- ✓A clear problem statement is the foundation of an IoT project plan; it defines what the application is intended to achieve and for whom.
- ✓Sensor selection requires balancing accuracy, power consumption, cost, and compatibility with the chosen microcontroller or platform.
- ✓Data flow mapping describes how data moves from sensors through gateways and processing layers to the application and user interface.
- ✓Connectivity options including Wi-Fi, Bluetooth, Zigbee, LoRaWAN, and cellular each have different range, power, and bandwidth characteristics suited to different scenarios.
- ✓A project plan for an IoT application should include a risk assessment that addresses connectivity failures, hardware faults, and security vulnerabilities.
Listen to the full episode inside the course. Enrol to access all 80 episodes, plus assignments, tutor support and Student Finance funding.
Start learning →Alex: Today we're looking at how to plan an IoT application. Sam, where does planning start?
Sam: It starts with a clearly defined problem. The most important question is: what physical phenomenon do you need to measure or control, why, and for whom? Without a clear answer to those questions, every subsequent decision is poorly grounded. An IoT application that monitors soil moisture for a small farm to automate irrigation has very different requirements from one that monitors the condition of industrial machinery to predict maintenance needs.
Alex: How do you turn that problem definition into a technical plan?
Sam: You work through several key questions systematically. What sensors are needed to measure the relevant physical quantities? What actuators are needed to take action based on those measurements? What are the data rates and volume: how frequently do you need to sample, and how much data will that generate? What connectivity options are available at the deployment location, and what are their range, power, and bandwidth trade-offs? How long must the device operate on a single charge, if battery-powered? And what processing needs to happen on the device versus in the cloud?
Alex: Let's talk about sensor selection. What are the key considerations?
Sam: Accuracy: does the sensor measure the physical quantity precisely enough for the application's needs? Resolution: can it detect the smallest change that matters? Operating range: does it work across the full range of conditions the device will encounter? Power consumption: for battery-powered devices, sensor power draw is a critical constraint. Interface compatibility: does it communicate using a protocol your microcontroller supports, such as I2C, SPI, or UART? And cost: for deployments at scale, sensor cost per unit is significant.
Alex: And connectivity options?
Sam: Wi-Fi provides high bandwidth and is familiar, but has limited range and relatively high power consumption, making it suitable for indoor applications near a power source. Bluetooth Low Energy is excellent for short-range communication with very low power consumption, ideal for wearables and home devices. LoRaWAN provides very long range, potentially kilometres, with very low power consumption, at the cost of very low bandwidth, making it ideal for agricultural and smart city applications where devices report infrequently. Cellular connectivity provides wide coverage but has higher cost and power consumption.
Alex: What should a formal IoT project plan include?
Sam: The plan should document the problem statement and objectives, a description of the system architecture including device selection, connectivity approach, and back-end platform, a data flow diagram showing how data moves from sensors through to the application, a security design covering device authentication, data encryption, and update mechanisms, a risk register covering technical risks, connectivity risks, and security risks, and a project timeline with clear milestones for hardware prototyping, software development, integration testing, and deployment.
Alex: Brilliant. Thanks Sam. Next we look at the development phase of an IoT application.