MMWave Radar HA Card(WIP)

Multi-model millimeter-wave radar calibration & live visualization card for Home Assistant.

---

Supported Models

Model	Freq	Targets	Z axis	Breathing	Heart rate	Sleep
MicRadar R60ABD1	60 GHz	1	✅	✅	✅	✅
Hi-Link LD2450	24 GHz	3	❌	❌	❌	❌

Adding a new model requires only creating one file — see Adding a New Model.

---

Features

Panel	Description
① Geometry & Boundary	Installation coordinates (X/Y/height), orientation (yaw/pitch/roll), optional room boundary polygon
② Yaw Calibration	Two-point geometric method — no IMU required
③ Live View	Real-time top-down map, time-faded trail, boundary filtering, multi-target support

• Full 3-axis orientation: yaw · pitch · roll (for non-horizontal installations)
• Room boundary polygon: targets outside are shown with a dashed indicator and excluded from the trail
• Multi-target: LD2450 shows up to 3 targets simultaneously with index labels
• Per-model calibration saved separately in localStorage
• Visual config editor with model-aware entity picker

---

Screenshots

Tab ① — Geometry & Boundary

Tab ② — Yaw Calibration

Tab ③ — Live View

---

Installation

Via HACS (recommended)

Or: HACS → Frontend → ⋮ → Custom repositories → add this repo URL → category Lovelace.

Manual

1. Download mmwave-card.js from the latest release.
2. Copy to <config>/www/mmwave-card.js.
3. Settings → Dashboards → ⋮ → Resources → add /local/mmwave-card.js (type: module).

---

Configuration

R60ABD1

type: custom:mmwave-card
radar_model: r60abd1
presence_entity: binary_sensor.r60abd1_presence
x_entity: sensor.r60abd1_x
y_entity: sensor.r60abd1_y
z_entity: sensor.r60abd1_z        # optional
breath_entity: sensor.r60abd1_breath  # optional
heart_entity:  sensor.r60abd1_heart   # optional
sleep_entity:  sensor.r60abd1_sleep   # optional
room_w: 400   # room width (cm)
room_d: 350   # room depth (cm)

LD2450

type: custom:mmwave-card
radar_model: ld2450
presence_entity: binary_sensor.ld2450_presence
target1_x: sensor.ld2450_target_1_x
target1_y: sensor.ld2450_target_1_y
target2_x: sensor.ld2450_target_2_x  # optional
target2_y: sensor.ld2450_target_2_y
target3_x: sensor.ld2450_target_3_x  # optional
target3_y: sensor.ld2450_target_3_y
room_w: 500
room_d: 400

All options

Option	Type	Default	Description
radar_model	string	required	Model ID: r60abd1, ld2450, …
room_w	number	400	Room width for canvas scaling (cm)
room_d	number	350	Room depth for canvas scaling (cm)
presence_entity	string	required	Binary sensor for presence
(model entities)	string	varies	See model tables above

---

Adding a New Model

1. Create src/models/<your_model>/index.ts and implement RadarModelAdapter:

import type { RadarModelAdapter } from "../base";

export const myRadarAdapter: RadarModelAdapter = {
  info: {
    id: "my_radar",
    displayName: "My Radar XYZ",
    fovDegrees: 90,
    maxRangeM: 5,
    updateRateHz: 10,
    maxTargets: 1,
    hasZAxis: false,
    hasBreathing: false,
    hasHeartRate: false,
    hasSleep: false,
  },

  getEntitySchema: () => [
    { key: "presence_entity", labelKey: "editor.presence_entity", required: true,  domain: "binary_sensor" },
    { key: "x_entity",        labelKey: "editor.x_entity",        required: true,  domain: "sensor" },
    { key: "y_entity",        labelKey: "editor.y_entity",        required: true,  domain: "sensor" },
  ],

  validateConfig(config) {
    return this.getEntitySchema()
      .filter(f => f.required && !config[f.key])
      .map(f => `Missing: ${f.key}`);
  },

  readFromHass(hass, config) {
    const pres = hass.states[config.presence_entity as string];
    if (!pres || pres.state !== "on") return { present: false, targets: [] };
    const x = parseFloat(hass.states[config.x_entity as string]?.state) || 0;
    const y = parseFloat(hass.states[config.y_entity as string]?.state) || 0;
    return { present: true, targets: [{ index: 0, rawX: x, rawY: y, rawZ: 0 }] };
  },

  getDefaultCalibration() {
    return { ...DEFAULT_CALIBRATION, radar_height: 200 };
  },
};

2. Register it in src/models/index.ts:

import { myRadarAdapter } from "./my_radar";

export const RADAR_MODELS: Record<string, RadarModelAdapter> = {
  r60abd1: r60abd1Adapter,
  ld2450:  ld2450Adapter,
  my_radar: myRadarAdapter,   // ← add here
};

3. Build: npm run build

That's all. The editor model picker, all three panels, and the calibration storage update automatically.

---

Building from Source

git clone https://github.com/zomco/mmwave-card.git
cd mmwave-card
npm install
npm run build      # → dist/mmwave-card.js
npm start          # watch mode

Requires Node.js ≥ 18.

---

Related

• ESPHome R60ABD1 Component

---

License

MIT © zomco