Project

# Title Team Members TA Documents Sponsor
47 Automatic cake decorator
Honorable Mention
James Zhu
Muye Yuan
Rui Gong
Jason Zhang design_document1.pdf
design_document2.pdf
final_paper1.pdf
presentation1.pdf
proposal2.pdf
proposal1.pdf
video
# Team Members:
Muye Yuan(muyey2)
Rui Gong(ruigong5)
James Zhu (tianyi9)

# Problem

The current challenge lies in manual application of cream on cakes, prompting the need for an automated solution. Traditional methods often result in variations in cream thickness, coverage, and overall quality due to the nature of manual application. This not only demands skilled workers but also leads to increased production costs and the potential for human errors. Moreover, labor costs can be a significant factor in the overall production costs.

# Solution

We decided to make an automatic cake decorator, which puts creams with shapes and curves around the edge of the top surface of the cake. By automating this process, we aim to eliminate the inconsistencies associated with manual application, enhance the overall quality of decorated cakes, and reduce production costs. Ultimately, this device can offer a more efficient and cost-effective solution for the baking industry. The decorator can move along the edge of the cake detected by the camera. According to the input, the movement will be divided by x and y components which can lead the stepper motor to the appropriate position. This system differs from existing food printer solutions, which only print pixelated images on the food. It leaves a vectorised, continuous trail of cream. So it requires a more dedicated CV algorithm to recognize the shape of cakes.

# Solution Components
##Subsystem1 Computer vision and detector:
1x 1080p usb camera, laptop
A frame holds the camera hanging it on the top of our decorator machine, looking down to the cake in it. It’s connected to a laptop running our recognition program. The program would recognize the edge of the camera with a CV algorithm. It could identify the cake successfully even with other distractions (like the machine itself) in the view, and fit the edge into a set of waypoints for the cream extruder to follow. The program presents a preview of it for the user to confirm. The laptop is connected to MCU PCB with USB. Once a key is pressed, it would send out a waypoint to the MCU and signal for it to start moving the mechanical parts.

##Subsystem2 MCU and PCB
1x ATmega328P MCU, 1x self designed PCB with the MCU and the motor driving circuit
Input: Usb connected from the laptop
Output: Control signal to the step motors driving the extruder and the cream syringe.
Once a set of waypoints is received, the trajectory following the waypoint would be converted into its projections on the x and y axis, and the function of x and y position over time would be calculated. (these calculations might be done on the laptop as well). Then the program on the MCU would start and drive the two sliding rail motors, as well as the motor pushing the syringe.

##Subsystem3 Mechanical structure
3x 42-40 Stepper Motor, Cake Decorating Tools Cupcake Injector, rectangular frame, and 2x Linear Rail Guide, height adjustable base (placing the cake)
The structure of the machine resembles that of a cartesian robot, or a 3D printer, which is two perpendicular sliding rails (powered by motors) connected to each other, able to move its tips to arbitrary x-y positions. A large syringe with cream inside is mounted at the tip, extruding the cream uniformly when pushed by a motor.

# Q&A
##1.Decide whether to implement a 2D or 3D movement system.

We want to implement the 3D movement system, but we don’t know how complex it is. Thus, if the 3D system is too complicated for us to implement, we will change to implement a 2D movement system.

##2.Clarify the mechanisms you plan to use for x, y, and z movements. Will they be similar to those in a 3D printer, and how will you ensure movements, when working with a medium like cream?

Yes. It is similar to 3D printers with two perpendicular sliding rails. And we are planning on putting a rubber hose on the syringe and the end factor of the mechanism grabbing the other end of the hose, keeping the relatively heavy syringe static.

##3.Determine the dimensions of the machine(syringe size, etc). Are you considering a vertical actuator to push the cream out of the syringe? Detail out all the electrical components required for this idea.

We want to start from a small size, so the amount of cream will not be large. For example, we start from using the Cake Decorating Tools Cupcake Injector and a step motor pushing it to get the cream out.

##4. The incorporation of a camera for position detection adds complexity. How do you plan to convert the camera inputs into xyz position? The coding required to convert camera output into g-code(x,y,z) is critical.

The z position is fixed for a cake. We first require the user to place height of the cake manually so that its top surface is near the extruder. Later we might add an ultrasound system and an automatically adjustable base for the cake. For x,y coordinates, we might first try to mount the camera high enough, so that we can assume it’s a planar projection from the pixel coordinates to the physical. We would first fix the relative position of the machine and the camera and do calibration (mapping from pixel coordinate to physical) manually. But later we could try adding some marks on the edges on the machine, the camera can automatically figure out the linear translation without the need to calibrate every time. If the error of assuming planar projection turns out to be too large, we could still figure out the intrinsic of the camera and do unprojection with formulas.


# Criterion For Success
-CV system recognize the edge of the target successfully

-Moving system can successfully follow the input instruction

-Put cream with a curve around the edge of the top surface of the cake.

Recovery-Monitoring Knee Brace

Dong Hyun Lee, Jong Yoon Lee, Dennis Ryu

Featured Project

Problem:

Thanks to modern technology, it is easy to encounter a wide variety of wearable fitness devices such as Fitbit and Apple Watch in the market. Such devices are designed for average consumers who wish to track their lifestyle by counting steps or measuring heartbeats. However, it is rare to find a product for the actual patients who require both the real-time monitoring of a wearable device and the hard protection of a brace.

Personally, one of our teammates ruptured his front knee ACL and received reconstruction surgery a few years ago. After ACL surgery, it is common to wear a knee brace for about two to three months for protection from outside impacts, fast recovery, and restriction of movement. For a patient who is situated in rehabilitation after surgery, knee protection is an imperative recovery stage, but is often overlooked. One cannot deny that such a brace is also cumbersome to put on in the first place.

--------

Solution:

Our group aims to make a wearable device for people who require a knee brace by adding a health monitoring system onto an existing knee brace. The fundamental purpose is to protect the knee, but by adding a monitoring system we want to provide data and a platform for both doctor and patients so they can easily check the current status/progress of the injury.

---------

Audience:

1) Average person with leg problems

2) Athletes with leg injuries

3) Elderly people with discomforts

-----------

Equipment:

Temperature sensors : perhaps in the form of electrodes, they will be used to measure the temperature of the swelling of the knee, which will indicate if recovery is going smoothly.

Pressure sensors : they will be calibrated such that a certain threshold of force must be applied by the brace to the leg. A snug fit is required for the brace to fulfill its job.

EMG circuit : we plan on constructing an EMG circuit based on op-amps, resistors, and capacitors. This will be the circuit that is intended for doctors, as it will detect muscle movement.

Development board: our main board will transmit the data from each of the sensors to a mobile interface via. Bluetooth. The user will be notified when the pressure sensors are not tight enough. For our purposes, the battery on the development will suffice, and we will not need additional dry cells.

The data will be transmitted to a mobile system, where it would also remind the user to wear the brace if taken off. To make sure the brace has a secure enough fit, pressure sensors will be calibrated to determine accordingly. We want to emphasize the hardware circuits that will be supplemented onto the leg brace.

We want to emphasize on the hardware circuit portion this brace contains. We have tested the temperature and pressure resistors on a breadboard by soldering them to resistors, and confirmed they work as intended by checking with a multimeter.

Project Videos