Team Vision for System-Level Design Phase
The plan for this phase was to elaborate upon the problem defined in phase 1 using a variety of tools, in order to address customer and engineering requirements, at the system level.
A functional decomposition was used to identify the material and information inputs and outputs into functional components starting at the whole system level moving down to the irreducible subsystem functions. Next, original benchmarking was revisited in order to identify potential solutions which could be utilized to provide the functions identified in the function decomposition. A system architecture block diagram was constructed to graphically show the interconnected of system functions and components.
Based on necessary functions and benchmarking, concepts were generated by subteams which could be used to perform functions required by the final design. In order to help select concepts which best perform the necessary functions, a morphological table was constructed. Using this table, concepts for system components were selected based on a large number of selection criteria formulated such that the selected design meets as many customer and engineering requirements as possible.
Selected concepts were tested via feasibility analysis wherein basic modeling or calculations were performed in order to assess as to whether or not the selected concepts were in fact feasible.
Functional Decomposition
Purpose
In order to define a complete list of functions, and establish how they are achieved (without referencing explicit solutions), a functional decomposition chart was created. This chart made it clear specifically where concepts will need to be generated in order to properly achieve engineering and customer requirements.Snapshot
A link to the live Functional Decomposition can be found here.
From top to bottom, functions and subfunctions are established. Nearing the bottom, the more specific and quantifiable functions of the design are indicated in their relation to the overall function in this flow diagram.
Inputs and Source
- Template and Example.
- Project Objectives.
- Engineering Requirements.
- Benchmarking Data.
Outputs and Destination
- Complete list of functions that must be provided in order to satisfy the functional requirements of the entire project.
- Engineering Requirements are written to measure how well the functions are performed.
- Concepts are generated around functions identified here.
Benchmarking
Purpose
In order to avoid redundant work, already available solutions and concept options were explored which could be utilized or exploited.
Class: Object Detection - Ultrasonic Ranger
Cost: $30
These lost-cost analog-output ultrasonic rangers can be made waterproof and used in underwater applications. They have great spatial resolution and should be able to detect obstacles easily. The analog voltage output can be connected to one of the IOs on the microcontroller which can convert the input signal to a range. The ranger will likely detect the walls of the pool and the setup will have to be programmed as such.
Pros: Low Cost, Easy-ish to use
Cons: Not waterproof
Maximum Range of 254 inches (645 cm or 12 feet) Operates from 2.5-5.5V Low 2.0mA average current requirement
Class: Object Detection - Laser Rangefinder
Cost: $125
These lasers can be used in pairs as rangers. They may be better than the ultrasonic rangers in terms of resolution and range, however, they are likely more expensive. The implementation of said rangers may also be more difficult.
Pros: Precise, Long Range
Cons: Expensive
Class: Enclosure - Waterproof
Enclosure
This company, Intex, sells a wide variety of enclosures which could be used for housing electrical components.
Pros: Meets standards for waterproof
Cons: Limited customizability, Wiring needs to somehow
leave the box
Class: Tail Material - Molding
Rubber
This rubber will allow us to custom design the back tail section and the caudal fin. The properties of this material are well characterized and should make it much easy to design a well characterized tail. Urethane rubber has an elastic modulus of between 2 and 10 MPa. Silicone rubber has an elastic modulus of between 1 and 5 MPa.
Pros: Cheap, Highly customizable, Easy to use
Class: Microcontroller -
Arduino Mega Microcontroller
Microcontroller would provide full control of the Robofish. 54 digital I/O allows for control of all solenoid valves while 16 analog inputs will read sensor outputs.
Pros: Easy to program (C or MATLAB)
Cons: Slower processing power
Class: Microcontroller - Raspberry
Pi 2 Model B Desktop (Quad Core CPU 900 MHz, 1 GB RAM,
Linux)
Raspberry Pi allow for Arduino alternative with more processing power.
Pros: Faster processing power, once coded more efficient
program execution
Con: LINUX required, Unfamiliarity
Current Robotic Fish Designs
Robotic Fish designs created by other research teams and companies were research in order to assess currently applied solutions, materials, and techniques used in created unmanned underwater vehicles (UUV). The following are three selections from this research in order of complexity.
The Robofish children's toy made by the company Zuru uses an electromagnet to control the oscillation of the tail fin. An alternating current is passed into the electromagnet with a coil surrounding it that moves the tail. The fish turns off automatically when it leaves the water due to a control circuit which is closed by a water bridge.
Key Observations:
- Tail Fin sole thrust generating member (thunniform)
- Automatically runs when placed in water
- Turns off automatically out of water
- Sinks to the bottom due to angle
- No sensor systems
This robotic fish is extremely maneuverable, however, it does not move fast. This fish is powered by servomotors and is broken into many sections lending to its maneuverability. The carangiform locomotion style is utilized by this design with nearly 50% of the body contributing to thrust. The walls of the tank in which it was tested where detected using cameras and image processing. Fish has powered pectoral fins for depth control.
Key Observations:
- Nearly 50% of the body is generating thrust (carangiform)
- Highly maneuverable (carangiform)
- Slow (carangiform)
- Powered pectoral fins for depth control (stability)
- Uses image processing for wall avoidance (sensor)
- Two cameras side angle mounted - corresponding avoidance algorithm based on camera activated (sensor)
A fairly fast and maneuverable fish, this Korean made robotic fish sports a strong outer shell constructed of plastic and silicone. The fish utilizes the subcarangiform method of locomotion and employs the back third of its body for generating thrust. The tail section is divided into two resulting in a three section fish. To turn, the fish employs tail motion in the form a C-turn. Passive yet rigid dorsal and pectoral fins help to minimize the lateral yaw of the fish. In order to detect obstancles and walls, the fish has a single centered front-mounted red laser.
Key Observations:
- 33% of the body generates thrust (subcarangiform)
- Balance of maneuverability and speed (subcarangiform)
- Rigid passive dorsal and pecoral fins for stability (stability control)
- Centered front-mounted red laser avoidance (sensor)
- Hard plastic "skin" connected with silicone at joints (asthetics)
Inputs and Source
- PRP.
- Project Objective.
- Function Decomposition.
- Prior MSD EDGE sites.
- Customers.
- Vendors and users of existing or similar solutions.
- Literature search.
- Friends and associates.
Output and Destination
- List of best available concepts and embodiments.
- Criteria used for selection of best.
Concept Development
Purpose
Investigate and research existing or new concepts used on previous iterations of Robofish and other projects.Electronic Systems Schematics
The past Robofish (P15029) electronic schematic design
file was updated to a more readable and applicable
format. Past
Robofish P15029 Schematic Update
Electronic System Concept
This proposed electric system diagram was created referencing the previous Robofish iteration. It also incorporates new ideas into the design. The gate drive circuitry has been kept intact as well as parts of the voltage regulation and depth sensor. New features include the consolidation of the digital circuitry into a Arduino Mega microcontroller, buck and boost converters to eliminated uneven battery discharging, and the LASER/sonar emission and detection system.
Tail Concepts
In order to avoid detected obstacles the designed solution must be maneuverable enough while still maintaining a reasonable forward locomotion speed. Using background research on fish locomotion methods and swimming styles, preliminary tail concepts were generated. Based on the the previously mentioned constraints (see Problem Definition section) the tail was designed to be between 30-40cm or 1-1.31ft
Customer and Engineering Requirements:
- CR17/ER03 - Body Length; Target: 2.5ft, Ideal: 1.75 ft
- CR01/CS01 - Tail Design and McKibben Muscles
- CR09/ER04 - Turning Radius; Target: 2.5ft, Ideal: 1.75 ft
- CR09/ER06 - Swim Speed; Target: 15 ft/min, Ideal: 20 ft/min
- CR13/ER10 - Weight; Target: 17 lbs(mass), Ideal: 12 lbs(mass)
In addition to tail concepts, caudal fin concepts were generated. These concepts were based off of examples of caudal fins found in nature. Designs were based off of fish which utilize the same locomotion methods and swimming styles used to generate the tail concepts.
Found through research, the aspect ratio of the caudal fin is a crucial parameter which defines different fins. For this reason concepts were ordered based upon their presumed aspect ratio as follows:
Ascending Aspect Ratio (lowest to highest):
- Design 14
- Design 12
- Design 13
- Design 11
- Design 9
- Design 10
Sensor Concepts
In order to autonomously avoid obstacles while swimming the designed solution must be able to detect obstacles far enough from the solution that the solution has enough time and space to avoid the obstacle. Concepts utilize different sensor types and positioning that may be used for the proposed solution.
Customer and Engineering Requirements:
- CR16/ER05 - Object Detection Distance; Target: 3.75 ft, Ideal: 10 ft
Input and Source
- Function Decomposition.
- Results of Benchmarking.
- Analysis and prototyping.
Outputs and Destination
List of multiple viable concept options to provide each function.Morphological Chart
Purpose
In order to develop concept options to meet the required list of functions a morphological chart was constructed. By doing this, it was determined whether or not all customer requirement and functions of the final design are achievable. This chart will be used in later steps to select an optimal set of concepts to meet customer and engineering requirements.Snapshot
A link to the live Morphological Chart can be found here.
Inputs and Source
- Project Objectives.
- Engineering Requirements.
- Function Decomposition.
- Concept Development.
- Benchmarking criteria.
- System Architecture.
Outputs and Destination
- A complete set of concepts that provide all of the functionality required.
- Fall back options to substitute for the most risky concepts.
Systems Architecture
Purpose
In order to investigate the proposed architecture of the system a high level system diagram was generated. This diagram graphical shows the flow of energy, info, material, and structural quantity flow. The system architecture leads directly into the design of subsystems and generation of concepts.High-Level System Architecture
The high-level systems diagram gives a overview of how
the electrical, mechanical, and other systems interact
with eachother. It is especially useful in determining
which components lie in the path of two independent
components.
Inputs and Source
- Engineering Requirements.
- Functional Decomposition.
Outputs and Destination
- High level description of the total system that support concept selection.
- Interface definition for subsystem design.
- Concept Development.
Concept Selection
Concept Selection - Design Solutions
Derived from the morphological chart, the team developed 4 leading concepts. Each design consisted of a different sensor system to which a seemingly logical array of concepts were chosen as this was deemed the most critical system component for completing the problem at hand.
- A - Directional Laser Ranger, Analog Filtering, Eyes Mount, Arduino, Subcarangiform, Tail-Based Steering, Ballast Tank, Torpediform, Forked, Carp-like, Direct McKibben
- B - Camera, Image Processing, Eye Mount, Arduino, Carangiform BCF, Tail-based Steering, Ballast Tank, Torpediform, Emarginated, Headstander-like, Direct McKibben
- C - Ultrasonic Ranger, Analog Filtering, Underwater Mount, Arduino, Thunniform BCF, Dorsal Fin, Ballast Tank, Laterally Flattened, Lunate, Tiger Fish-like, Spring
- D - Infrared, Analog Filtering, Eyes mount, Analog Processing & Delay Circuits, Subcarangiform BCF, Tail-Based Steering, Ballast tank, Torpediform, Forked, Carp-like, Fishing Line
Concept Screening - Selection Criteria
Seen below are the selection criteria broken down within regards to electrical, mechanical, and general needs, constraints, and requirements.
Electrical Based Criteria:
- Accuracy of identified object type
- Amount of falsely identified objects (%)
- Amount of correctly identified objects (%)
- Computing power/speed necessary
- Algorithm complexity
- Microprocessor cost
- Range of sensor
- Required Power (electrical)
- Robustness (electrical)
Mechanical Based Criteria:
- Swimming Speed
- Maneuverability
- Depth Control
- Body Stability (yaw)
- Required Power (pump pressure, valves)
- Versatility
- Robustness (mechanical)
- Power consumption efficiency (tail motion)
General Criteria:
- Size (aspect ratio)
- Weight
- Ease of maintenance
- Cost
Concept Screening - Pugh Analysis
Based on the aforementioned and the Pugh Analysis design solution A proved to be the best of the selected designs. There were many cases where the differences between these selections were rather trivial for certain characteristics and over the course of continued testing it may be that the final solution mixes and matches the best parts of the selected morphology. The most important features of design A are: maneuverability without compromising speed, obstacle detection range, energy efficiency, and versatility.
Selected Solution Summary
A high aspect ratio body comprised of three section, the first of which is the main body and the latter two which are the thrust generating sections (tail), which balances speed and maneuverability by utilizing subcaragiform locomotion. Thrust will be generated by McKibben Pneumatic Muscles, as per customer requirement, placed directly in the tail. Mounted on the head will be a laser used to detect obstacles. The design solution will read sensor data, implement evasive maneuvers to avoid obstacles, and navigate an obstacle course. This will all be accomplished autonomously, as per customer requirement.
Feasibility: Prototyping, Analysis, Simulation
Purpose
The feasibility section provides research and analysis efforts toward exploration of possible details which will be incorporated into Robofish. Proof of concept ideas are also discussed.MATLAB Swimming Analysis
A initial proof of concept was created to show a
suggested swimming algorithm and identify potential
issues. The analysis was performed in MATLAB assuming
inputs of swimming speed and overlap of valve timings.
The function outputs logic level commands which would
activate the valves.
Note that solenoid/valves 2 and 1 correspond to the tip
end of the tail while solenoid/values 3 and 4 correspond
to the section of the tail closer to the body.
Two sinusoidal signals form the basis of the timing and
overlap portion of the swimming analysis. One sinusoidal
signal controls a set of valves (2) and only the positive
portion of the sine wave is considered.
In order to implement the overlap of the valve actuation,
one of the sine waves was offset by a value between 0 and
1. This forces the sine wave to be moved a certain amount
above the normal configuration. Therefore make a larger
portion of the sine wave positive, corresponding to
longer valve on-time.
This valve timing configuration involves a speed of 1
swim stoke per second and 0% overlap.
This valve timing configuration involves a speed of 5
swim stokes per second and 0% overlap.
This valve timing configuration involves a speed of 5
swim stokes per second and 50% overlap. Note that the
transients/abnormalities in the solenoid signals should
be neglected as they result from the sine wave offset.
This will be addressed at the end of this section.
This valve timing configuration is most likely the ideal
case for the preferred method of swimming. The tip of the
tail remains fixed while the body-end of the tail
transitions. This valve timing configuration involves a
speed of 5 swim stokes per second and -100% overlap
(negative overlap refers to the other sine wave being
offset).
The completion of this analysis provided a look into how
the valves will be actuated and also provided glimpse
into how the algorithm will be implemented. Due to the
adjustable overlap functionality, using a sine wave
method may not be the best option. More likely,
implementing a strict timing analysis will be more
beneficial and easy to integrate on to a
micro-controller.
SolidWorks Tail Mock-Up
This model was generated from preliminary dimension discussions that were held during the initial concept generation and selection process, and is in no way exhaustively developed.
Depicted above is the front section of the tail, referred to as section A. The pillars seen in the hinges of the model were created in order to account for the requirements of the intended production method, 3D printing. The nature of the production method dictated the need to support so-called ‘bridges’, or portions of the component that would potentially sag during the printing process.
Depicted above is the back section of the tail, referred to as section B. The pillars seen in the hinge of the model were created in order to account for the requirements of the intended production method, 3D printing. The back portion of section B was designed to be inserted into a mold filled with a flexible medium which would set around and through it. The back portion of section B is intended as a rigid ‘spine’ that will help support the flexible portion of the tail.
Depicted above is the entirety of the tail ‘skeleton’ as it would be assembled, minus the fasteners and formed flexible medium which is to form the flexible portion of the tail.
The selected design for the tail skeleton assembled and in a bent configuration. This configuration shows how the team envisions the tail being positioned as it actuates through the water. In both images, it is assumed that the tail as a whole is moving through the water clockwise about the front end hinge of tail section A.
Underwater Laser Testing
Lasers Tested:
Staples Red Laser Pointer: 500 yards air distance 2.3 yards clear dot appearance underwater Wavelength: 650nm Power: <5mW
NcStar Green Laser: 740+ yards air distance 9 yards clear dot appearance underwater Wavelength: 532nm Power: 5mW
Test Results: Use of the recreation pool in the Gordon Field House limited measurements to opposing pool walls. Testing was done with all lights on simulating sunlight. Changes of distance in laser were observed when lights were all off. Scattering of laser light was not seen by naked eye so the use of a range finding laser would be recommended over photo-resistors if this option is chosen.
Battery Life Feasibility
In order to meet the customer requirement of a 10 hour run time a system of custom battery packs was devised. These battery packs would consist of NCR18650B Lithium Ion batteries arranged as follows:
A "cell" would be created which contains 4 NCR18650B Lithium Ion batteries in series. Any number of these could be used to create "packs" which support the required run time. Deemed to be most feasible is terms of meeting this requirement is the "9 - hour pack" comprised of 12 cells, weighing a total of 5.132 lbs(mass), providing approximately 501.12 Watt-hours of power.
Inputs and Source
- Engineering Requirements
- Concept Selection
Outputs and Destination
- A list of Design Parameters, Quantified Targets, and acceptable tolerances
- Sensitivity analysis
- Concept Selection
Future Testing
Testing Rig Tail and Muscle Characterization
A testing rig will be constructed which will be used to test the muscles and the fin and tail designs. The rig should consist of a force transducer on a rig which has been constrained initially to one dimensional motion. This will be used to characterize the tail and fin designs. Additionally, the force transducer can be used to characterize the muscles. Cameras will be necessary to capture displacement characteristics.Muscles Parameters to Measure:
- Lmin
- Lmax
- F
- dF/dt
- dx/dt
Other Necessary Parameters:
- Vfull
- Q of pump
- dF/dP
Fin Parameters to Measure:
- Aspect ratio
- Angular displacement
- Required bending force
Overall Parameters to Measure:
- F
- Yaw (x-directional force production)
Mathematical Models
Mathematical models will be generated for the tail which should help to characterize the thrust generating capabilities of the tail. Given a set of key assumptions and simplifications a rough estimate of the thrusting force will be determined which will be assumed to be an underestimate of the force which will actually be generated. This ensures that the actual fish should have more than enough thrust generating capability to meet the engineering and customer requirements.Designs and Flowcharts
This has not been started at this time.
Risk Assessment
Project Plan Revisited
REV as of 09/29/15
A link to the live document can be found here.
Plans for next phase
Looking forawrd to the next review the team would like to have a clearer picture of the functioning of the specific subsystems which drive the systems envisioned and selected in the Systems Design Level Phase (II). During the next phase specific questions such as the following will be addressed:
- What characteristics of the subsystem in question are critical for the overall success the system?
- How can we characterize the subsystem?
- How is the subsystem in question interconnected with other subsystems or systems?
- What are the possible failure modes at the subsystem level which would result in failure to meet the desired objective?
- What alternatives could be used to minimize risks and therefore maximize the probability of success?
On an individual level members of the team brainstormed how to help your team achieve this vision. In order to solidify these abstractions individual Three-Week Plans will be developed following the Systems Level Design Review in order to assist in achieving the aforementioned goals.
The team was previously divided into subteams based on background. An electrical team, comprised of Jon Nguyen, Corey Muench, and Ryan Selby, will focus their efforts on constructing and characterizing electrical components. A mechanical team, comprised of Nick Gulati, Matthew Yap, and Khalfan Alzaabi, will focus their efforts on constructing and characterizing structural and mechanical components. The subteams will collaborate on optimizing the swimming algorithm as this requires both mechanical and electrical inputs.
Home | Planning & Execution | Imagine RIT
Problem Definition | Systems Design | Subsystem Design | Preliminary Detailed Design | Detailed Design
Build & Test Prep | Subsystem Build & Test | Integrated System Build & Test | Integrated System Build & Test with Customer Demo | Customer Handoff & Final Project Documentation