In this exploration, students will determine the field of view of the AI Vision Sensor by testing how far and how wide the AI Vision can detect barrels and AprilTags. 

Ensure students have the materials needed in order to mark the field of view. Distribute the Field of View – Barrels task card  (Google / .docx / .pdf) first. You can use paper and pens, a field and masking tape or wet-erase markers, chart paper and markers as shown in the video, or whatever combination works best for you and your students. As they complete this exploration, they will need to measure the angle of their field of view and will need a protractor. 

As students share their angular measurements, expect numbers at 73° with a range of 2 degrees in either direction. This value was recorded in a controlled environment with a white surface and bright lighting—your classroom may produce slightly different results. If groups fall outside of the expected range, encourage them to revisit their testing process to ensure they are following the procedures correctly.

Once students have completed their field of view with a barrel, distribute the Field of View – AprilTags task card (Google / .docx / .pdf). Groups should clearly distinguish between the two fields of view. For more ideas on how to implement this activity with the materials available to you, reach out in the PD+ Community.

Circulate around the room as students are completing this activity. Ask students questions such as:

• What do you notice about the detection of the object as you move it away from the robot?
• Does your data match your hypothesis? What are you discovering about the AI Vision's field of view?
• Is there a difference between the field of view when detecting a barrel versus an AprilTag? Why do you think that is?

In this exploration, students will determine how different lighting conditions affect AI Vision. 

Ensure students have the materials needed in order to modify lighting around the robot. Distribute the task card  (Google / .docx / .pdf) first. Lighting can be modified by changing the light source (classroom lights, window, flashlights), the color of the lighting (warm, bright, colored), the brightness (covering the robot to make it darker, adding more light), or other ideas. If students need help determining what factors to test, provide them with a smaller list of potential factors.

As students test each variable, they should record: the lighting factor being changed, the barrel’s position, whether it is detected, and whether AI Vision correctly identifies its color.

Circulate around the room as students are completing this exploration. Ask students questions such as:

• How does the lighting affect the detection of the barrel when it is close to the robot? When it's further away?
• What lighting variable impacted the detection the most? Why do you think it had such a large impact? 
• How does light affect vision? Do the factors that impact the AI Vision match or differ from our vision as humans?

In this exploration, students will investigate how different objects can be used to trick the AI Vision sensor. They will make hypotheses about what the AI Vision will report as being cargo, AprilTags, or another robot, then test those hypotheses and record their data.

Ensure students have the materials needed in order to trick the AI Vision. Distribute the task card (Google / .docx / .pdf) first. This exploration can use a wide variety of materials like the One Stick Controller, construction paper reconstructions of objects, printouts of elements like barrels, or other objects from around the classroom. Encourage students to first try for objects around the classroom that have the most likely chance of being detected before creating brand new objects.

Encourage students to use the AI Vision Utility for this exploration. As students test, they will be able to see the trick object while also seeing what is being reported by the AI Vision Sensor.

Circulate around the room as students are completing this exploration. Ask students questions such as:

• What patterns do you see about the objects being detected as sports balls and barrels? Why do you think that is?
• Is there one object that is more often detected than others? (i.e. sports ball, orange barrel, blue barrel)
• Is there a difference in your data between the trick objects when they are closer or further away from the robot? Why do you think that is?

In this exploration, students will investigate how the surface the robot is on affects the data from the AI Vision. They will form hypotheses about the effect (or lack of effect) of the surface on object detection, then test their ideas and record the results.

Ensure students have the materials needed in order to test different surfaces. Distribute the task card (Google / .docx / .pdf) first. This exploration can use a wide variety of options including: carpet (of different colors and thicknesses), linoleum tiles, stone flooring, wood flooring, table tops, mirrors, etc. Make sure students know where in the classroom they can go to find and test those surfaces.

Encourage students to use the AI Vision Utility during this exploration. As students test, they will be able to see the differing surfaces while also seeing what is being reported by the robot's AI Vision.

Circulate around the room as students are completing this exploration. Ask students questions such as:

• How do the surfaces you are testing differ? (color, size, reflectivity, etc).
• What patterns do you see in your data as you test the different surfaces? Why do you think that is?