Table of Contents
Team Vision for Subsystem-Level Design PhaseSummarize: In the subsystem phase of the project the goal was to move from the system design, and determine the subsystems of the project and the feasability of particular solutions to each subsystem. The subsystems that were determined to be important in this phase of the project were:
- Air tank
- Jaw design
Additionally tasks to complete this phase with the subsystems that were determined were:
- Test water pump
- Finalize hydraulic system schematic
- Make Muscles
- Determine options for air tank
- Improve image processing
- Test additional dive sticks
Feasibility: Prototyping, Analysis, Simulation, Flow Charts, Schematics
Between now and the end of the semester, your team will continue to add detail to your design, which will require more detailed analysis and prototyping to answer feasibility questions. Consider:
- Iterative activities to demonstrate feasibility, including assumptions you made in your analyses, simulations, or prototypes.
- Have you completed sufficient preliminary analysis or prototyping to ensure that your design is likely to satisfy requirements?
- Have you included all usage scenarios in your modeling?
- Confirm that the selected concepts can deliver functionality defined by the System Architecture.
- Define the optimal values of the most sensitive design parameters.
- Support the evaluation of your team's concepts with quantitative information.
Body design was conducted with team P16229 in order to meet requirements for both team. Rear cross area is where tail is going to be attached and circular front will be used for jaw attachment. Curves on body is to allow fish to tilt forward incase of retrieving object that is on the floor. It is also accounted for side force that will be generated by tail's movement as well as necessary components it needs to hold. Once all the sensors and other components are finalized then area to secure other components will be created.
There are several qualities in our jaw design that were decided upon because it most benefited our fish as a whole. The jaw design will not allow water to flow into the body of the fish. This quality helps minimize weight, drag force, and water logging inner parts of the fish. The jaw mechanism also assists in an easy grab of the object, with a hall sensor detecting if the object has been obtained or not. The downside to this design is that it does not represent the structure of a fish.
Detecting the falling diving stick can be done using a simple webcam attached to a raspberry pi, and then processing of the frames can be done using opencv to determine the location of the object. A feasibility test was done with both pictures and video to prove that the code could detect a specified object and its location. When detecting the object the program will return the x and y position of the object, as well as the width and height of the object in pixels. It is recommended that to train the program to detect a specific object, 7000 images of the object should be used. In the case of the proof of concept test only six were used, which was still able to detect the object but also resulted in having quite a bit of noise. In further testing, more images will be used to see if it can properly clean up the detection of the object.
Determine Successful Retrieval
This Hall Sensor will be able to determine whether there is any object between each side of the mouth. This provides information regarding whether object was retrieved by the jaw successfully or not. If it recognizes that object is not caught then RoboFish will re-scan for the object until it retrieve the object successfully.
Determine Home Location
Determining Home location can be performed by knowing the distance between the "RoboFish" and "Home". As fish swims in certain direction, distance will change and simple algorithm can be used to swim towards the "Home" location. This algorithm will determine which direction allows distance to decrease and swim towards it until the distance reading become zero meter.
A proof of concept test was conducted to visually show the feasibility and accuracy of determining home using Bluetooth beacons. Using Android Studio, an application was written which provided location within the map using trilateration.
This provided a single beacon's RSSI(Received Signal Strength Indication) variation in different condition. Home location can be "determined" by RSSI which can be converted into distance data.
RSSI was monitored for 450ms then histogram was generated.
Simulation of approaching "Home" was conducted by moving cell phone towards the beacon in constant speed. It indicates that "Home" location can be approached by using RSSI of a single Bluetooth beacon within 5 meter radius.
Ballast Tank Schematic
This schematic shows the hydraulic and pneumatic systems used to operate the ballast tank and the actuate the McKibben muscle for the jaw.
Calculations for the ballast system are very complex and dependent on many different factors. This flowchart attempts to break down the steps needed to determine sinking and surfacing speeds, as well as the volume of the ballast tank.
This shows the theoretical flow rate from the pump as a function of pressure. When we order a pump we can replace this with actual data.
The following calculations show how the water flow rates for filling and emptying can be determined, and also the effects of limiting the size of the ballast tank.
P15029 Waterproof Pump Housing
The previous RoboFish team purchased a waterproof case to house their pump. The case is much larger than necessary and can potentially leak around where the tubing leaves the case.
3D Printed Case Concept
A mock-up was made of a waterproof case that could be 3D printed and would be designed to fit closely around the pump. Tank fittings could be used to provided waterproof outlets for tubing and electrical leads.
Unfortunately, this idea has been scrapped because the 3D printing process is not guaranteed to produce a waterproof surface. Without using an extremely high end printer or 100% infill, any printed object will contain tiny cracks that water may seep through. However, we do plan to use waterproof electrical fittings and tank fittings to create waterproof ports for the tubing and electrical leads.
There are very few companies that have mckibben muscles on the market today, the two different companies that I found on line are Images Scientific instruments and Festo. Those are all specs that I can find about their products and they are pretty expensive and too powerful for our project's objective so I looked up for DIY project to make our own mckibben muscles. They can come in many different size and strength so I picked out the size that would fit to our goal for our jaw. Our jaw only need at atleast 10 lbs force on the object and the parts that I picked the parts that would be able to deal with 120 psi of air/water pressure which should be able to meet to our goal.
We decided to use three way valve since we need a way to push the water out of the system. Those company that I found on line stated that the system is a electrical system that will switch the pathway of where water/air will flow to which is an ideal idea since we cannot physically touch the fish while it is running so the system will switch the pathway for us. Two different company that I compared on the table are MISOL and MEAD, their specs are relvently similar to each other but have a different size of inlet and outlet. The size of tube we are using was more than 1/10 inches diameter which means MEAD instruments would not be an ideal situtation for our project. I narrowed down to two different MISOL system and I chose the second one because of its size of inlet and outlet are much closer to the size of tube we are using right now.
Bill of Materials (BOM)
A bill of materials was constructed which began with the materials available to our team from the previous team and to this was added the components purchased by both tail and object retrieval team.
Updated assessment from Systems Design or link to other location.
Design Review MaterialsNot applicable
Plans for next phase
The RoboFish electrical team will be mainly focus on object detection and algorithm to control buoyancy. Object detection will be performed through image processing. Buoyancy testing will be conducted once ballast tank's construction completes and using depth sensor as a feedback source. Purchasing finalization will be made for electrical team for sensors.
The mechanical team will create jaw and ballast tank prototypes while creating body prototype and waterproofing system with team P16229. Once these prototypes are built, they will join electrical team to control buoyancy within required precision.