SimuLith Gen-1

Behavioral AIoT Virtual Room Simulator

Simulating realistic smart-room behavior through contextual human activity, adaptive devices, environmental interaction, and imperfect IoT sensors.

SimuLith Gen-1 is a behavioral AIoT simulation system designed to generate realistic smart-room datasets using virtual human behavior, environmental physics, adaptive devices, and contextual events.

Instead of producing random sensor values, SimuLith simulates:

• human routines
• environmental changes
• device interaction
• energy consumption
• sensor imperfections
• contextual behavioral patterns

The result is a rich, explainable, and reusable dataset suitable for:

• AIoT research
• smart-home simulation
• anomaly detection
• occupancy prediction
• behavioral analysis
• energy forecasting
• edge AI experimentation

◆ ◆ ◆

Main Idea

◆ ◆ ◆

Simulation Architecture

Contextual behavioral simulation pipeline inside SimuLith Gen-1.

◆ ◆ ◆

What SimuLith Simulates

Human Behavior

The virtual human is not random.

Behavior changes dynamically based on:

• time
• weather
• stress
• fatigue
• focus
• comfort
• financial pressure
• special days
• events

Simulated Activities

• Sleeping
• Gaming
• Side Job
• Cooking
• Watching TV
• Relaxing
• Outside Working
• Morning Routine

---

Dynamic Weather System

Weather affects:

• occupancy
• gaming probability
• room brightness
• comfort
• humidity
• AC usage

Weather Types

• Sunny
• Cloudy
• Light Rain
• Heavy Rain

---

Contextual Events

SimuLith includes contextual virtual-life events.

Events

• Weekend Gaming
• Late Night Work
• Laundry Day

Special Days

• Ramadan
• Lebaran
• New Year

These events influence:

• activity patterns
• occupancy
• energy usage
• sleep behavior
• device interaction

---

Adaptive Device Logic

Devices react to:

• indoor conditions
• human comfort
• activities
• weather
• financial behavior

Simulated Devices

• AC
• PC
• Main Lamp
• Desk Lamp
• TV
• Rice Cooker
• Washing Machine

---

Environmental Physics

The environment engine simulates:

• indoor temperature
• humidity
• brightness
• heat transfer
• AC cooling effects
• PC heat generation

---

Energy System

SimuLith calculates:

• total power consumption
• current
• voltage
• cumulative kWh

This enables:

• energy analytics
• smart-grid experiments
• AI forecasting
• occupancy estimation

---

Realistic Sensor Layer

Sensors are intentionally imperfect.

Features

• noise
• drift
• false positives
• imperfect readings

Simulated Sensors

• temperature sensor
• humidity sensor
• PIR sensor
• lux sensor

This creates more realistic AI training data.

◆ ◆ ◆

Dataset Output

SimuLith exports contextual CSV datasets.

Example columns:

activity
fatigue
stress
focus
comfort
inside_temp
humidity
power_consumption
occupancy
weather_type
special_day
sensor_values

The dataset is designed to preserve:

• temporal continuity
• behavioral causality
• realistic transitions
• environmental interaction

---

Why This Project Exists

Many public IoT datasets:

• are static
• have little context
• contain disconnected sensor values
• lack behavioral realism

SimuLith explores a different idea:

What if we simulate life itself first,
then generate the sensors from that world?

◆ ◆ ◆

Tech Stack

Current

• Python
• Pandas
• CSV-based simulation pipeline

Planned

• MQTT streaming
• Live dashboard
• Multi-room simulation
• AI prediction layer
• Firebase / MongoDB integration
• Realtime visualization

◆ ◆ ◆

Project Structure

simulith-gen1/
│
├── core/
│   ├── world_state.py
│   └── tick_engine.py
│
├── engines/
│   ├── time_engine.py
│   ├── weather_engine.py
│   ├── event_engine.py
│   ├── human_engine.py
│   ├── device_engine.py
│   ├── environment_engine.py
│   ├── energy_engine.py
│   ├── sensor_engine.py
│   └── export_engine.py
│
├── simulation/
│   └── simulated_data.csv
│
├── main.py
└── README.md

◆ ◆ ◆

Running SimuLith

Install dependencies

pip install pandas

Run simulation

python main.py

Generated CSV output:

simulation/simulated_data.csv

◆ ◆ ◆

Future Vision

SimuLith is planned to evolve into:

• realtime AIoT digital twin
• smart boarding-house ecosystem simulator
• MQTT-based streaming platform
• behavioral smart-environment dataset generator
• AI experimentation environment

---

Research Possibilities

Possible AI use-cases:

• occupancy prediction
• stress estimation from energy usage
• anomaly detection
• energy forecasting
• adaptive automation
• edge-AI behavior modeling

◆ ◆ ◆

Status

SimuLith Gen-1
Behavioral Core: COMPLETED

Current focus:

• realism refinement
• ecosystem expansion
• realtime architecture
• AI integration

◆ ◆ ◆

Developer Notes

SimuLith was built as an exploration of:

"How realistic can a virtual smart-room become
before AI even touches the real world?"

The project focuses heavily on:

• explainability
• realism
• interconnected systems
• behavioral causality
• contextual AIoT datasets