What is the Capstone Project?
The Capstone Project is the culminating assessment of the IoT & Smart Systems course. It requires you to independently design, build, test, and document a complete, functional IoT system of your own choosing. Everything you have learned across Units 0–9 is applied here: hardware selection, circuit design, firmware development, communication protocols, cloud integration, security, and data analytics.
Capstone Learning Outcomes
By completing the Capstone Project, you will demonstrate the ability to:
- Apply hardware, programming, communication, cloud, security, and analytics skills in an integrated, independently designed IoT system.
- Make and justify architectural decisions across all layers of an IoT system.
- Test a complete system against defined requirements and iteratively improve it.
- Communicate a technical system design and its performance to a non-technical audience.
- Produce professional-quality documentation and code that another engineer could maintain.
Capstone Project — Mandatory Requirements
Every Capstone project MUST include all of the following elements. Projects that are missing any mandatory element cannot score above 60%.
Example Capstone Project Ideas
You may propose your own project or choose from the following. All projects must be approved by the instructor in Phase 1.
| Project | Sensors, Actuators, and Features | Sector |
|---|---|---|
| Smart Greenhouse Monitor | DHT22 + soil moisture + LDR. Fan relay + irrigation pump + grow light. Automated climate control. ThingsBoard dashboard. ML anomaly detection on temperature trends. | Agriculture, Rwanda rural IoT |
| Hospital Medicine Cold Chain | DS18B20 waterproof probes in refrigerators. Alert when > 8°C. Email alarm. Cloud logging with audit trail. mTLS authentication. Rwanda Law 058/2021 privacy compliance documented. | Healthcare, regulatory compliance |
| Smart Office Energy Monitor | Current sensor (SCT013) + occupancy PIR. Relay for HVAC and lighting. Energy kWh accumulation. Diurnal analysis dashboard. Predictive occupancy ML model. | Building management, energy savings |
| Water Quality Station | pH sensor + turbidity sensor + DS18B20. LoRa or Wi-Fi gateway. ThingsBoard multi-sensor dashboard. Alert on out-of-range pH. Weekly analytics report. | Water management, rural infrastructure |
| Smart Parking System | Ultrasonic sensors (HC-SR04) per bay. LED indicators (green/red). Cloud dashboard showing occupancy %. Count entering/leaving vehicles. Anomaly detection for stuck sensors. | Urban IoT, smart city |
| Air Quality Monitor | MQ-135 CO2/VOC + DHT22 + PM2.5 sensor. TFT display for local readout. MQTT to ThingsBoard. Alert when AQI exceeds threshold. Daily analytics report emailed automatically. | Environmental monitoring, schools |
| Smart Irrigation System | Soil moisture (capacitive) + DHT22 + rain sensor. Solenoid valve control. Event-driven irrigation with hysteresis. LoRa remote sensors. Weather API integration. | Precision agriculture, water conservation |
| Industrial Vibration Monitor | MPU-6050 accelerometer on pump motor. Anomaly detection for bearing wear. Predictive maintenance alerts before failure. REST API for maintenance log integration. | Industrial IoT, predictive maintenance |
Lesson 10.1 — Project Planning, Proposal, and System Design
The quality of your Capstone project is largely determined before you write a single line of code. A clear problem definition, well-chosen architecture, and documented design prevent costly rework later.
Writing a Strong Project Proposal
Your proposal is a one-page document that defines what you will build and why. It must answer four questions:
| Section | What to write |
|---|---|
| Problem statement | What real-world problem does your system solve? Who benefits? How does it currently work without IoT? Write 2–3 sentences describing the problem clearly. |
| Proposed solution | Describe your IoT system at a high level. Which sensors? Which actuators? Where does data go? What decisions does the system make automatically? |
| Success criteria | How will you know if your project succeeded? List 4–6 measurable criteria (e.g. ‘Temperature alert triggered within 30 seconds of threshold breach’, ‘Dashboard updates every 15 seconds’). |
| Risk assessment | List 3–5 risks (hardware fault, sensor accuracy, connectivity issues, time constraint) and your mitigation plan for each. Shows you have thought ahead. |
System Architecture Design
The architecture document is the engineering blueprint for your system. It must include all of the following:
1. Hardware Component List
| Component | Description | Connection / Purpose |
|---|---|---|
| Microcontroller | e.g. ESP32 DevKit V1 | Central processing, Wi-Fi, GPIO control |
| Sensor 1 | e.g. DHT22 temperature/humidity | GPIO 4, 1-Wire. Reads environment. |
| Sensor 2 | e.g. HC-SR04 ultrasonic | TRIG GPIO 9, ECHO GPIO 10. Measures distance. |
| Actuator 1 | e.g. 5 V relay module | GPIO 16 (active-LOW). Controls pump. |
| Power supply | e.g. 5 V 2 A USB adapter | Powers ESP32 and relay module. |
2. System Architecture Diagram
Draw a diagram showing how all components connect. Include: sensor connections (GPIO pins, protocols), actuator connections, ESP32 to cloud path (Wi-Fi → MQTT broker → ThingsBoard), cloud to dashboard, and any remote access (REST API, RPC). You may draw this by hand and photograph it, or use draw.io (free, browser-based).
3. State Machine or Decision Logic
Document the automation logic your system implements. For each automated action, specify:
- Trigger condition: which sensor value, threshold, and hysteresis gap
- Action: which actuator, what state (on/off/speed), for how long
- Override: how RPC commands interact with automatic control
- Safety rule: what happens if the sensor fails (NaN, out of range)
- Timeout: maximum duration for any actuator that could cause harm if stuck on
4. MQTT Topic Hierarchy
| Topic | Direction | Purpose |
|---|---|---|
| nedu/capstone/{project}/{device}/telemetry | ESP32 → Cloud | JSON telemetry payload (all sensor readings + actuator states) |
| nedu/capstone/{project}/{device}/attributes | ESP32 → Cloud | Static device info: firmware version, location, sensor types |
| nedu/capstone/{project}/{device}/rpc/req/+ | Cloud → ESP32 | RPC command subscription (control actuators) |
| nedu/capstone/{project}/{device}/status | ESP32 → Cloud | Last Will: ‘online’ on connect, ‘offline’ on LWT |
5. Security Design
- Authentication method: username/password or access token (stored in NVS)
- TLS: MQTT port 8883, CA certificate source
- ACL: list topic permissions for each device role
- Data retention: how long will telemetry be stored (ThingsBoard TTL setting)
- Privacy: which data fields are personal data under Rwanda Law 058/2021 and how they are protected
6. Analytics Plan
- Which ThingsBoard REST API endpoint you will use to retrieve data
- Preprocessing steps required (likely: resample to 5-min, fill gaps, remove outliers)
- Which features you will engineer and why
- Which EDA visualisations you will produce (time-series, diurnal, correlation)
- Whether you will implement anomaly detection, classification, or regression ML, and why
Design Review Checklist
Before proceeding to implementation, your design must pass this review. Check each item:
You may also like
Nyereka Tech and Tumba College Explore Opportunities for Collaboration
The Launch of our IoT Innovation Hub
