The courses for the capstone projects in the Department of Electrical Engineering and Computer Science (EECS 159A/B) are run by Assistant Professor Maxim Shcherbakov (maxim.shcherbakov@uci.edu) and Associate Professor Hung Cao (hungcao@uci.edu). They run the capstone project sequence during the fall and winter quarters. Interested industry is encouraged to reach out to Prof. Hung Cao at hungcao@uci.edu.
The purpose of our project is to make it easier for people to throw away their trash while providing a way to organize different types of trash accurately. Our project uses a combination of mechanical hardware parts, software components, and machine learning to classify trash and throw them away automatically. The user just has to place their item on our machine's platform and it will throw it into either a recycle bin, compost bin, or landfill bin.
The goal of our project is to develop a system that monitors energy consumption and performance for high energy usage/high waste devices in commercial environments. This power management system addresses power usage by devices with high energy consumption. It will provide power analysis to make the user aware of how much energy the device is consuming and providing the user and device with appropriate corrective action.
Design a system to analyze respiratory waveform. The clinical application of monitoring respiratory wave patterns may be helpful in identifying the infant’s readiness to initiate and progress feeds orally. This measure of variability may also be used as a screening tool clinically to identify coordination patterns in infants with oral feeding challenges.
The goal of the project is to design and manufacture a hand-portable platform that applies hyperspectral 3D photogrammetry for a biomedical purpose. The project integrates the platform with both software and hardware. Specifically, this platform utilizes NVIDIA Jetson TX2 and Raspberry Pi3 as a central controller, connect and control 3 Raspberry Pi0s through hot spot ethernet. Each Raspberry Pi0 connects and controls one LED light ring and one Noir V2 Infrared Camera. LED light rings and infrared light sets are used to control light conditions when collecting image data from Raspberry Pi0 cameras. Image data can be automatically transferred from Raspberry Pi0 to the central controller in order to synthesis the 3D model. We use NVIDIA Jetson TX2 as our graphics processing unit to run Alice Vision Meshroom, which is an open-source 3D photogrammetry software for building 3D models. Also, we can polish and modify models by...
The most crucial step of the engineering design process is collecting and verifying data. Collecting data on a system allows engineers to concretely measure the performance of their system as well as making adjustments based on data. Telemetry systems provide sensory data in real-time which is necessary for engineers to evaluate the current state of their systems.
The goal of this project is to design a functional, intuitive telemetry system that collects data to ensure efficient and safe operation within engineering projects. The system is adaptable using modular sensor packages and the use of industry standards including: On Board Diagnostics (OBD2), which is an international vehicle standard, and Controller Area Network (CAN) Bus, another efficient vehicle standard for data transmission. Projects can utilize this system throughout the engineering design process to monitor and record the performance of their project’s components.
The On Your Mark project aims to create and affordable Smart starting block to donate to a high school track an field team. This project will be able to collect pressure from the runners feet on the plates and create a gradient map of the pressure. The force at take off will also be collected. The data will be compared with controlled/desired data and stored for reference of the runners progress.
Background:
The Autonomous Follower drone is a system in which a "follower" drone will be capable of following a single "leader" drone. The follower drone will be capable of maintaining a threshold distance away from the leader drone, as well as maintain same elevation as the leader drone. This project aims to improve current drone communication systems so that surveillance operations conducted by drones can be more coordinated.
Goals & Objectives
Manipulate a commercial drone so that it may follow another drone of any type autonomously. Develop a system in order to calculate the location of both drones and provide the necessary adjustments to the follower drone via a Raspberry Pi.
Team Members
Andy Vo
David Phan
Hugh Fong
Lawrence Dizon
Advisors
Professor Zak Kassas
Dr.Mahdi Maaref
A wideband spectrum monitor is a device that monitors a selected area of the RF spectrum so that the operator can get an idea for the sorts of signals that regularly exist in the area. These devices are used to make determinations about what areas of spectrum are congested or to identify potential sources of interference. This system can be implemented in NI LabVIEW and it is capable of scanning an arbitrary segment of the RF spectrum. It is only constrained by the frequency range of the SDR in use and its antenna. The system is also capable of saving signals that exceed a user-defined power threshold for later analysis. This is an effective low-cost alternative to more sophisticated, single-purpose monitoring solutions.
Our project, the autonomous target practice robot, will serve as an aid for the Navy in their training exercises. Through the use of a Computer, Raspberry Pi, Arduino, and various sensors we are creating a robot that will drive to a coordinate, hold up a target, orient it to a shooter, and register if it has been hit or not. It will then proceed to the next target. This is achieved through the development of a Graphical User Interface (GUI) and a robot capable of driving across terrain. It also requires us to use data processing techniques to isolate data for proper use within our discrete control system. So far we have a program with a user-friendly GUI to generate random points and send that data to the robot. Once the Raspberry Pi has received the data we are able to process and format it in a way that the...
Our goal is to use a simulation environment to run cybersecurity tests on autonomous vehicles. The goal is to have ways to test AV security outside of purely field tests.
So far, we have placed a sticker below a stop sign and had it recognized as a person. We would like to expand this to have objects in the road avoid recognition.
In rural area, traditional way of package delivery methods can be costly because of limitation of road condition and resources. A lot of companies today are working on this approach.[1][2] With the drone, it would save both time and money for small amount of deliveries. We are designing an autonomous drone for the food/medicine delivery. The drone we designed would be able to lift things up to 2kg via using a claw which is being controlled by Raspberry pi. The landing would be accomplished with the GPS instruction inserted in the controller (the address would come from the app of users) and IR beacon/IR sensor for accurate landing. We would also use the camera in the drone to verify if the object is sent to the correct location.
The goal of Intellidriver is to create a safer driving experience by collecting physical data from the driver. Using the data, Intellidriver will use machine learning to provide various services such as risk management and safe driving suggestions.
This project adds a dynamic element to the traditional Smart Mirror design by incorporating several peripherals, such as a microphone and sensors. These additional mediums of interactivity allow the user to customize their mirror hands-off. The GUI of the Mirror is built from scratch in Python, allowing for more flexible interaction and displays than market tools. More to come soon.
Design an OPAMP by hand to get a feel for the process. Create a tool that solves the design problem for Miller Compensated Two-Stage OPAMP topology. Formulate the design problem as a geometric program [1]. Reformulate geo- metric program as a convex optimization prob- lem [1]. Solve efficiently for a globally optimal solution [1].
According to the National Spinal Cord Injury Statistical Center, every year, there are approximately 17,700 new incidences of spinal cord injuries in the United States, many of which result in either partial or complete loss of mobility. Our EyeDrive systems aims to reliably and affordably help patients retain their mobility by empowering their eyes and enabling them to control a first-person-view RC car with nothing but their eye-gaze!
This is achieved using:
-The OpenCV framework in Python;
-2 Arduino UNOs
- A 2.4GHz Antenna (NRF24L01)
- A Dual Motor Motor Driver (L298N)
- 4 AA Batteries
- 1 Raspberry Pi
And the help of PyGaze and TheDroneBotWorkshop on YouTube
PyGaze Channel: https://www.youtube.com/channel/UC5hHNks012Ca2o_MPLRUuJw
DroneBotWorkshop: https://www.youtube.com/channel/UCzml9bXoEM0itbcE96CB03w
Thank you, guys!
The goal of the project is to design, develop, and test a functional torque vectoring algorithm to increase the performance of electric vehicles that utilize two independent rear motors.
So far we have implemented a simple torque vectoring algorithm that is responsive to accelerator position and steering angle. We have implemented this algorithm on a hardware testbed and verified it is functioning as intended.
This quarter we are working on incorporating more data into the algorithm, particularly inertia and suspension travel data. In addition we are working on tuning the parameters of the algorithm to improve overall performance. Finally, we are working on implementing the algorithm on a full-sized FSAE Electric vehicle so we can test our algorithm empirically on an actual vehicle.
The Surrounding Environement Scanner is a scanner that uses lidar sensors, a gyroscope/ accelerometer, and motors to obtain geometrical data from the surrouding area. This data will be used to create PCD files that save the points obtained by the sensor. These stored points are then used to create a 3D representation on a web appplicaiton. The SJNC group hopes to expand the reach of 3D scanning technologies with the development of this sensor.
The MASQ Motorcycle Safety System has different functionalities:
1 - Warns the rider if the helmet is not buckled. The helmet has a minicomputer installed. At start up the system checks if the helmet is buckled.
2 - Keeps the rider aware if his/her surroundings. Using stereoscopic cameras and ultrasonic sensors, an embedded system calculates positiions and distances of vehicles behind the motorcycle. It then sends the data to the helmet mini computer which will display it using LEDs mounted below the visor.
3 - In case of accident it will contact an emergency number. Acceleration is monitored on the helmet. An algorithm is used to detect accidents based on decelerations. The helmet sends the status to the motorcycle embedded system which will text the GPS coordinates of the current location to an emergency number.
4 - Extra features : Battery monitor with charge display, LEDs brightness adjusted automatically, Emergency...
The goal of this project is to build an autonomous car. The milestone goals to achieve this are to obtain live video footage from a Raspberry Pi, detect road lanes and recognize objects using OpenCV, and control the movement of the RC car with the use of an Arduino and L293D chip. Development of an iOS app is currently in progress, which will allow us to control the movement of the car manually.
Instagram: @ampeaters
Youtube: https://www.youtube.com/channel/UCwAYPD4DxD-ceooR0akPOFw
Reminder aims to aid people in taking the right medications at the right times. The project creates a mostly unmanned line of communication that will intelligently remind users via smart phone to take their pills. To accomplish this, our design proposes that pharmacies use an RFID communication module to write simple medication usage guidelines to RFID tagged prescription bottles so that patients can scan the tags at home and have their machine assign periodic reminders through a smart phone app. With Raspberry Pis we’ve implemented the RFID reading and writing requirements as well as the database management. Then, using Android Studio, we’ve created an Android compatible app that downloads medication information from the same database to display for the user’s convenience. A finger print scanner addition to the dispensing unit enables the system to track a patient's medication adherence so that notifications can adapt depending on how well the user has been...
A cloud-deployed, functioning telemetry tracking of an unmanned vehicle and controlling it in waypoint mode through a web-app deployed on AWS. User can register with our service and connect to multiple vehicles simultaneously. Our end goal is to fully deploy the CloudStation app and demo it on vehicles across the globe.
Our project is open-source, check out our GitHub page here: https://github.com/CloudStationTeam
Our project is focused on the creation of a dynamic mesh network that is specifically implemented for networks with highly mobile nodes, for example drone swarms. This implementation is unique in its ability to reconfigure itself as necessary when devices and components of the network drop in and out of the network. This project is associated with Professor Marco Levorato's HYDRA project.
For more information: https://github.com/ericcai9907/Dynamic-Network-Project-Documentation.
FiAR is an Augmented Reality (AR) helmet for firefighters that provides the ambient temperature, heart rate of the firefighter, GPS location of the firefighter, and radio communications.
Our project goal is to design a simulation for the performance, primarily from the signal processing perspective, of a communication system that transmits WiFi to eight different aircraft in a given aircraft. In our project we calibrate for errors in phase and gain that arise from hardware inconsistencies by estimating the error and thermal noise. We then must shape our beams in order to minimize the interference each beam observes from the other beams. This is because they are all using the same carrier frequency. We also are designing an over the air real time calibration method that will be able to calibrate for hardware errors that arise over time due to deterioration.
SmartFarmer is bringing Internet-of-Things to farms. Utilizing sensors and the Cloud, SmartFarmer is developing UwUPlants -- a hardware-as-a-service (HaaS) to monitor the state of users' farms. Utilizing microcontrollers and WiFi mesh networking, SmartFarmer aims to minimize data costs by linking multiple sensors to a single bridge node, where data can be streamlined and hardware complexity can be reduced on average. Various sensors are used to track farm vitals, such as soil moisture and local temperature, and those vitals are analyzed for users to view at a single dashboard. To see more details, please visit uwuplants.org
Mind your control is a multidisciplinary senior design project with the goal of helping paraplegic people control daily appliances. It uses EEG (electroencephalogram) signals from a headset to read a user's attention levels. These attention levels are then used to control a mobile application, which is, in turn, connect to daily appliance i.e a speaker, door, etc.
Whiteboards are generally considered to be an effective tool to assist in the educational and professional industries. With the advancement of technology, interactive whiteboards that use a projector as a screen and a board that registers user input have modernized the concept. However, traditional devices of this type tend to have a high cost, be limited by a fixed boardsize, and must be invasively installed using wall mounts. We created a cost-effective and portable alternative to conventional interactive whiteboards that can be used on many surfaces. By using small infrared sensors to calculate the position of the stylus, we remove the limitation of a predetermined board size, and by adhering to a wall temporarily, the board interface can be set up rapidly. The movement of the stylus in the board will be replicated on the computer.
Our project is called efficient deep racing via environmental contraining, our goal is to create a cost efficient yet high performance autonomous driving system by limiting the variables needed to process through using mario kart as the environment.
We are team Influence and our project is to detect the influenza virus via our model. The model is based on four different data parameters from human subjects: ECG, PPG, heart rate and body temperature. Our goal is to limit the spread of the flu by detecting early carriers and encourage better self-care if one is found with the influenza. As such, we will reduce the impact of the influenza by reducing its ability to spread from unknowing carriers.
This project is about the Autonomous Vehicle, our goal is to build an autonomous car which it can avoid obstacles automatically, recognize and follow roads automatically, recognize traffic lights and signs automatically. In order to achieve the above requirements, we need the following materials: a modified RC car, battery, driver board, raspberry pi, lidar, and camera. Battery-powered, driver board can control moving and stop of car, raspberry pi used to control lidar and camera, lidar recognizes and avoids obstacles, and cameras are responsible for identifying traffic lights, signals and making vehicles follow the road. The reason for designing autonomous cars is because autonomous driving technology is the development direction of future driving modes. Mature applications of autonomous driving technology will bring more convenience to people's lives. At the end of the first quarter, we completed the movement and stop of the car as well as the lidar part as planned...
