Skip to content
OpticWorks Intranet

hardwareOS Overview

hardwareOS is an open-source embedded platform for building connected sensor devices. It powers the OpticWorks RS-1, combining mmWave radar with camera vision for real-time occupancy tracking.

What is hardwareOS?

Section titled “What is hardwareOS?”

hardwareOS provides the complete software stack for the RS-1 device, emphasizing:

  • Low-latency streaming via WebRTC DataChannel
  • Sensor fusion combining radar and vision data
  • Over-the-air updates for field deployments
  • Spatial awareness through room calibration

The RS-1 Device

Section titled “The RS-1 Device”

The RS-1 is OpticWorks’ flagship sensor fusion device:

ComponentSpecification
SoCRockchip RV1106G3 (ARM Cortex-A7, 1.2GHz, 256MB RAM)
CameraSC3336 MIPI (3MP, LDCH correction, 4-lane CSI)
NPURKNN (1.0 TOPS for YOLOv8 inference)
RadarHLK-LD2450 (24GHz mmWave, 3 targets, 6m range)
ConnectivityWiFi 6, Bluetooth 5.0

Key Features

Section titled “Key Features”

Sensor Fusion

Section titled “Sensor Fusion”

Integrates 24GHz mmWave radar with camera vision to provide robust presence detection that works in any lighting condition.

Real-Time Performance

Section titled “Real-Time Performance”
  • 30Hz WorldState updates via WebRTC DataChannel
  • Sub-100ms end-to-end latency from detection to client
  • Hardware-accelerated inference on the RKNN NPU

Advanced Tracking

Section titled “Advanced Tracking”
  • Kalman filtering with 4-state tracking and velocity estimation
  • Hungarian algorithm for optimal detection-to-track matching
  • Multi-target support tracking up to 3 simultaneous targets

Room Calibration

Section titled “Room Calibration”

Integration with iPhone RoomPlan for easy spatial calibration and coordinate mapping.

Repository Structure

Section titled “Repository Structure”
hardwareOS/
├── products/rs1/
│   ├── fusion/           # Kalman filter, Hungarian algorithm
│   └── radar/            # LD2450 driver
├── targets/rv1106/
│   └── native/           # Native C code, gRPC integration
├── ui/                   # React frontend
├── docs/                 # Technical documentation
└── tools/                # Build and deployment utilities

Technology Stack

Section titled “Technology Stack”
LayerTechnology
LanguagesC (41.9%), TypeScript (34.4%), Go (20.7%)
Build SystemCMake, Makefiles
StreamingWebRTC (DataChannel)
InferenceRKNN, YOLOv8
CommunicationgRPC, Protocol Buffers

Data Flow

Section titled “Data Flow”
                    ┌─────────────┐
                    │  LD2450     │
                    │   Radar     │
                    └──────┬──────┘
                           │ Raw detections
                           ▼
                    ┌─────────────┐
                    │  Hungarian  │
                    │  Algorithm  │
                    └──────┬──────┘
                           │ Associated tracks
                           ▼
                    ┌─────────────┐
                    │   Kalman    │
                    │   Filter    │
                    └──────┬──────┘
                           │ Smoothed states
                           ▼
                    ┌─────────────┐
                    │ WorldState  │
                    │  (30Hz)     │
                    └──────┬──────┘
                           │ WebRTC DataChannel
                           ▼
                    ┌─────────────┐
                    │  Browser/   │
                    │    App      │
                    └─────────────┘

Who Works on This

Section titled “Who Works on This”
RoleResponsibility
Hardware EngineersBoard bring-up, driver development, sensor integration
Firmware EngineersCore C code, RTOS integration, performance optimization
Backend EngineersGo services, gRPC, WebRTC infrastructure
Frontend EngineersReact UI, visualization, user configuration

Getting Started

Section titled “Getting Started”
  1. Review the Architecture documentation
  2. Set up your Development Environment
  3. Understand the Deployment process

Resources

Section titled “Resources”