Nail/elmer
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

refactor: move legacy docs to doc/legacy/ with opus/sonnet subfolders

Browse Source
This commit is contained in:
Repinoid
2026-07-10 13:13:42 +04:00
parent 3ea7fc66d5
commit 95834a4245
46 changed files with 0 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
doc/legacy/research/competitors.md
+173
View File
@@ -0,0 +1,173 @@
# Конкуренты (2026-05-25)
## Резюме
На рынке есть 4 буржуазных конкурента. **Русскоязычных — ноль.** RuStore пустой по запросу «OBD2 + AI диагностика».
Это главное преимущество: мы не «ещё один», а **первый на русском рынке**.
---
## DiagnostiX AI
| | |
|---|---|
| **Разработчик** | Exza / Ontario Analytics (Канада) |
| **Платформа** | Android (Google Play) |
| **ID** | `com.exza.diagnostixai` |
| **Описание** | OBD2 scanner + AI mechanic assistant. Чтение/сброс DTC, live data, сенсоры в реальном времени, AI-помощь |
| **LLM** | Не раскрыт |
| **Слабость** | Всё в одном app, locked-in |
## OBDAI
| | |
|---|---|
| **Разработчик** | Ontario Analytics (Канада/США) |
| **Платформа** | iOS + Android |
| **ID** | `com.ontarioanalytics.obdai` |
| **Описание** | AI-агент ARIA (Automotive Reasoning & Intelligence Agent). Профессиональные диагностические отчёты. «Поговори с машиной» |
| **LLM** | ARIA (вероятно свой/кастомный) |
| **Слабость** | Закрытый, платный, английский |
## Zero Touch Car Diagnostics
| | |
|---|---|
| **Разработчик** | kellinheller (GitHub) |
| **Платформа** | Кроссплатформа (Flutter) |
| **Репозиторий** | `github.com/kellinheller/zero_touch_car_diagnostics` |
| **Описание** | Open source. OBD2 + датчики телефона + GPS + AI-анализ через Google Gemini 2.5 Pro |
| **LLM** | Google Gemini 2.5 Pro |
| **Слабость** | Gemini платный. Flutter = тяжёлый. Не для ELM327 v1.5 (клонов) |
## MUCAR 892BT MUAI (THINKCAR)
| | |
|---|---|
| **Производитель** | THINKCAR (Китай) |
| **Тип** | Железка + софт |
| **Описание** | Полноценный диагностический сканер со своим экраном. 8 диагностических модулей, отчёты |
| **LLM** | **DeepSeek** (единственный конкурент на DeepSeek) |
| **Слабость** | Свой сканер (~$150-300). Не работает с ELM327 |
---
## Сравнение с Elmer
| Критерий | DiagnostiX AI | OBDAI | Zero Touch | MUCAR MUAI | **Elmer** |
|---|---|---|---|---|---|
| Open source клиент | ❌ | ❌ | ✅ | ❌ | ✅ |
| Тонкий клиент (свой сервер) | ❌ | ❌ | ❌ | ❌ | ✅ |
| Русский язык | ❌ | ❌ | ❌ | ❌ | ✅ |
| RuStore | ❌ | ❌ | ❌ | ❌ | ✅ |
| Работает с дешёвым ELM327 | ✅ | ✅ | 🟡 | ❌ | ✅ |
| DeepSeek | ❌ | ❌ | ❌ | ✅ | ✅ |
| Бесплатно | 🟡 | ❌ | ✅ | ❌ | ✅ |
---
## Вывод
Ниша существует и растёт. MUCAR с DeepSeek — сигнал что LLM-подход правильный.
Наш козырь: **открытость + русский рынок + дешёвый ELM327**. Никто не сочетает эти три фактора.
---
## 2026-05-27: Дополнительные находки
### Automotive AI (Eloquent-Algorithmics)
| | |
|---|---|
| **Репозиторий** | `github.com/Eloquent-Algorithmics/Automotive-AI` |
| **Платформа** | Windows / Linux (Python) |
| **Описание** | Экспериментальный проект. ELM327 (BT или эмулятор) → OpenAI GPT. Голосовой ввод/вывод: «считай коды», «сделай отчёт». |
| **Технологии** | Python, OpenAI API |
| **LLM** | OpenAI GPT (заменяемый) |
| **Обновление** | Май 2025 |
| **Плюсы** | Прямой конвейер ELM327→LLM. Умеет DTC, параметры, стоп-кадры. Голос — интересная фича для гаража (руки грязные). |
| **Минусы** | Только десктоп, не Android. Голос бесполезен в движении. Английский. |
| **Вывод** | Смотреть архитектуру: как парсит ответы, как формирует промпт. Голос — возможно позже для нашей веб-панели. |
### OBD2AI
| | |
|---|---|
| **Платформа** | Android (готовое приложение) |
| **Описание** | ELM327 по BT → ChatGPT → анализ ошибок. Простая схема без промежуточного сервера. |
| **LLM** | ChatGPT |
| **Плюсы** | Быстрый старт — установил и работает. Можно использовать как референс перед поездкой к машине. |
| **Минусы** | Только ChatGPT. Без сервера — нет batch-загрузки, нет истории, нет своей LLM. |
| **Вывод** | Проверить как референс Android-клиента. Но наша архитектура (fat-client + свой сервер) гибче. |
### Car OBD2 Diagnostics (kagibson)
| | |
|---|---|
| **Репозиторий** | `github.com/kagibson/car_obd2_diagnostics` |
| **Платформа** | Linux/Windows (Docker) |
| **Описание** | «Vibe coding» эксперимент. ELM327 → веб-панель с RPM, скоростью, температурой, DTC, стоп-кадрами. |
| **Технологии** | Docker, Python (бэкенд), React (фронтенд) |
| **LLM** | Не указан (вероятно подключаемый) |
| **Плюсы** | Чистый читаемый код. Веб-панель с графиками — хороший референс для нашей веб-части. Docker — лёгкий деплой. |
| **Минусы** | Только десктоп, не Android. Нет LLM-диагноза из коробки. |
| **Вывод** | Изучить архитектуру веб-панели (React-компоненты для отображения PID). Docker — возможно для production-деплоя elmer-сервера. |
---
## Обновлённое сравнение (2026-05-27)
| Критерий | DiagnostiX | OBDAI | ZeroTouch | MUCAR | AutoAI | OBD2AI | CarOBD2 | **Elmer** |
|---|---|---|---|---|---|---|---|---|
| **Android** | ✅ | ✅ | ✅ | ❌ | ❌ | ✅ | ❌ | ✅ |
| **Offline-скрипты** | ❌ | ❌ | ❌ | ✅ | ❌ | ❌ | ❌ | ✅ |
| **Свой сервер** | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ✅ | ✅ |
| **Русский язык** | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ✅ |
| **Жалобы водителя** | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ✅ (план) |
| **Cross-val LLM** | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ✅ (план) |
| **Любой LLM** | ❌ | ❌ | ❌ | ❌ | 🟡 | ❌ | ❌ | ✅ |
| **Open source** | ❌ | ❌ | ✅ | ❌ | ✅ | ❌ | ✅ | ✅ |
| **Дешёвый ELM327** | ✅ | ✅ | 🟡 | ❌ | ✅ | ✅ | ✅ | ✅ |
**Уникальных фич Elmer:** 7 из 11 (отмечены ✅ только у нас).
---
## 2026-05-27 (2): Масштабный поиск — 50+ проектов
Найдено через GitHub Code Search: 50+ open-source проектов на стыке ELM327 + LLM.
### 🔥 Проекты с рабочим Android→ELM→LLM (можно изучать код)
| Проект | Язык | LLM | Звёзды | Обновлён | Ценность для нас |
|--------|------|-----|--------|----------|------------------|
| **Wal33D/OBD-Droid** | Java | ChatGPT | 4 | Фев 2025 | ⭐ **ЛУЧШИЙ.** Рабочий Android→ELM→LLM. AI copilot, GPS, NHTSA |
| **catsmoker/OBD2AI** | Kotlin | gpt-5-mini | 3 | Окт 2025 | ⭐ Android Kotlin. BT SPP + BLE |
| **petrpatek/obd2-mcp-server** | Python | Claude MCP | 2 | 10 дней назад | Свежий. 1937 Ford DTC |
| **castlebbs/Vehicle-Diagnostic-Assistant** | Python | DeepSeek/Claude | 2 | Дек 2025 | Hackathon-победитель. W600 MCU |
### 🧠 Что украдено из OBD-Droid (Java Android)
**Подтверждённые паттерны ELM327-коммуникации:**
1. **UUID SPP:** `00001101-0000-1000-8000-00805F9B34FB` ✅ как у нас
2. **Чтение:** побайтово, сон 1мс между проверками (не sleep(250)!)
3. **Детекция промпта `>`:** КРИТИЧНО — не слать следующую команду пока не получили `>`
4. **Обработка ошибок:** SEARCHING → ждать, NO DATA → пропустить, UNABLE → reconnect, CAN ERROR → warm start
5. **Мульти-фрейм ISO-TP:** строки с префиксом `:`, буферизация до полного ответа
6. **Адаптивный таймаут:** базовый 5000мс, увеличивается для медленных шин
7. **`flush()` после каждой команды** — иначе данные теряются
8. **ATH1/ATH0:** управление заголовками CAN
**Что у нас НЕ так по сравнению с OBD-Droid:**
- ❌ Мы ждём `sleep(250)` вместо ожидания `>` промпта
- ❌ Нет адаптивного таймаута
- ❌ Нет recovery при CAN ERROR / UNABLE TO CONNECT
- ❌ Нет `flush()` после отправки
### 📊 Итоговая картина рынка
| Тип | Количество |
|-----|-----------|
| ELM327 + LLM (всего) | 50+ |
| Рабочий Android→ELM→LLM | 4 |
| Русский язык | **0 (только мы)** |
| Offline fat-client | **0 (только мы)** |
| Жалобы водителя | **0 (только мы)** |
doc/legacy/research/elm-reference.md
+710
View File
@@ -0,0 +1,710 @@
# ELM327 Communication Patterns — анализ 5+14 проектов
> **Цель:** понять как РЕАЛЬНО работают проекты с ELM327, выбрать лучшие паттерны для Elmer.
> **Дата:** 2026-05-27
> **Источники:** 5 LLM-проектов + 14 традиционных OBD2 Android-проектов
---
## Часть 1: Проекты с LLM (ELM → LLM → диагноз)
### Сводная таблица
| | OBD-Droid | OBD2AI | Automotive-AI | obd2-mcp-server | Vehicle-Diag-Assist |
|---|---|---|---|---|---|
| **Язык** | Java | Kotlin | Python | Python | C (W600) + Python |
| **Платформа** | Android | Android | Desktop | Desktop/Claude MCP | Embedded (MCU) |
| **LLM** | ChatGPT | gpt-5-mini | GPT-3.5/4 | Claude (MCP) | DeepSeek/Claude |
| **Чтение** | Побайтово, 1мс | kotlin-obd lib | readline() | Побайтово (BLE/SPP) | UART, семафор |
| **UUID** | 00001101... | 00001101... | N/A (pyserial) | BLE + serial | N/A (UART) |
| **Baud** | — | — | config.py | 38400 (auto-retry) | 38400 8N1 |
| **Timeout** | Адаптивный 5с | 400мс fix | 1с | 20с connect / 2с config | 2000мс |
| **Инит** | ATD→ATE0→ATL0→ATS0→ATH1→... | ATZ→ATE0→ATL0→ATSP0 | N/A | ATZ→ATE0→ATL0→ATS0→ATH1→ATCAF1→ATAT1→ATST64→ATSP0 | ATZ→... |
| **Ретраи** | requeue + SETPROT | 3 strikes → stop | Нет | [2,5,10]с backoff | Нет |
| **Simulator** | Встроенный demo | Нет | ELM327-emulator | Mock mode (Ford) | Gradio + HW sim |
| **DTC база** | Встроенная | Нет | Нет | 1937 Ford + generic | Нет |
---
## 1. OBD-Droid (Wal33D) — Java Android ⭐ ЛУЧШИЙ
### 1.1. StreamHandler.java — побайтовый I/O
```java
// ЧТЕНИЕ: побайтово, сон 1мс между проверками
public void run() {
while (true) {
if (in.available() > 0) {
if ((chr = in.read()) > 0) {
processRxChar(chr);
} else break;
} else {
Thread.sleep(1); // ← 1 миллисекунда!
}
}
}
// ОБРАБОТКА СИМВОЛОВ: '>' = такой же разделитель как CR/LF!
private void processRxChar(int chr) {
switch (chr) {
case 32: break; // пробел — игнорируем
case '>': // промпт ELM
message += (char) chr;
// fall through — НЕ отдельный случай!
case 10: // LF
case 13: // CR
messageHandler.handleTelegram(message.toCharArray());
message = "";
break;
default:
message += (char) chr;
}
}
// ОТПРАВКА: BufferedWriter с буфером 1 байт = flush на каждом байте
out = new BufferedWriter(new OutputStreamWriter(outStream), 1);
public int writeTelegram(final char[] buffer, int type, Object id) {
new Thread(() -> {
String msg = new String(buffer) + "\r"; // ELM ждёт CR
out.write(msg.toCharArray());
out.flush(); // немедленный flush из-за буфера 1 байт
}).start();
return buffer.length;
}
```
**Ключевые выводы:**
- `>` — НЕ спецсигнал «можно слать дальше». Это просто разделитель строк, как CR/LF.
- Буфер 1 байт на запись = каждый байт сразу уходит в порт.
- Отправка в отдельном потоке (не блокирует чтение).
### 1.2. ElmProt.java — стейт-машина протокола
**RSP_ID — все возможные ответы ELM327:**
```java
enum RSP_ID {
PROMPT(">"), OK("OK"), MODEL("ELM"),
NODATA("NODATA"), SEARCH("SEARCHING"),
ERROR("ERROR"), NOCONN("UNABLE"), NOCONN2("NABLETO"),
CANERROR("CANERROR"), BUSBUSY("BUSBUSY"),
BUSERROR("BUSERROR"), BUSINIERR("BUSINIT:ERR"),
BUSINIERR2("BUSINIT:BUS"), BUSINIERR3("BUSINIT:...ERR"),
FBERROR("FBERROR"), DATAERROR("DATAERROR"),
BUFFERFULL("BUFFERFULL"), STOPPED("STOPPED"),
RXERROR("<"), QMARK("?"),
UNKNOWN("");
}
```
**STAT — состояния соединения:**
```java
UNDEFINED → INITIALIZING → INITIALIZED → ECU_DETECT → ECU_DETECTED
→ ECU_SELECTED → CONNECTING → CONNECTED
// Ошибки:
NODATA, STOPPED, DISCONNECTED, BUSERROR, DATAERROR, RXERROR, ERROR
```
**Инициализация (после ATZ → MODEL):**
```
ATD // defaults
ATE0 // echo off
ATL0 // line feeds off
ATS0 // spaces off
ATH1 // headers ON (для обнаружения ЭБУ)
ATDP // узнать протокол
ATSPA1 // протокол AUTO
ATAT1 // adaptive timing ON
ATST<value> // установить таймаут
```
**Обработка ошибок — детально:**
```
SEARCHING → статус CONNECTING (не ошибка!)
NODATA → увеличить OBD timeout + переустановить протокол
ERROR → WARMSTART (ATWS)
DATAERROR → WARMSTART
RXERROR → WARMSTART
BUFFERFULL→ WARMSTART
BUS ERROR → DISCONNECTED → переустановить протокол + ретрай последней команды
UNABLE → DISCONNECTED → переустановить протокол + ретрай
```
**Мульти-фрейм ISO-TP:**
```
Формат: "0:4100..." — первая строка с длиной
"1:4100..." — продолжение
charsExpected = байт_длины * 2 (каждый байт = 2 hex символа)
```
### 1.3. BluetoothCommService.java — BT SPP
```java
final UUID SPP_UUID = UUID.fromString("00001101-0000-1000-8000-00805F9B34FB");
// Первая попытка: стандартный RFCOMM
tmp = device.createRfcommSocketToServiceRecord(SPP_UUID); // secure
// или
tmp = device.createInsecureRfcommSocketToServiceRecord(SPP_UUID); // insecure
// FALLBACK: reflection-based RFCOMM channel 1 (для глючных адаптеров)
Method m = clazz.getMethod("createRfcommSocket", paramTypes);
Object[] params = new Object[]{1}; // channel 1
sockFallback = (BluetoothSocket) m.invoke(device, params);
```
**Ключевой вывод:** Есть fallback на reflection-based RFCOMM channel 1 — для дешёвых китайских клонов!
---
## 2. OBD2AI (catsmoker) — Kotlin Android
### 2.1. BluetoothHelper
```kotlin
val sppUuid: UUID = UUID.fromString("00001101-0000-1000-8000-00805F9B34FB")
suspend fun connectToDevice(deviceAddress: String): Pair<InputStream, OutputStream> {
val device = bluetoothAdapter?.getRemoteDevice(deviceAddress)
bluetoothSocket = device.createRfcommSocketToServiceRecord(sppUuid).apply {
bluetoothAdapter.cancelDiscovery()
connect()
}
return Pair(socket.inputStream, socket.outputStream)
}
```
### 2.2. ObdHelper — инициализация и команды
```kotlin
// Инициализация: фиксированные задержки, БЕЗ ожидания '>'
suspend fun initializeObd() = withContext(Dispatchers.IO) {
suspend fun sendRawCommand(command: String) {
out.write((command + "\r").toByteArray())
out.flush()
delay(400) // ← 400мс после КАЖДОЙ команды
}
sendRawCommand("ATZ") // сброс
sendRawCommand("ATE0") // эхо выкл
sendRawCommand("ATL0") // line feeds выкл
sendRawCommand("ATSP0") // авто-протокол
delay(1000) // дополнительная пауза после инита
// Очистка буфера
if (`in`.available() > 0) {
val buffer = ByteArray(`in`.available())
`in`.read(buffer)
}
}
```
**Используется библиотека `kotlin-obd` (eltonvs):**
```kotlin
// Для стандартных команд — библиотека
obdConnection = ObdDeviceConnection(inputStream, outputStream)
val result = connection.run(TroubleCodesCommand())
// Для нестандартных — ручной парсинг
class MyRPMCommand : ObdCommand() {
override val pid = "0C"
override val handler = { it: ObdRawResponse ->
val rawValue = it.processedValue
val identifier = "410C"
val aHex = rawValue.substring(index + 4, index + 6)
val bHex = rawValue.substring(index + 6, index + 8)
((a * 256) + b) / 4 // формула RPM
}
}
```
### 2.3. Live Data Monitoring
```kotlin
suspend fun startLiveDataMonitoring() = withContext(Dispatchers.IO) {
var errorCount = 0
while (isMonitoring.get()) {
try {
val speed = runCommand(MySpeedCommand())
val rpm = runCommand(MyRPMCommand())
val temp = runCommand(MyCoolantTempCommand())
errorCount = 0
delay(800) // 800мс между циклами
} catch (e: Exception) {
errorCount++
if (errorCount >= 3) break // 3 ошибки подряд = стоп
delay(1000)
}
}
}
```
---
## 3. Automotive-AI (Eloquent-Algorithmics) — Python Desktop
### 3.1. ELM327 через pyserial
```python
# config.py
SERIAL_PORT = "/dev/ttyUSB0" # или COM3 на Windows
BAUD_RATE = 38400
# Подключение
ser = serial.Serial(port=SERIAL_PORT, baudrate=BAUD_RATE, timeout=1)
# Отправка команды
def send_command(ser, command):
ser.write((command + "\r\n").encode()) # CRLF терминатор
response = ser.readline().decode().strip()
response = response.replace("\r", "").replace(">", "")
return response
```
**Ключевые отличия от OBD-Droid:**
- `readline()` вместо побайтового чтения — ПРОЩЕ, но менее надёжно
- `\r\n` вместо просто `\r`
- `timeout=1` — ждёт 1 секунду на readline
- Убирает `>` из ответа (не использует как разделитель)
### 3.2. Парсинг ответов
```python
# RPM: 010C → 41 0C HH LL
if cmd == "010C":
value = (int(response.split()[2], 16) * 256 +
int(response.split()[3], 16)) / 4
# Coolant: 0105 → 41 05 XX
if cmd == "0105":
value = int(response.split()[2], 16) - 40 # -40 offset
# VIN: 0902
vin_response = parse_vin_response(response)
vehicle_data = decode_vin(vin_response)
```
---
## 4. obd2-mcp-server (petrpatek) — Python Claude MCP ⭐ САМЫЙ СВЕЖИЙ
### 4.1. BLE + Serial подключение
```
Поддерживает:
- BLE (vLinker FD, STN чип) — асинхронный, asyncio.Lock
- Serial (classic Bluetooth SPP) — синхронный, pyserial
Baud rate auto-retry: [500k, 115.2k, 38.4k, 9.6k]
BLE: 30-секундный keepalive heartbeat (без него адаптер засыпает через ~120с)
```
### 4.2. Инициализация (САМАЯ ПОЛНАЯ)
```python
ATZ # сброс
ATE0 # эхо выкл
ATL0 # line feeds выкл
ATS0 # пробелы выкл
ATH1 # заголовки CAN ВКЛ (для обнаружения ЭБУ)
ATCAF1 # CAN auto-formatting ON
ATAT1 # adaptive timing ON
ATST64 # timeout = 64*4ms = 256ms
ATSP0 # авто-протокол
# Для STN адаптеров (OBDlink):
ATPP 0E SV 00 # отключить сон
ATPP 0E ON # включить
```
### 4.3. Ретраи и таймауты
```python
MAX_RETRIES = 3
RETRY_BACKOFF = [2, 5, 10] # секунды
CONNECT_TIMEOUT = 20 # секунд
PROTOCOL_TIMEOUT = 12 # секунд
CONFIG_TIMEOUT = 2 # секунды
```
### 4.4. Очистка ответа
```python
def _clean_elm_response(raw: str) -> str:
# Убирает: промпт ">", эхо команд, "SEARCHING...", пустые строки
...
```
### 4.5. DTC база данных
```
- 1937 Ford-специфичных кодов
- Generic OBD-II коды (P, B, C, U)
- Ленивая загрузка по бренду
- Скрапинг с troublecodes.net
```
---
## 5. Vehicle-Diagnostic-Assistant (castlebbs) — Embedded C + Python
### 5.1. Аппаратная архитектура
```
W600 MCU ←UART1 38400 8N1→ ELM327 чип → CAN → Авто
↕ HTTP/MCP
LangChain Agent (Python) → DeepSeek / Claude
```
### 5.2. ELM327 Driver (C)
```c
// elm327.c
int elm327_send_command(const char* cmd, char* resp, int len, int timeout) {
// Пишет команду + \r в UART1
// Ждёт ответ через FreeRTOS semaphore (прерывание по приёму)
// Таймаут по умолчанию: 2000мс
// Макс. длина ответа: 512 байт
}
// Hybrid simulation mode:
// AT команды → реальный ELM327
// OBD команды → симуляция (если включена)
```
### 5.3. Поддерживаемые режимы OBD
```
Mode 01: live data (30+ PID)
Mode 03: stored DTC (формат 43 XX XX XX XX)
Mode 04: clear DTC (44)
Mode 07: pending DTC (47)
Mode 09: vehicle info (VIN, calibration ID)
```
### 5.4. PID формулы (Mode 01)
| PID | Формула | Пример |
|-----|---------|--------|
| 0C (RPM) | `(A*256+B)/4` | 0x1AF8 → 1726 |
| 0D (Speed) | `A` (km/h) | 0x00 → 0 |
| 05 (ECT) | `A-40` (°C) | 0x5A → 50 |
| 04 (Load) | `(A*100)/255` (%) | 0x40 → 25.1 |
| 10 (MAF) | `((A*256)+B)/100` (g/s) | — |
| 2F (Fuel) | `(A*100)/255` (%) | — |
### 5.5. Safe formula evaluation
```python
def calculate_obd_value(raw_response, formula):
# LLM вызывает этот tool для расчёта значений
# safe_eval() — ограниченный eval (только +-*/ и переменные A,B,C,D)
hex_bytes = raw_response.replace("41 XX ", "").split()
A, B, C, D = [int(x, 16) for x in hex_bytes]
return safe_eval(formula, {"A": A, "B": B, "C": C, "D": D})
```
---
## СРАВНИТЕЛЬНЫЙ АНАЛИЗ: Что взять для Elmer
### Инициализация ELM327
| Проект | Последовательность | Задержки |
|--------|-------------------|----------|
| OBD-Droid | ATD→ATE0→ATL0→ATS0→ATH1→ATDP→ATSPA1→ATAT1→ATST | Стейт-машина, нет фикс. задержек |
| OBD2AI | ATZ→ATE0→ATL0→ATSP0 | 400мс после каждой |
| obd2-mcp | ATZ→ATE0→ATL0→ATS0→ATH1→ATCAF1→ATAT1→ATST64→ATSP0 | async, по ответам |
| Automotive-AI | Нет явной инициализации | — |
**Рекомендация для Elmer:** взять последовательность obd2-mcp-server (самая полная) + задержки OBD2AI (400мс) + ATH0 вместо ATH1 (для чистых ответов без CAN-заголовков).
### Чтение ответов
| Проект | Метод | Плюсы | Минусы |
|--------|-------|-------|--------|
| OBD-Droid | Побайтово, 1мс sleep | Макс. контроль | Сложный код |
| OBD2AI | kotlin-obd lib | Готовое решение | Зависимость от библиотеки |
| Automotive-AI | `ser.readline()` | Простой код | Менее надёжно |
**Рекомендация для Elmer:** для Android — побайтовое чтение как у OBD-Droid (уже есть в TestService). Для Python-мока/сервера — `readline()` достаточно для тестов.
### Обработка ошибок
| Ошибка | OBD-Droid | OBD2AI | obd2-mcp |
|--------|-----------|--------|----------|
| SEARCHING | Статус CONNECTING | — | Пропустить, ждать |
| NO DATA | Увеличить timeout | — | Вернуть пусто |
| BUS ERROR | DISCONNECTED + retry | — | — |
| UNABLE | DISCONNECTED + retry | — | — |
| ERROR | WARMSTART (ATWS) | — | — |
| RX ERROR | WARMSTART | 3 strikes → stop | — |
**Рекомендация для Elmer:** SEARCHING = ждать + увеличить таймаут. NO DATA = пропустить PID. BUS ERROR/UNABLE = одна попытка reconnect + retry. ERROR = WARMSTART.
### Тайминги
| Проект | Между командами | Инит | Таймаут ответа |
|--------|-----------------|------|----------------|
| OBD-Droid | Нет (стейт-машина) | Стейт-машина | 5000мс адаптивный |
| OBD2AI | 400мс fix | 1000мс после всех | ? (внутри lib) |
| Automotive-AI | Нет | Нет | 1000мс (readline) |
| obd2-mcp | По ответам | По ответам | 2000-20000мс |
| castlebbs | По семафору | — | 2000мс |
**Рекомендация для Elmer:** 400мс между командами (как OBD2AI) + адаптивный таймаут от 2000мс с возможностью увеличения (как OBD-Droid).
### BT подключение (Android)
| Проект | Метод | Fallback |
|--------|-------|----------|
| OBD-Droid | `createRfcommSocketToServiceRecord` secure + insecure | Reflection RFCOMM channel 1 |
| OBD2AI | `createRfcommSocketToServiceRecord` | Нет |
**Рекомендация для Elmer:** взять fallback на reflection channel 1 из OBD-Droid — критично для дешёвых клонов.
---
## ИТОГ: Что реализовать в elmer-android
### Приоритет 1 (обязательно)
- [ ] Побайтовое чтение с паузой 1мс (StreamHandler.java)
- [ ] `>` = разделитель строк, НЕ спецсигнал
- [ ] Fallback RFCOMM channel 1 (BluetoothCommService.java)
- [ ] Фиксированные задержки 400мс между командами (OBD2AI)
- [ ] Очистка буфера после инициализации
### Приоритет 2 (важно)
- [ ] Обработка SEARCHING, NO DATA, BUS ERROR
- [ ] 3-strike retry для live monitoring
- [ ] Адаптивный таймаут (базовый 5000мс)
### Приоритет 3 (для production)
- [ ] WARMSTART при ERROR/DATAERROR
- [ ] Мульти-фрейм ISO-TP
- [ ] DTC база (можно с obd2-mcp-server)
- [ ] Экспоненциальный backoff для ретраев
---
## Часть 2: Традиционные OBD2 Android-проекты (БЕЗ LLM)
> Только ELM327 ↔ Android Bluetooth SPP. Именно они интересны для слоя коммуникации — оттестированы годами на тысячах машин.
### Топ-10 батл-тестед проектов
| # | Проект | URL | Язык | Создатель | Обновлён | ⭐ | DTC | Live | VIN | Примечания |
|---|--------|-----|------|-----------|----------|-------|-----|------|-----|-----------|
| 1 | **AndrOBD** | [fr3ts0n/AndrOBD](https://github.com/fr3ts0n/AndrOBD) | Java | Sepp Seidel | Апр 2025 | 1993 | ✅ | ✅ | ✅ | **Король.** 10+ лет, MQTT, графики, плагины, многоязычный |
| 2 | **AndroidOBD** | [barnhill/AndroidOBD](https://github.com/barnhill/AndroidOBD) | Kotlin | Brian Barnhill | Ноя 2025 | 79 | ✅ | ✅ | ✅ | **Библиотека.** Хороший API для интеграции |
| 3 | **Java OBD** | [Tomiwa-Ot/obd](https://github.com/Tomiwa-Ot/obd) | Java | Tomiwa O. | Март 2025 | 30 | ✅ | ✅ | ✓ | **Библиотека.** SPP + USB. Опубликована на JitPack. Async |
| 4 | **ObdGraphs** | [tzebrowski/ObdGraphs](https://github.com/tzebrowski/ObdGraphs) | Kotlin | Tomek Żebrowski | Апр 2025 | 53 | ✅ | ✅ | ✅ | Графики в реальном времени. Проф. дизайн |
| 5 | **CarScanApp** | [midnightyoff/CarScanApp](https://github.com/midnightyoff/CarScanApp) | Kotlin | midnightyoff | Окт 2025 | 2 | ✅ | ✅ | ✅ | **Свежий.** Compose, DTC read/clear, maintenance log |
| 6 | **Elm327** | [takyonxxx/Elm327](https://github.com/takyonxxx/Elm327) | Java | Türkay Biliyor | Окт 2023 | 48 | ✅ | ✅ | ✅ | В Play Store. WiFi + BT. 22 форка |
| 7 | **CarBusInterface** | [theksmith/CarBusInterface](https://github.com/theksmith/CarBusInterface) | Java | Kristoffer Smith | Янв 2016 | 291 | ✅ | ✅ | ✅ | Старый, боевой. 77 форков. Raw command interface |
| 8 | **BT OBD-II Diag Tool** | [fussek/Bluetooth-OBD-II-Diagnostic-Tool](https://github.com/fussek/Bluetooth-OBD-II-Diagnostic-Tool) | Java | Sebastian Fussek | Сент 2022 | 21 | ✅ | ✅ | ✅ | Bachelor thesis. Хорошая документация + расчёты PID |
| 9 | **obd-scanner-android** | [ETSoftwareStudio/obd-scanner-android](https://github.com/ETSoftwareStudio/obd-scanner-android) | Kotlin | ETSoftware | Апр 2025 | 2 | ✅ | ✅ | ✓ | Jetpack Compose + Hilt. Образец архитектуры |
| 10 | **KWP Logger** | [bri3d/kwp-android-logger](https://github.com/bri3d/kwp-android-logger) | Java | Brian Ledbetter | Фев 2016 | 28 | ✅ | ✅ | ✓ | **KWP2000 для VW/Audi.** Специалист по протоколам |
### Что между ними общего (подтверждено всеми 14 проектами)
**Bluetooth SPP:**
```java
UUID SPP_UUID = UUID.fromString("00001101-0000-1000-8000-00805F9B34FB");
BluetoothSocket socket = device.createRfcommSocketToServiceRecord(SPP_UUID);
socket.connect();
```
**Инициализация ELM327 (единый стандарт):**
```
ATZ → сброс
ATE0 → эхо выкл
ATL0 → line feeds выкл
ATS0 → пробелы выкл (или ATS1 — с пробелами)
ATH0/1 → заголовки CAN
ATSP0 → авто-протокол
```
**Отправка команд (единый стандарт):**
```java
outputStream.write((cmd + "\r").getBytes());
outputStream.flush();
```
**Чтение ответов — два лагеря:**
- **Лагерь 1 (AndrOBD, CarBusInterface):** побайтово, по 1мс
- **Лагерь 2 (AndroidOBD, Elm327):** `BufferedReader.readLine()`
### Кого изучать в первую очередь для Elmer
| Для чего | Проект | Почему |
|----------|--------|--------|
| **BT ↔ ELM слой** | AndrOBD + CarBusInterface | 10 лет отладки, все краевые случаи |
| **Kotlin-интеграция** | AndroidOBD + CarScanApp | Чистый API, современный код |
| **PID/DTC парсинг** | AndrOBD | Все формулы, все режимы |
| **KWP2000 (VW)** | KWP Logger | Если Phaeton на старом протоколе |
---
## Часть 3: СЫРОЙ КОД AndrOBD — золотой стандарт (1993 ⭐, 10 лет в продакшене)
### 3.1. Bluetooth SPP подключение (BtCommService.java)
```java
final UUID SPP_UUID = UUID.fromString("00001101-0000-1000-8000-00805F9B34FB");
// Стандартный socket
if (secure) {
tmp = device.createRfcommSocketToServiceRecord(SPP_UUID);
} else {
tmp = device.createInsecureRfcommSocketToServiceRecord(SPP_UUID);
}
// FALLBACK: reflection RFCOMM channel 1 (для китайских клонов!)
catch (IOException e) {
Class<?> clazz = mmSocket.getRemoteDevice().getClass();
Method m = clazz.getMethod("createRfcommSocket", Integer.TYPE);
sockFallback = (BluetoothSocket) m.invoke(mmSocket.getRemoteDevice(), 1);
mmSocket = sockFallback;
mmSocket.connect(); // пробуем снова
}
```
### 3.2. 🚨 КРИТИЧЕСКИ: 500мс пауза после BT connect
```java
// AndrOBD issue #233 — без этой паузы Android теряет данные!
Thread.sleep(500); // CRITICAL: Fix for Android Bluetooth timing
// Только после этого запускаем worker thread
mBtWorkerThread = new BtWorkerThread(socket, socketType);
mBtWorkerThread.start();
```
### 3.3. Инициализация ELM327 (точный порядок)
```java
private void initialize() {
setStatus(STAT.INITIALIZING);
// 1. Кастомные init-команды (если есть)
cmdQueue.addAll(customInitCommands);
// 2. Установить протокол (AUTO или конкретный)
pushCommand(CMD.SETPROT, preferredProtocol.ordinal()); // ATSP<n>
// 3. Инициализировать адаптивный тайминг
mAdaptiveTiming.initialize(); // ATAT1
// 4. Ускорить протокол (убрать пробелы и переводы строк)
pushCommand(CMD.SETSPACES, 0); // ATS0
pushCommand(CMD.SETLINEFEED, 0); // ATL0
// 5. Выключить эхо
pushCommand(CMD.ECHO, 0); // ATE0
}
// → ждёт "ELM" ответ → затем 0100 (запрос поддерживаемых PID)
```
### 3.4. Адаптивный таймаут (ГЕНИАЛЬНО)
```java
// Константы
ELM_TIMEOUT_MAX = 1000 // мс — максимум
ELM_TIMEOUT_DEFAULT = 200 // мс — стартовое значение
ELM_TIMEOUT_RES = 4 // мс — шаг изменения
ELM_TIMEOUT_MIN = 12 // мс — минимум
// На каждый NODATA → увеличить таймаут на 4мс
// На каждый успех → уменьшить таймаут на 4мс (но не ниже learned_min)
void adapt(boolean increaseTimeout) {
if (increaseTimeout) {
if (elmMsgTimeout + 4 < 1000)
setElmMsgTimeout(elmMsgTimeout + 4);
} else {
if (elmMsgTimeout - 4 >= learnedMin)
setElmMsgTimeout(elmMsgTimeout - 4);
}
}
// Каждое изменение отправляет: ATST<timeout/4>
```
**Диапазон: 12мс → 1000мс, шаг 4мс. Старт: 200мс.**
### 3.5. Обработка ответов (handleTelegram)
```java
// 1. Фильтр эха
if (lastTxMsg.equalsIgnoreCase(bufferStr))
return 0; // проигнорировать
// 2. Определить тип ответа
switch (getResponseId(bufferStr)) {
case PROMPT: // ">" — конец ответа
case MODEL: // "ELM" → вызвать initialize()
case SEARCHING: // идёт поиск
case NODATA: // нет данных
case ERROR: // ошибка
case NOCONN: // "UNABLE"
case BUSERROR: // ошибка шины
// ... ещё 10+
}
```
### 3.6. Обработка ошибок (production-grade)
```
BUS ERROR / UNABLE / CAN ERROR:
→ DISCONNECTED
→ переставить в очередь последнюю команду
→ сбросить протокол (ATSP)
→ переинициализировать тайминг
→ закрыть протокол (ATPC)
DATA ERROR / RX ERROR / BUFFER FULL:
→ WARM START (ATWS) — мягкий перезапуск
NO DATA:
→ увеличить адаптивный таймаут
→ переустановить протокол
STOPPED:
→ переставить последнюю команду в очередь
```
### 3.7. Мульти-фрейм (ISO-TP для VIN и длинных ответов)
```java
// Формат: "014" = 1 строка, 14 hex символов
// "0:4902015756..." — первая строка с длиной
// "1:5A5A314B5A41..." — продолжение
if (buffer[0] == '0' && buffer.length == 3) {
charsExpected = Integer.valueOf(bufferStr, 16) * 2; // 0x14 = 20 байт
lastRxMsg = "";
responsePending = true;
}
if (bufferStr.indexOf(':') >= 0) {
lastRxMsg += bufferStr.substring(idx + 1);
}
if (lastRxMsg.length() >= charsExpected) {
result = handleDataMessage(lastRxMsg); // готово
}
```
### 3.8. Итоговый процесс подключения (хронология)
```
1. BT socket connect()
2. Thread.sleep(500) ← КРИТИЧЕСКИ! Без этого — потеря данных.
3. BtWorkerThread.start()
4. StreamHandler.run() — побайтовое чтение
5. ATSP0 → ATAT1 → ATS0 → ATL0 → ATE0
6. Ждём "ELM" → статус INITIALIZED
7. 0100 → опрос поддерживаемых PID
8. ECU detect → выбор ECU → готов к работе
```
doc/legacy/research/mpscholar-automotive-sensing-actuators.md
+396
View File
@@ -0,0 +1,396 @@
# Automotive Sensing and Actuators
> Источник: [MPScholar — Monolithic Power Systems](https://www.monolithicpower.com/en/learning/mpscholar/automotive-electronics/automotive-sensing-and-actuators)
> Дата сохранения: 2026-06-10
---
## Содержание
1. [Introduction to Automotive Sensors and Actuators](#1-introduction-to-automotive-sensors-and-actuators)
2. [Types and Functions of Sensors in Automotive Systems](#2-types-and-functions-of-sensors-in-automotive-systems)
3. [Types and Functions of Actuators in Automotive Systems](#3-types-and-functions-of-actuators-in-automotive-systems)
4. [Power Management for Sensors and Actuators](#4-power-management-for-sensors-and-actuators)
5. [Integration and Interfacing of Sensors and Actuators](#5-integration-and-interfacing-of-sensors-and-actuators)
---
## 1. Introduction to Automotive Sensors and Actuators
### The Role of Sensors and Actuators in Modern Vehicles
A new era of unheard-of performance, safety, and control in automobiles has begun with the introduction of sensors and actuators in automotive engineering. The future of mobility can be understood by comprehending the complex functions that these devices play, especially at a time when we are on the verge of a revolution in transportation.
#### Overview of Vehicle Automation and Control
The 21st-century automobile is changing from a mechanical device to an extremely complex electrical system on wheels. This change has been made possible in large part by the growing integration of actuators and sensors, which work together to enhance vehicle functioning.
- **Role of Sensors:** In essence, sensors are the eyes and ears of a vehicle. They keep an eye on a number of variables, including proximity, temperature, acceleration, and speed. Numerous control systems rely on this data to provide them with real-time information about the vehicle and its surroundings.
- **Role of Actuators:** If sensors are the information gatherers, actuators are the doers. Actuators receive signals and respond with specified actions, including changing the air-fuel ratio in the engine, tightening up the suspension, or even applying the brakes. They convert electrical information into mechanical action, directly influencing and controlling a variety of vehicle components.
#### Improving Safety, Efficiency, and Performance
The ultimate goal of sensor and actuator integration is to improve driving in three critical areas: performance, efficiency, and safety.
- **Safety Enhancements:** In order to provide power to advanced driver-assistance systems (ADAS), sensors such as radar, lidar, and cameras collaborate with one another. Meticulous sensor input and actuator reaction enable features like automated emergency braking, adaptive cruise control, and lane-keeping assistance. Through anticipatory threat detection and proactive measures, these technologies significantly lower accident rates and save lives.
- **Efficiency Optimization:** In today's automotive world, fuel economy and pollution management are critical. Onboard computers can modify combustion settings due to sensors that track pollutants and engine data. Actuators then put these adjustments into practice, maximizing fuel efficiency and lowering dangerous emissions. In a similar vein, sensors aid in the best possible battery utilization in electric cars, guaranteeing optimal range and longevity.
- **Performance Upgrades:** Today's drivers need a car that is strong, nimble, and responsive. Sensors evaluate performance metrics like grip, acceleration, and aerodynamic drag through continuous feedback loops. Actuators then modify components such as the suspension, engine, and gearbox to improve the vehicle's performance and provide for a thrilling ride.
To sum up, the integration of actuators and sensors in contemporary automobiles has completely reshaped the concepts of automotive engineering. These elements will become even more crucial as we approach the future of autonomous driving and smart transportation, spurring innovation and setting new standards for performance, safety, and efficiency.
### Basic Principles of Sensing and Actuation
The two main pillars that support the current vehicle control system are actuation and sensing. The sophisticated and sensitive behavior of today's cars, which allows them to easily interact with constantly changing environments, depends on both of these components.
#### Sensing as Information Gathering
In the context of automobiles, sensing can be conceptualized as the means by which the vehicle perceives its internal states and external environment. Similar to how our senses of sight, touch, and hearing feed us vital information about the world around us, automobile sensors pick up on particular factors that affect how well vehicles operate.
- **Types of Sensors:** Sensors vary widely based on their functional requirement. Common varieties include position sensors (for crankshaft or throttle position), temperature sensors (for engine and interior conditions), pressure sensors (in tire monitoring systems or fuel lines), and more sophisticated devices (such as cameras and radars for ADAS functions).
- **Data Acquisition:** Every sensor operates on the principle of converting a physical quantity into an electrical signal. Electronic control units (ECUs) interpret and analyze these electrical impulses, making real-time analysis possible. For this reason, this conversion is essential.
- **Feedback Mechanism:** Continuous data collection guarantees that a feedback loop is maintained at all times, which in turn supplies the control systems of the vehicle with the most recent information. This ongoing cycle enables the behavior of the vehicle to be improved and adjusted.
#### Actuation as Control Execution
Actuation takes over to make the required adjustments after the sensors have collected the crucial data. Actuators essentially function as the vehicle's "muscles," translating the electrical impulses that are processed back into motion.
- **Types of Actuators:** Actuators in vehicles are diverse, including components like fuel injectors (which control fuel delivery), electric motors (steering, braking, or throttle control), and solenoids (for valve operation or gear shifts).
- **Signal Interpretation:** The ECUs of the car send signals to the actuators, which decipher the sensor data. These signals specify the precise action that the actuator must do in order to accomplish the intended result.
- **Responsive and Adaptive Actions:** Vehicles that use actuators can be made to be both responsive and adaptable. When an obstruction is detected, responsive actions take rapid action, such as automated braking. Adaptive actions, like adaptive cruise control, which modifies vehicle speed based on traffic circumstances, change over time based on continuous sensor data.
In conclusion, the modern vehicle's intelligence is defined by the combination of sensing and actuation. Actuators implement the necessary modifications to maximize safety, performance, and efficiency, while sensors offer a thorough understanding of the surroundings and the condition of the vehicle.
### Historical Development of Automotive Sensors and Actuators
When one looks at the realm of sensing and actuation, the evolution of the automobile is a fascinating tapestry of engineering achievements and discoveries.
#### Evolution of Sensing Technologies in Vehicles
The early autos' basic mechanical and electro-mechanical systems are where sensing in cars first appeared.
- **Mechanical Era:** The nascent stages of automotive development predominantly employed mechanical systems. An example of an early speedometer was a cable-driven device that sent speed through a rotating cable and was directly attached to the gearbox.
- **Electro-Mechanical Onset:** Transitioning into the 20th century, electro-mechanical components began surfacing. For instance, bimetallic strips and Bourdon tubes were utilized in temperature and oil pressure gauges, respectively.
- **Electronic Revolution:** Thanks to developments in semiconductor technologies, electronic sensing saw a boom after the 1970s. The advent of sensors such as oxygen sensors, manifold absolute pressure sensors, and throttle position sensors during this era laid the foundation for advanced engine management and electronic fuel injection systems.
- **Advent of ADAS and Connectivity:** Advanced driver-assistance systems (ADAS) were introduced in the late 20th and early 21st centuries. Advances in autonomous driving, collision avoidance, and lane departure warning systems were made possible by technological innovations, including radar, LIDAR, and cameras.
#### Trends and Future Directions
The scope of sensing and actuation in the automobile industry is expanding in step with the constant advancement of technology.
- **Miniaturization and Integration:** Miniaturization is a trend in modern sensors, making them smaller without compromising on functionality. Integrated sensor systems are increasingly widely used; they combine several sensing functions into a single unit.
- **Self-Diagnostics and Predictive Maintenance:** The upcoming generation of sensors and actuators are not only operational devices but also self-aware. They are able to keep an eye on their performance, anticipate malfunctions, and notify the driver or the car's central system of possible problems.
- **Holistic Vehicle Sensing:** An automobile that senses its environment holistically is the automotive industry's vision of the future. To ensure peak performance, safety, and comfort, a confluence of internal and external sensors must cooperate.
- **Actuators in Electric and Autonomous Vehicles:** With electric cars (EVs) gaining pace, specialized actuators customized for EVs are on the horizon. Actuators will also become increasingly important as autonomous driving technologies advance, guaranteeing precise, split-second responses to sensor input.
- **Material Innovations:** Actuators can now respond faster, with greater precision, and for longer periods of time thanks to new materials including shape-memory alloys and piezoelectric compounds.
---
## 2. Types and Functions of Sensors in Automotive Systems
### Classification of Automotive Sensors
Automotive sensors are essential to the smooth operation of modern automobiles. These sensors provide information about numerous vehicle parameters to the Electronic Control Unit (ECU) so that safety, efficiency, and performance are maximized. They do this by translating physical quantities into electrical impulses. These sensors can be categorized along two main lines: first, by the physical characteristics they measure, and second, by the underlying technology they use.
#### Classification Based on Physical Properties
- **Pressure Sensors:** These devices identify and gauge the pressure of the car's various fluids, including air, fuel, and oil. They make sure that the pressures stay within predetermined limits for ideal functioning and are frequently utilized in fuel injection and brake systems. They are predicated either on differential pressure sensing or absolute pressure sensing theory.
- **Temperature Sensors:** Integral to engine management, temperature sensors monitor the engine's coolant, oil, and air temperatures. By doing this, possible harm is avoided and the engine is guaranteed to run within a safe temperature range. Furthermore, temperature sensors are integrated into all power electronic controllers so that, in the event that the temperature rises above safe limits, the power can be derated or switched off.
- **Position Sensors:** These sensors determine where different parts are located. Examples are the Camshaft/Crankshaft Position Sensors, which help with engine timing, and the Throttle Position Sensor (TPS), which senses the position of the throttle in internal combustion engines.
- **Speed Sensors:** These sensors detect the rotational speed of the wheels and axis and are frequently used in the Anti-lock Braking System (ABS) and Transmission Control Units (TCU). This information enables the ECU, for example, to make real-time changes to prevent wheel lockup while braking.
- **Level Sensors:** These sensors keep an eye on the fluid levels in a variety of reservoirs, such as engine oil sump pumps, braking fluid reservoirs, and gasoline tanks.
#### Classification Based on Technology
- **Capacitive Sensors:** When a physical quantity varies, they work on the basis of capacitance alteration. In capacitive proximity sensors, for example, an object's approach modifies the capacitance, which is then detected. Certain fluid-level sensors rely on the fluid's capacitance.
- **Ultrasonic Sensors:** These sensors produce ultrasonic waves and are mostly utilized in parking assistance and obstacle detection. The sensor measures the distance by measuring the time it takes for the waves to reflect back after hitting an obstruction and receiving the information.
- **Infrared Sensors:** These sensors use the infrared spectrum to detect obstacles and provide night vision, particularly in low-light situations.
- **Piezoelectric Sensors:** These sensors produce a voltage in response to mechanical stress. Engine knock sensors use this feature to identify engine knock or pinging.
- **Hall-Effect Sensors:** Operating on the principle of the Hall Effect, these sensors can detect magnetic fields and are commonly employed for position detection, notably in the context of camshaft and crankshaft positions.
- **Resistive Sensors:** These sensors, such as temperature sensors, whose resistance varies inversely with temperature, alter their resistance in response to the physical quantity they detect.
### Applications of Sensors in Automotive Systems
#### Engine Management and Control
The engine management system's core components are the sensors, they enable peak performance, fuel economy, and emission control:
- **Fuel/Air Mixture Control:** By measuring the amount of oxygen in exhaust gasses through the use of oxygen sensors installed inside the exhaust system, the engine control module is able to modify the fuel-air mixture for the best possible combustion.
- **Ignition Timing:** Crankshaft and camshaft position sensors help establish the engine's phase and speed. This information helps the engine control unit (ECU) to time the spark for combustion exactly.
- **Cooling System:** Temperature sensors monitor the engine's coolant temperature. If the temperature crosses a defined threshold, the ECU can modify the functioning of the cooling fan or communicate a potential overheating issue to the driver.
- **Turbocharger Control:** Pressure sensors are used in turbocharged engines to monitor the boost pressure and ensure that it remains within the safe operating parameters established for the engine.
#### Safety Systems
Safety is fundamental in vehicle design, and sensors play a critical part in numerous safety-enhancing systems:
- **Airbag Deployment:** Accelerometers detect fast deceleration characteristics of a collision. The sensor alerts the airbag control unit to activate the airbags, which cushion the occupants and lower the possibility of injury in the event of a large accident.
- **Anti-Lock Braking System (ABS):** Wheel speed sensors constantly track the rotational speed of each wheel in the anti-lock braking system (ABS). The ABS adjusts brake pressure to prevent wheel lockup when it senses it is about to happen, preserving steering control.
- **Traction Control System:** This system detects when one or more wheels lose grip by using wheel speed sensors. In order to regain traction, the ECU can then lower engine power or apply brake force to particular wheels.
- **Collision Sensors:** These are particularly crucial for battery electric vehicles (BEVs), as they ensure that all high-voltage parts are deactivated in the event of a collision. This is accomplished via the collision sensor circuit, which modifies the crash signal state that high-voltage components expect in the case of a crash and ensures that any circuits that may have become accessible to persons due to the collision and vehicle damage are de-energized.
#### Driver-Assistance Systems
- **Parking Assistance:** This is provided by ultrasonic sensors installed all around the car to identify nearby obstructions. By giving the driver input regarding the distance to objects, these sensors help make parking in confined places easier to handle.
- **Lane-Keeping Assistance:** Roadside lane markers are detected by optical or infrared sensors. Depending on how sophisticated the system is, it may alert the driver or even take corrective action if it detects an inadvertent lane departure without signaling.
- **Adaptive Cruise Control:** This technology keeps a safe following distance between itself and the car in front of you using radar or LIDAR sensors. The mechanism automatically lowers speed to preserve the predetermined gap if the car in front of it slows down.
- **Blind Spot Detection:** This system lowers the likelihood of side-swiping accidents by alerting drivers to cars in their blind spots, usually through the use of radar or ultrasonic sensors.
### Key Specifications and Performance Criteria
#### Accuracy and Resolution
- **Accuracy:** This indicates the degree to which the sensor's reading agrees with the real value. A temperature sensor that is precise to within 0.5°C of the real temperature, for example, is more reliable than one that could be 2°C off.
- **Resolution:** The smallest change in the quantity being measured that causes the related output signal to alter noticeably is referred to as this. For example, a pressure sensor is said to have 0.01 psi resolution if it can measure variations as small as 0.01 psi.
#### Sensitivity and Range
- **Sensitivity:** This is defined as the sensor's response, or change in output, to a change in the input or amount being measured.
- **Range:** The physical quantity that the sensor is capable of measuring is shown, along with its minimum and maximum values.
#### Environmental Considerations
- **Temperature Stability:** Because cars operate in a variety of conditions, sensors need to be able to function accurately and consistently across a wide temperature range.
- **Resistance to Contaminants:** To ensure lifetime and reliable operation, automotive sensors should be resistant to fuel, oil, dust, moisture, and other contaminants.
- **Vibration Resistance:** Cars can cause a lot of vibrations and shock, especially in rough terrain. For constant readings, sensors must be unaffected by these vibrations.
#### Type of Sensor Errors
- **Offset Error:** An ongoing inaccuracy injected into the sensor data.
- **Gain Error:** Errors proportionate to the input signal are called gain errors.
- **Drift Error:** Errors that gradually change over time.
- **Random Error:** Typically indicative of noise in the sensor circuit, random errors lack a clear pattern.
- **Quantization Error:** This kind of error is caused by the sensor's restricted resolution.
#### Fault Diagnostics
Modern car systems have built-in self-diagnostic features to keep an eye on the condition and performance of their sensors.
- **5V Output Sensors:** Sensors with a 5-volt output voltage range frequently use the lower voltage band (below 0.5V) and upper voltage band (above 4.5V) to indicate a fault.
- **Digital Temperature Sensors:** High safety-rated temperature sensors frequently display a false, implausible temperature value, such as -200°C, to signify that a chip internal problem has occurred.
#### Detection of Faults
- **Redundancy:** Making use of several sensors to make a single measurement.
- **Self-Test Mechanisms:** Modern sensors are equipped with self-test functions.
- **Plausibility Checks:** Comparing sensor outputs to established physical models to make sure they are consistent.
---
## 3. Types and Functions of Actuators in Automotive Systems
### Classification of Automotive Actuators
In automotive systems, actuators operate as a conduit between the physical actions occurring inside a car and the control systems. They convert incoming energy into motion in order to carry out commands.
#### Classification Based on Control Action
**Linear Actuators**
- **Description:** These actuators produce linear motion, usually in the form of push or pull actions.
- **Application:** An example of an application is the operation of the brake master cylinder, in which the hydraulic fluid is pushed through the system by the actuator to engage the brake pads.
**Rotary Actuators**
- **Description:** These produce rotational motion, which is usually expressed in terms of angles or whole revolutions.
- **Application:** An example of an application is the fuel injection system's throttle plate adjustment, where the actuator spins the plate to regulate airflow. The liquid-cooled systems pressure pump serves as an additional illustration.
#### Classification Based on Technology
**Electric Motors**
- **Description:** Produce motion by means of electrical energy. Their working principle is based on electromagnetic principles, in which motion is produced by a magnetic field created by current flowing through a coil.
- **Application:** One example of such application is electric power steering systems, which, in response to driver input, use motors to help in steering.
**Solenoids**
- **Description:** These are electromagnetic devices that, when powered on, create a regulated magnetic field. Subsequently, a plunger or rod experiences linear motion due to the magnetic field.
- **Application:** An example of an application is transmission shift control, in which a solenoid engages or disengages gears in response to commands from the driver or computer.
**Piezoelectric Actuators**
- **Description:** Use the piezoelectric effect. When mechanical stress is applied, some materials generate an electric charge. In contrast, these materials undergo a shape-changing process that results in mechanical motion when voltage is given to them.
- **Application:** Fuel injector systems in some sophisticated engines. Because of their high-frequency response, piezoelectric actuators can provide injections that are extremely rapid and precise.
### Applications with Actuators in Automotive Systems
#### Throttle Control
- **Role of Actuators:** The throttle actuator controls how much air enters the engine. In the past, this operation was mainly mechanical. On the other hand, "drive-by-wire" or electronic throttle control (ETC) systems are used in modern systems.
- **How It Works:** Rather than physically pulling a cable, depressing the gas pedal in an ETC system delivers an electrical signal. This signal is interpreted by an actuator at the throttle body, which then modifies the throttle plate to control engine airflow.
#### Transmission Shift Control
- **Role of Actuators:** In both automated and manual transmission systems, transmission actuators help with gear shifting.
- **How It Works:** Solenoid actuators in contemporary automatic transmissions decode electrical signals from the transmission control module. By regulating the hydraulic fluid flow to various transmission tunnels, these solenoids regulate which gear set is in operation.
#### Active Suspension Systems
- **Role of Actuators:** Active suspensions are cutting-edge devices that instantly adjust to changing road conditions and driving demands to improve handling dynamics and ride comfort.
- **How It Works:** The system uses a mix of actuators and sensors to identify cornering forces, vehicle speed, and road defects. Actuators quickly change the ride height or damper stiffness. They are typically electromagnetic or electro-hydraulic.
### Key Specifications and Performance Criteria
#### Force and Torque Capabilities
- **Definition:** Two essential indicators of an actuator's performance are force and torque. Torque, which is typically linked with rotary actuators, represents rotational force, whereas force is a push or pull action that is linear in nature.
- **Measurement:** Generally, torque is expressed in Newton-meters (Nm) or foot-pounds (ft-lb), while force is expressed in Newton's (N) or pounds-force (lbf).
#### Speed and Response Time
- **Definition:** Response time is the amount of time an actuator takes to begin moving after receiving a command, whereas speed is the fastest an actuator may move to reach its desired location.
- **Measurement:** For linear motions, speed can be stated in mm/sec, while for rotating actuators, it can be given in RPM. Milliseconds (ms) are commonly used to indicate response time.
#### Reliability and Durability
- **Definition:** Durability is the number of operational cycles an actuator can withstand before wearing out or malfunctioning, whereas reliability is the capacity to perform consistently over time without failure.
- **Measurement:** While durability may be described in terms of operating cycles or hours of operation under specific conditions, reliability is frequently measured using metrics like Mean Time Between Failures (MTBF).
---
## 4. Power Management for Sensors and Actuators
### Power Requirements for Sensors and Actuators
#### Operating Voltage and Current Ranges
- **Definition:** Specific voltage and current ranges are intended for the operation of each sensor and actuator.
- **Importance:** Staying within these parameters guarantees that the sensor or actuator operates as intended without running the risk of damage or malfunction.
- **Measurement:** Common operating voltages for automotive applications may be between 5V and 24V.
#### Power Consumption and Efficiency
- **Definition:** Power consumption measures the total amount of energy that a sensor or actuator uses over time. Efficiency quantifies how well a device transforms the power it consumes into useful output.
- **Importance:** Energy is a limited resource in automobiles, particularly in electric or hybrid versions.
- **Factors Affecting Consumption and Efficiency:** The device's design, the materials utilized, the working environment, and operation frequency.
- **Measurement:** For smaller devices, power consumption is commonly expressed in milliwatts (mW) or watts (W). Efficiency is the ratio of usable power output to total power input, stated as a percentage.
### Power Optimization Strategies
#### Power-Saving Modes for Sensors
- **Sleep Mode:** In sleep mode, the sensor uses very little power and is largely inactive. It can become "awakened" when its purpose is required.
- **Idle Mode:** The sensor keeps working but at a reduced capacity, ready to go back to full operation when needed.
- **Interrupt-Driven Mode:** Until an external trigger or interrupt activates the sensor, it stays in low-power mode.
#### Always-Awake Sensors in Vehicles
- **Theft-Detection Sensors:** They keep a close eye out for any indications of tampering or illegal access.
- **Key Fob Detection Sensors:** These sensors are always on the lookout for signals from the key fob in cars with keyless entry systems.
#### Energy Efficient Actuation Techniques
- **Adaptive Control:** The actuator modifies its actions in response to immediate feedback.
- *Variable Displacement Pumps:* Modify the fluid flow rate in accordance with the system's present requirements.
- *Dynamic Brake Energy Recovery:* The energy generated during braking is recovered and transformed back into useful electrical energy.
- *Electric Motors with Load Sensing:* The motor can adjust its power output according to the required torque.
- **Pulse-Width Modulation (PWM):** Enables more precise control over the amount of energy utilized by altering the width of the electrical pulse delivered to the actuator.
- **Optimized Drive Circuits:** Energy efficiency can be achieved in the design of the electronic circuits that drive actuators.
- **Variable Load Sensing:** Certain sophisticated actuators have the ability to detect the load they are experiencing and modify their energy usage accordingly.
---
## 5. Integration and Interfacing of Sensors and Actuators
### Sensor and Actuator Interfaces
#### Analog vs. Digital Interfaces
**Analog Interfaces**
- **Nature:** Use a continuous signal that fluctuates in frequency or amplitude to transmit data.
- **Pros:** They offer a clear representation of a measured or controlled quantity and can be easy to use and reasonably priced.
- **Cons:** Limited range and susceptibility to noise interference. The connecting ECU must supply a distinct sensor ground specifically for that sensor.
- **Usage:** Commonly seen in simple sensors like pressure or temperature sensors.
**Digital Interfaces**
- **Nature:** Discrete signals, mostly binary (0s and 1s), are used to transmit data.
- **Pros:** They provide accurate and strong noise immunity.
- **Cons:** Their cost is usually higher than that of analog sensors. They require additional computational power from the DSPs and microcontroller interface.
- **Usage:** Common in contemporary automobile systems where accurate control and data collection are essential.
#### Communication Protocols for Sensors
**Inter-Integrated Circuit (I²C)**
- A packet-switched, single-ended, multi-master, multi-slave serial communication protocol. Frequently used to connect slower peripheral integrated circuits (ICs) to microcontrollers and processors.
- **Example:** Ambient light sensors in cars.
**Single-Edge Nibble Transmission (SENT)**
- A point-to-point protocol that allows sensor readings to be sent from a controller to a sensor. Designed with low power consumption and the fewest possible sensor connection pins.
- **Example:** Throttle position sensors.
**One-Wire**
- This protocol just needs one wire to communicate. Intended for low-speed data transmission.
- **Example:** Tire pressure monitoring sensors.
**Serial Peripheral Interface (SPI)**
- A synchronous serial communication protocol that selects the target device using a select line in addition to distinct clock and data lines.
- **Example:** High-speed gyroscopic sensors used in advanced stability control systems.
**Controller Area Network (CAN)**
- A common protocol for higher-level vehicle communications. Reliable, able to function in noisy settings, and appropriate for real-time applications.
- **Example:** Wheel speed sensors for ABS and traction control.
**Local Interconnect Network (LIN)**
- For non-critical sub-networks inside a car, a more affordable option to CAN.
- **Example:** Rain or light-detecting modules.
### Integration Challenges and Solutions
#### Ensuring Compatibility Between Components
**Challenge:** The variety of sensors and actuators that may originate from different manufacturers, different eras of technology, or different design paradigms.
**Solutions:**
- **Standardization:** Using standardized interfaces, voltages, and communication protocols (SAE, ISO standards).
- **ISO:** ISO 14229, ISO 15765 (vehicular communication), ISO 26262 (functional safety).
- **SAE:** SAE J1979 (OBD systems), SAE J1939 (heavy-duty communication).
- **Interfacing Modules:** Use interface modules or gateways that can translate between different protocols.
- **Unified Development Platforms:** Develop and test on the same platform or environment.
- **Comprehensive Documentation:** Keep detailed documentation for every component.
### Procedures for Sensors and Actuators
#### Sensor Calibration
- **Procedure:** Recording the sensor's reaction after subjecting it to a variety of known situations. The output is modified to match the anticipated values.
- **Example:** A temperature sensor may be subjected to a range of exact temperatures while modifications are made to guarantee that its output corresponds to the input values that are known.
#### Actuator Calibration
- **Procedure:** Change the control signal that is supplied to the actuator, measure its reaction, and make adjustments as needed to get the desired result.
- **Example:** To make sure a solenoid delivers the appropriate force or displacement for each level, it may be driven at different current levels. The correlation between current and displacement can be used as an integrated look-up table in the DSP or Microcontroller of the ECU.
---
*Сохранено с MPScholar (Monolithic Power Systems) — Automotive Electronics / Automotive Sensing and Actuators*
doc/legacy/research/research.md
+89
View File
@@ -0,0 +1,89 @@
# Заметки и находки
## 2026-05-25 — Исследование конкурентов
- Найдено 4 конкурента: DiagnostiX AI, OBDAI, Zero Touch, MUCAR MUAI
- Русскоязычных нет, RuStore пустой
- MUCAR уже использует DeepSeek → подтверждение правильности выбора LLM
- Никто не делает открытый клиент (кроме Zero Touch, но там Flutter/Gemini)
## 2026-05-25 — Подтверждение гипотезы
DeepSeek дал полный подробный анализ по логам VCDS с тестового проезда.
Лучше любого гугла. Гипотеза подтверждена.
## 2026-05-25 — Домен
- Куплен **obdai.ru** (обыгрывается: «обдай грязью» + OBD + AI)
- Зона `.ai` дорогая ($60-80/год), не брали
- Сервер: `https://obdai.ru/api/v1/raw-obd`
- HTTPS: Let's Encrypt (обязательно для RuStore/Google Play)
## 2026-05-25 — AndrOBD (fr3ts0n)
- **Разработчик:** Erwin Scheuch-Heilig (fr3ts0n), Австрия, `erwin.scheuch-heilig@gmx.at`
- **Репозиторий:** `github.com/fr3ts0n/AndrOBD`
- **Лицензия:** GPLv2
- **Год начала:** 2015
- **Архитектура:** библиотека (`library/`) + приложение (`androbd/`) + плагины
### Ключевые файлы для форка
| Файл | Что делает |
|---|---|
| `library/.../ElmProt.java` | Протокол ELM327: AT-команды, парсинг ответов, DTC, PID, VIN, адаптивные таймауты |
| `library/.../ObdProt.java` | OBD2-сервисы, декодирование PID, битовые маски, NRC-коды |
| `androbd/.../BtCommService.java` | Bluetooth SPP (классический, не BLE) |
| `androbd/.../BleCommService.java` | Bluetooth BLE (запасной вариант) |
| `androbd/.../NetworkCommService.java` | WiFi OBD адаптеры |
### Что НЕ берём
- `androbd/` — весь UI: графики, дашборды, CSV-экспорт, плагины (MQTT, GPS, сенсоры)
- `fastlane/` — метаданные для магазинов
### Стратегия форка
- **НЕ копипаст.** Используем Git submodule на AndrOBD
- Подключаем только `library/` модуль
- Наш код — отдельный `app/` модуль с тонким UI
- Лицензия GPLv2 на весь клиент (совместимо с открытостью)
- AndrOBD протестирован с 2015 года — экономим недели отладки краевых случаев
### Решение
Библиотека AndrOBD (GPLv2) покроет 100% ELM327-клоунов и их глюки.
Наш тонкий UI + HTTP-forward — пишем сами.
Перед RuStore — форкнуть обязательно. Для MVP (одна машина) — опционально, наш `elm.py` справится.
- Chrome Android: Web Bluetooth API — только BLE. ELM327 использует Bluetooth Classic SPP → НЕСОВМЕСТИМЫ
- Web Serial API — не поддерживается на Android вообще
- Termux + Python + pyserial — теоретически возможно, но Bluetooth-доступ в Termux сложен
- Вывод: нативное Android-приложение обязательно
## 2026-05-25 — ELM327 v1.5 (PIC18F25K80)
- Китайский клон, не оригинальный чип PIC18F2480
- Протокол неполный, возможны глюки на高速 CAN
- Держать в уме при тестировании
## 2026-05-25 — RuStore
- Бесплатная регистрация разработчика
- Модерация 1-3 дня
- Проверяет: вредоносный код, подозрительные permissions
- Не проверяет: скрытые закладки в легальном API
- Наши permissions (BLUETOOTH + INTERNET) — минимальны, вопросов не вызовут
## 2026-05-25 — Sideload (установка APK напрямую)
- Проверок нет совсем
- Android показывает список permissions перед установкой
- Пользователь видит только BLUETOOTH + INTERNET → доверие
## 2026-05-25 — Целевая аудитория
- Технически любопытный автовладелец с ELM327
- Не профессионал, но и не «глубинарий» (это шутка)
- Хочет понять проблему, а не просто получить код
- Требования к ответам: честная уверенность, пояснения, предупреждения