FuzzyBear Spring 2020
Object Detection Trade-Off Study
Author/s: Arnel Cunanan (PM & MST)
Table of Contents
Introduction
In this blog post we will be choosing how FuzzyBear will detect objects in its mission. Our goal is FuzzyBear to be able to detect a “bee” when it is within 40 mm in distance. We will look at the weight, cost, various conditions in which the object detection device can operate, and other research that contribute to our selection for object detection.
Choosing the Object Detection Device
The main types of object detection we will be looking at are camera module, inductive, capacitive, ultrasonic, TOF (Time-of-Flight), and RFID sensors. We will be looking at the cost, dimensions, operating voltage, and measuring distance of the specific devices. Below is a table of the possible object detection device that we might be able to use.
First, we will address specific requirements for our FuzzyBear robot. FuzzyBear is required to be able to detect an object within its path when it comes within 40 mm of the object. If the object detecting device fails to meet this requirement, we will have failed the functional requirement that determines our grade. Also, FuzzyBear has an operating voltage of 3.3 V internal battery and a, unregulated 5 V external battery. Our required cost for the device is $20 based on the allocated budget. Finally, the device should be able to fit in the head of our FuzzyBear skin. The maximum dimension that the device can be is 45 mm x 30 mm x 30 mm.
All the parts are well within our budget, but other factors will be looked at. Looking at the numbers we can already rule out the RC522 RFID Module and the LJ123-4/BY Inductive Sensor because the dimensions exceed that of our FuzzyBear head. Next, we will be looking at the operating voltage of all devices compared to what FuzzyBear can supply. Operating voltages for all components are in range of the project because we have an external battery supplying about 5V and a battery on the 3Dot supplying 3.3V. Lastly, we will look at the the measuring distance of the sensors. The only forms of object detection that will help us achieve our goal of at least 40 mm distance is the VL53L0X TOF Distance Sensor and the HC-SR04 Ultrasonic Sensor as they can measure distances of 30 mm to 1000 mm and 20 mm to 4000 mm, respectively.
Conclusion
The other factors that affected our decision on which object detection device to select are that the RC522 RFID Module’s Antenna can only read the corresponding RFID fob from about 3 cm away. This is not ideal because of the short distance and also we need to be able to detect at least 3 objects in our path. The OV7670 Camera Module can perform edge detection but can be affected by environmental factors such as lighting and object characteristics.
We were able to narrow down to HC-SR04 Ultrasonic sensor and VL53L0X TOF Distance Sensor.
Both are commonly used for distance measurements, which is ideal for detecting obstacles in FuzzyBear’s path, both have contactless detection, not affected by external environmental conditions, can be used in day and night. However, HC-SR04 Ultrasonic sensor becomes less accurate when the angle to an object increases. We decided to use Adafruit VL6180 TOF because it fits all the requirements for FuzzyBear. Some added notes are that it has its own unique address and won’t interfere with servo shield address (only uses two pins SDA+SCL). The two pins make it much simpler to use.
References/Resources
- https://cdn.sparkfun.com/datasheets/Sensors/Proximity/HCSR04.pdf
- https://cdn-learn.adafruit.com/downloads/pdf/adafruit-vl53l0x-micro-lidar-distance-sensor-breakout.pdf
- http://www.datasheetcafe.com/lj12a3-4-z-by-datasheet-pdf-sensor/
- http://web.mit.edu/6.111/www/f2016/tools/OV7670_2006.pdf
- https://www.hobbytronics.co.uk/datasheets/sensors/MFRC522.pdf
- https://www.seeedstudio.com/blog/2019/12/19/all-about-proximity-sensors-which-type-to-use/