By Max Strand, Imani Willis, Vinny Larson, Dylan Grizzell, Aaron Lopez Hernandez
Click on the link below for project details?
An endangered species is a plant or animal that is in serious danger of becoming extinct, meaning it could disappear from the Earth entirely. This happens when a species' population dwindles to a dangerously low level or when its habitat is severely threatened, making it difficult for the species to survive and reproduce.
E.g., Plants and animals such as the Arizona Hedgehog Cactus and the Giant Panda, respectively.
Population count mostly depends on user input, such as a human individually counting and marking each population. Current modern methods include drones that use either AI animal detection or human supervision for population count.
Drones are an effective way to quickly recognize a specific population at a given time without human interference. However, they usually have short battery life, disrupt the environment with noise, and are powered by lithium-ion batteries.
Currently, there are many downsides to using a drone to count populations. As mentioned, drones make too much noise, which disrupts the ecosystem. Another downside is the use of lithium-ion battery, which has many ethical concerns, as well as the drone's high power consumption with very little battery.
Shall be able to be controlled with a controller wirelessly.
Shall be able to create a virtual environment with a lidar.
Shall be able to recognize specific object's
Shall be able to cost less than a drone to manufacture.
Shall be able to transmit a video stream wirelessly.
May be able to generate real-time data of position and the number of objects detected.
May be able to withstand rain and sunny weather conditions.
May be able to be all purpose terrain with a suspension system.
May be able to drive autonomous without user input.
May be able to be powered with solar panels.
Mind maps are regularly used to create and develop ideas. In this case, many factors are in play, and it is important to consider every option.
Below is how we chose to develop and find a problem that could use improvement.
Having time as a constraint helps with overall motivation to finish the project. Some of the key dates for this project are:
Week 4: Prototype Presentation 1 (group presentation of progress and receive feedback)
Week 8: Prototype Presentation 2 (group presentation of progress and receive feedback)
Week 10: Exposition (present to faculty, companies, & students)
The concept of the design was to create a small-medium size vehicle that used three wheels. The back two turn forward/backward, and the back ones turn either left or right. Also, the original design included a suspension system for the vehicle only for the back wheels, which would allow the vehicle to navigate all terrains.
This Arduino handles 5V therefore used only to connect the wireless controller logic to the stepper motor.
This Arduino can handle higher voltages and is used to connect with the MKR IMU,
Camera sensor is used to collect data by counting the amount of pixels and color. This data is sent to an AI model trough Wi-Fi.
This is a small lidar used to collect data points of distances to create a virtual environment of the space.
This shield includes an accelormeter that is used to calibrate the camera and lidar to make accurate maps while vehicle is in motion.
A medium-sized stepper motor is used to turn all wheels. Two back wheels turn forward and backward. The front only does turning.
A medium-sized servo motor is used to turn the lidar sensor 180° in the x-axis and 90°in the y-axis.
The prototype used a total of four base parts that connected to each other. In addition there was three wheel, one axle, two mounts, and two suspension system that were 3d printed for this project.
Very affordable
Biodegradable
Decreased risk of misprints
Shrink and deform under high exposure to UV light
Decomposes easily and short lifespan
More prone to cracking
Increase heat protection - UV resistant
Increased durability
Impact resistant
less prone to cracking
Increased risk of misprint failures
Not biodegradable, only recyclable
Slightly more expensive
For future innovation of the design, we plan to use PETG plastic due to its great benefits, such as handling more UV light for more extended periods of time. Besides, it won't shrink due to the amount of heat. Also, to manufacture it would be around the same cost as PLA plastic; therefore, there is really no downside to cost.
This image depicts a real test of the system gathering data and plotting it in a real room. We also printed out papers of real animal and placed them around the room to test the accuracy of the system. The system is successful if it plots the animal in the virtual map within reasonable distance.
This system uses the ESP32 to transmit a video stream wirelessly through Wi-Fi. The data of the pixels and frames are sent to an external device which runs AI model to identify object. This scours the internet to find similar animal which then is used to recognize the animal and label the animal with a box around it.
While assembling the first prototype, it had come to our attention that our system was not strong enough. We had to change the plastic stem and replace it with a steel one due to a 3d printing misprint. Currently, the design does not include a suspension system due to time constraints.
An annotated drawing of the physical prototype with key dimensions, with sensor and actuator locations.
An annotated drawing helps to share key information with other engineers for manufacturing and assembling purposes.
The physical model is a porotype that is used to test the functionality of the product and usually change key details before either another iteration or for mass production.
An isometric view of the prototype in CAD. This gives a general idea of the shape and function of the vehicle without much technical detail. These can usually be a single piece or an assembly of the product.