button>";
  h += "<button onclick=\"fetch('/spk?m=roar')\">老虎大叫</button>";
  h += "<div>spk_ok：<span id='spkok'>--</span>　mic_ok：<span id='micok'>--</span></div>";
  h += "</div>";

  h += "<div class='card'><h3>TFT</h3>";
  h += "<button onclick=\"fetch('/tft?m=hello')\">Hello</button>";
  h += "<button onclick=\"fetch('/tft?m=mouth_open')\">虎嘴張開</button>";
  h += "<button onclick=\"fetch('/tft?m=mouth_close')\">虎嘴閉合</button>";
  h += "<button onclick=\"fetch('/tft?m=clear')\">清屏</button>";
  h += "</div>";

  h += "<div class='card'><h3>超音波測距</h3>";
  h += "<div>距離：<span id='dist'>--</span> cm</div>";
  h += "</div>";

  h += "<div class='card'><h3>Servo</h3>";
  h += "<button onclick=\"fetch('/servo?all=90')\">四顆 90 度校正</button><br/>";
  for (int i=1;i<=4;i++){
    h += "<div class='row'>S"; h += i; h += " 角度：";
    h += "<input type='range' min='0' max='180' value='90' id='s"; h += i; h += "' ";
    h += "oninput=\"setServo("; h += i; h += ",this.value)\"/>";
    h += " <span id='sv"; h += i; h += "'>90</span></div>";
  }
  h += "</div>";

  h += "<script>";
  h += "function setServo(ch,v){document.getElementById('sv'+ch).innerText=v;fetch('/servo?ch='+ch+'&ang='+v);}";

  h += "async function tick(){";
  h += "try{let r=await fetch('/status'); let j=await r.json();";
  h += "document.getElementById('mic').innerText=(+j.energy).toFixed(0);";
  h += "document.getElementById('cmd').innerText=j.cmd;";
  h += "document.getElementById('th').innerText=j.th;";
  h += "document.getElementById('spkok').innerText=j.spk_ok;";
  h += "document.getElementById('micok').innerText=j.mic_ok;";
  h += "}catch(e){}";
  h += "try{let r2=await fetch('/ultra'); let j2=await r2.json(); document.getElementById('dist').innerText=j2.cm;}catch(e){}";
  h += "setTimeout(tick,500);"; // 網頁顯示 500ms 更新即可
  h += "}";
  h += "tick();</script>";

  h += "</body></html>";
  return h;
}

// ===================== Web handlers =====================
void handleRoot() {
  server.send(200, "text/html; charset=utf-8", htmlPage());
}

void handleTFT() {
  String m = server.arg("m");
  if (m == "hello") tftHello();
  else if (m == "mouth_open") { mouthOpenState = true; lastCmd="WEB_OPEN"; drawTigerMouth(true); }
  else if (m == "mouth_close"){ mouthOpenState = false; lastCmd="WEB_CLOSE"; drawTigerMouth(false); }
  else if (m == "clear") tft.fillScreen(ST77XX_BLACK);
  server.send(200, "text/plain", "OK");
}

void handleSPK() {
  String m = server.arg("m");
  lockI2S();
  if (m == "beep") playBeep1kHz_200ms_locked();
  else if (m == "roar") playRoarOnce_locked();
  unlockI2S();
  server.send(200, "text/plain", "OK");
}

void handleUltra() {
  long d1 = readDistanceCM();
  long d2 = readDistanceCM();
  long d3 = readDistanceCM();
  long d  = median3(d1,d2,d3);
  if (d < 0) d = -1;
  String json = "{\"cm\":" + String(d) + "}";
  server.send(200, "application/json", json);
}

void handleServo() {
  if (server.hasArg("all")) {
    int a = constrain(server.arg("all").toInt(), 0, 180);
    s1.write(a); s2.write(a); s3.write(a); s4.write(a);
    server.send(200, "text/plain", "OK");
    return;
  }
  int ch = server.arg("ch").toInt();
  int ang = constrain(server.arg("ang").toInt(), 0, 180);
  switch(ch){
    case 1: s1.write(ang); break;
    case 2: s2.write(ang); break;
    case 3: s3.write(ang); break;
    case 4: s4.write(ang); break;
    default: break;
  }
  server.send(200, "text/plain", "OK");
}

void handleStatus() {
  // 注意：lastMicEnergy/lastCmd 是背景更新
  String th = String((int)VOICE_ON_TH) + "/" + String((int)VOICE_OFF_TH);
  String json = "{";
  json += "\"energy\":" + String((float)lastMicEnergy, 1) + ",";
  json += "\"cmd\":\"" + lastCmd + "\",";
  json += "\"th\":\"" + th + "\",";
  json += "\"spk_ok\":" + String(spk_ok ? 1 : 0) + ",";
  json += "\"mic_ok\":" + String(mic_ok ? 1 : 0);
  json += "}";
  server.send(200, "application/json", json);
}

// ===================== Setup/Loop =====================
void setup() {
  Serial.begin(115200);

  i2sMutex = xSemaphoreCreateMutex();

  pinMode(US_TRIG, OUTPUT);
  pinMode(US_ECHO, INPUT);

  SPI.begin(TFT_SCLK, -1, TFT_MOSI, TFT_CS);
  tft.init(240, 320);
  tft.setRotation(1);
  tftHello();

  s1.attach(SERVO1, 600, 2400);
  s2.attach(SERVO2, 600, 2400);
  s3.attach(SERVO3, 600, 2400);
  s4.attach(SERVO4, 600, 2400);
  s1.write(90); s2.write(90); s3.write(90); s4.write(90);

  // 開機先穩定進入 Speaker 模式並 beep
  lockI2S();
  spk_ok = i2sInitSpeaker();
  if (spk_ok) playBeep1kHz_200ms_locked();
  unlockI2S();

  // WiFi AP
  WiFi.mode(WIFI_AP);
  WiFi.softAP(AP_SSID, AP_PASS);
  Serial.print("AP IP: ");
  Serial.println(WiFi.softAPIP());

  // Web routes
  server.on("/", handleRoot);
  server.on("/tft", handleTFT);
  server.on("/spk", handleSPK);
  server.on("/ultra", handleUltra);
  server.on("/servo", handleServo);
  server.on("/status", handleStatus);
  server.begin();

  Serial.println("Web server started.");
}

void loop() {
  server.handleClient();

  // 背景離線語音控制（方案B）
  voiceControlTick();
}
智慧生活科技專業社群

2026年2月13日星期五

[OTTO GO] 離線版語音控制，運用長短音

沒有留言:

張貼留言

2026年2月13日 星期五

[OTTO GO] 離線版語音控制，運用長短音

沒有留言:

張貼留言

2026年2月13日星期五
