4.1.2.3 Lab – Design a Prototype of an AI Application Answers

4.1.2.3 Lab – Design a Prototype of an AI Application Answers

Objectives

  • Part 1: Consider an IoT Application with AI/ML Technology
  • Part 2: Design Components Required for an IoT Application with AI/ML Technology
  • Part 3: Describe the Process and Operation for the IoT Application with Flowcharts

Background / Scenario

With IoT devices and IoT networks getting popular, technologies for IoT devices are also improving rapidly. IoT devices are no longer limited to sensors and actuators; they may have the capability to think and act accordingly according to environment changes, thanks to the development of artificial intelligence (AI) and machine learning (ML). This lab is designed as a group project. The ideal group size is 3 to 4 people.

Required Resources

  • Device with internet access

Instructions

Part 1: Consider an IoT Application with AI/ML Technology

In Part 1, the group members will list the functions and features of a smart home thermostat and controller device with capability of self-learning, making decision based on the environment changes, and acting accordingly.

Step 1: List the desired features and function for such a device.

The list will vary. Possible items may include the following:
• Understanding commands given by humans, such as “Raise/lower temperature.”
• Automatic adjustment of the temperature of the house according to environmental changes.
• Communication with cloud computing for autonomous learning and improvement of algorithms.

Step 2: List the factors that may influence the perception of temperature.

Answers may vary. Possible items may include the following:

1. What is a comfortable temperature? Different people can react differently to a single set of temperatures.
2. Current room temperature
3. Humidity.
4. Proximity to air outlets
5. Time of day

Step 3: List the ways the smart device can get information about those factors.

Answers may vary. Possible items may include the following:

1. Internal clock to keep time/date
2. Thermostat and humidity sensor
3. Commands from humans
4. Network access for constant weather forecasts

Part 2: Design components required for the device with AI/ML Technology

In Part 2, the group will explore and design the functions of the key components required for the smart thermostat/controller.

Step 1: What are the key components for a smart thermostat/controller?

Answers will vary. The possible element must include the following:

1. Temperature and humidity sensor
2. Access to a cloud computing application through a Home Gateway
3. Microphone and speaker to interact with a human being
4. Microprocessor/controller to turn on/off the heater or air conditioning unit
5. Electrical connection and battery backup

Step 2: List the process and operation of the smart thermostat/controller?

Answers may vary. Possible elements should include the following:

1. Cloud computing relies on a model of normal temperature in a room through the analysis of data from a large sample set.
2. Initially the device will adjust the room temperature according to the model with respect to the historical data and the neighborhood region.
3. Initial factors may include the time of day and the current ambient temperature.
4. Room temperatures can be adjusted by receiving commands from humans.
5. The data is sent to the cloud computing application for continuous refinement of the model.

Part 3: Describe the Process and Operation for the IoT Application with a Flowchart

In Part 3, the group will use flowcharts to describe the logic flow for data collection, data analysis, human being interaction, and taking proper actions.

Step 1: Using a flowchart, describe how data is collected, communicated to the AI/ML application in the cloud computing, and processed.

Actual flowcharts will vary

Step 2: Using a flowchart to illustrate the general operation of the smart thermostat/controller

Actual flow chart will vary

Reflection

1. Which component provides the brain power for the learning and then adjusting accordingly?
The AI/ML model is in cloud computing.

2. Can you think of other IoT devices that will learn over time and improve their operations?
Answers may vary.


guest
0 Comments
Inline Feedbacks
View all comments