For this project we decided to investigate some of the sensory properties of the ultrasonic sensor. We went ahead and tested some of the various limits of the sensor. These include:
- Maximum detection range (Two different sized papers)
- Minimum detection range
- Detection of different materials (Glass pane, window screen, cloth, and tissue paper)
- Detection of angled surfaces
- Max peripheral vision.
We created a small Lego structure to support the ultrasonic sensor so the table under it wouldn't interfere with the data collection. After the data was collected in LabView we exported it to Excel so the graphs could be easily interpreted.
We used the Log Red Dataset command in LabVIEW in order to collect the data from the ultrasonic sensor during the different trials. Here is a pretty simple program set to read signals sent every 0.2 seconds for a total of 60 seconds so we could have enough time to do the trials.
This short video is a compilation of all of the trials we conducted using the ultrasonic sensor.
Interesting Analysis of Trials:
- Angled Detection: The ultrasound can sense things at an angle as long as it is less than 40 degrees. Once the angle surpasses that, the sound waves stop being reflected directly back to the sensor so the readings are off.
- Fabric Detection: The sensing capabilities for fabric, such as a jacket or coat, tend to be more innaccurate than other surfaces. This is most likely due to the fact that the fabric is not perfectly straight like paper or a wall.
- Minimum Distance: The ultrasound can reliably sense a minimum of 3 cm from the sensors. Anything closer than that causes the sensor to read 255 cm which is an error.
- Max Distance: The size of the object you are trying to detect heavily impacts the accuracy of the readings at far distances. While both the small paper and the large paper had inconsitencies in the data, the large paper had less and was still able to be detected up to 240 cm from the sensor while the smaller paper lost detection at around 155 cm.
- Tissue Detection: A strange discovery was that the ultrasonic is unable to detect tissue paper at all. At first we thought it was because of the thickness of the tissue but then we folded up to 4 times and it was still undetectable to the sensor. Therefore, it can be concluded that the material of the tissue is unable to reflect back the sound waves making it non-existant to the sensor.
- Window Pane Detection: While it may not be able to detect tissue cloth the ultrasonic sensor can accurately receive readings from clear glass. This is because the glass is solid enough to reflect back the sound waves.
- Window Screen Detection: The window screen, on the other hand, was undetectable because of the many tiny holes in the screen. Apparently the the sensor cannot detect sound waves sent at the screen because not enough are sent back to make a reading.
- Peripheral Vision (Sideways): From the center of the two eyes, the sensor can detect objects 15 degrees to either side accurately. It can then go up to 20 degrees with some inconsitencies. After 20 degrees however, it cannot detect objects.
- Peripheral Vision (Up/Down): From the center of the two eyes, the sensor is able to detect objects 10 degrees above if it or below it accurately. It can go up to 15 degrees but it's accuracy begins to decrease. After the 15 degrees the objects are not able to be detection.
From the trials it can be concluded that the sensor works relatively well at closer distances. Once the distances begin to excede 100 cm then the objects need to be quite large to be detected accurately. The type of material being detected also has a high effect on the accuracy. The best materials are flat structured objects such as paper, walls, and cardboard. Another thing is the peripheral vision up/down is quite lacking when compared to its sideways vision. All of these affect how the ultrasonic sensor can be used for autonomous navigation. The discrepancy in different material detection can be a major issue when you have an autonomous robot in the real world rather than a lab setting. It needs to accurately be able to detect a wide range of materials in order not to bump into them. Also the whole inability to detect objects at an angle greater than 40 degrees is also very dangerous because not everything an autonomous robot will encounter will be flat and right in front of it. The peripheral vision limits will result in the sensor having to be placed near the center of the front of the robot. The reason being that in that position it is able to detect obstacles near the floor but at the same time to ensure that there isn't anything coming from above that will hit the top of the robot. In the end, however the ultrasonic sensor is a very useful sensor for an autonomous robot if its limits are able to be upgraded.