BME/CENG/ELEC/SENG 499
Design Project

 

2012 499 Spring Design/Technical Project Summaries


Group 1: Lazer Harp for Kinect

For our final undergraduate project we will design a virtual reality musical instrument. This instrument will simulate a variety of string instruments, from real instruments to the physically impossible variety. The LazerHarp will be a fun controller-free freeform musical exploration tool built on the Xbox360 Kinect and a home brewed 3D visual interface reminiscent of a bob ross episode.

Of physically un-constructible dimensions by detecting finger gestures in a user modifiable 3 dimensional space and physically modeling the effect those gestures would have on a predefined array of virtual strings. The finger pluck/strum recognition system uses a Microsoft Kinect and OpenCV, a visual feedback interface which helps the performer consistently with a consistent input-output control using a micro-controller-based hardware system, and a physically realistic synthesis of the sound of plucked strings of impractical physical dimensions using string instrument physics models and audio signal processing techniques. These components will be integrated using the OpenFrameworks toolkit. The practical scope of our project includes programming audio signal processing and computer vision.


Group 2: μMesh

The goal of the μMesh project is to develop a small device which may be used to form a wireless network of sensors. Several μMesh boards will communicate with one another, and route information between the wireless network and a PC. This portable device will be powered by an on-board lithium-ion battery cell, which is recharged using the USB port.

The unique feature of a mesh network topology is that network nodes which are not directly within range of each other may still establish communications by routing data through intermediate nodes. In this manner, the intermediate nodes repeat the messages using radio frequency (wireless) communication until they reach the destination.

The μMesh leverages some of the latest wireless communications technology for low-rate wireless personal area networks. In order to demonstrate the operation of this network, some basic sensors are used to provide telemetry. Ambient air temperature, ambient light level, and acceleration sensors are included on the board. Expansion capability is provided so that the network nodes may control other devices. Other practical applications of this technology include energy monitoring, indoor and outdoor environmental monitoring, remote controls, building automation and access control.


Group 3: Universal Hydrogen Injection System

To meet today's fuel efficiency regulations many measures are being taken by automotive companies to reduce fuel consumption and produce cleaner burning engines. Our project focuses on both aspects of this by not only reducing the amount of fuel burned but also producing less carbon emissions. Furthermore, the system will only require water and a simple salt to produce clean fuel to be utilized.

To achieve this goal it is proposed that hydrogen is injected with gasoline into an engine therefore reducing the amount of gasoline burned. The target reduction is 35-45% in any internal combustion engine utilizing gasoline as its primary energy source. Reducing fuel consumption will ultimately lead to increased gas mileage but at what cost? We want to create a system that can be fitted in any vehicle at a low cost, with a return on investment in 2 years.


Group #4: Garden Gnome Drone

Unwanted pests, which consume plants, fruits, and vegetables, are a common problem for gardeners. An effective system needs to be developed to deter these pests while minimizing disruptions to non-target animals. It should be able to operate on low power, work in a variety of environments, and be easy to setup. The system must also be environmentally friendly and relatively low-cost.

This project includes designing a UAV that will be used to autonomously protect a garden. Two or more sensors to detect movement will surround the garden. If the sensors detect an intrusion, the UAV will take off from its charging station, complete its flight path and land back on its charging station. The sensors will constantly poll for intrusion, and the UAV will be ready at any time. The weather conditions that the UAV will be designed to operate in are limited to light wind only. Once the UAV has been tested for stability under these conditions, there will be opportunity to test the UAV with stronger winds (weather permitting.) To avoid damaging the UAV, testing in strong winds will be very controlled and limited. There are no plans to test the UAV in the rain at this time. The UAV will be able to protect small to medium sized gardens, as the battery life is a limiting factor (discussed below). Given the size and slight noise generated by the UAV, it is expected to be able to deter most pests, including deer, without disturbing people in neighboring houses.


Group 5: Assistive Technology for Monitoring and Managing Diabetes

Diabetes is a disease that affects millions of people in North America. Medical research suggests that by maintaining a healthy lifestyle and being physically active can reduce insulin resistance and better regulate insulin levels. GlucoFit hopes to help by providing blood-glucose monitoring software combined with exercise monitoring via the FitBit device to suggest user-specific exercise goals. In addition, GlucoFit will bridge the gap between patient and doctor by providing secure access of user data to their medical health-care professionals.


Group 6: Predictive Keyboard with Learning

Currently the technological world is moving towards a more mobile environment. Smart phones and other mobile devices are becoming a crucial part in this technological era. Most users of smart phones use them for a variety of things, including texting, writing emails and using a multitude of social media applications. All of these actions involve typing on a mobile keyboard. Touch screen keyboard are standard in the majority of the new smart phones. In addition to the keyboard users are also given predictive software. This software is run while the users are typing messages and it will suggest alternatives to the user if they typed something wrong or if it thinks it can complete the current word the user is typing. As this software is rolled out to millions of users, it is by no means targeted or tailored to an individual user. With an un-targeted approach, the software will not be able to pick up on individual user's habits and even learn from them.

We plan to develop a predictive text Android keyboard that uses AI concepts, such as machine learning, to accurately predict the users future input. By analyzing previous messages in addition to basic text and dictionary data, the hope is to build a keyboard that is fast, and requires minimal keystrokes to complete a message. The components of this project include a base data-mined statistical model for text prediction, in addition to the formation of new pathways based on user input. This will be the most in depth portion of the project. It will also include the mobile keyboard and installer required to setup the application. Current keyboards include some of the aforementioned features to varying degrees of success and completion. Our hope for this project would be to create a robust application that is fast, easy to use, and requires minimal keystrokes, while also having a low error rate. By using data mining and other AI concepts in addition to mobile application development, we hope to accomplish this goal.

Key features such as a clean UI and user-controlled settings are vital to any mobile application. Settings and information will need to be clearly laid out and accessible to the user at all times. For the application itself, the main features will be learning and predicting on a word by word basis. These features will be the main focus of the application. The success of the prediction and learning will impact how well the app does in the Android market. In addition to learning, the application will need to start with a baseline statistical model and dictionary that can be used by all users effectively and efficiently. Another feature that could be beneficial would be plug and play dictionaries that would modify the base model to include specific information pertaining to the interest of the user, such as sports or medicine. To start the learning process off, users will have options for the application to read some of their past messages to modify the baseline model. This reading feature will check the users past texts/emails and additionally some social media if the user agrees to give access. Overall, the main focus of the application will be on providing the users with a tool that can accurately and efficiently predict their text in a manner that will save them time and keystrokes.


Group 7: Autonomous Hexapod

Our team plan to build an autonomous 6 legs robotics walker (Hexapod Robot). The robot is to study and simulate the walking motion of a 6 legs insect. The hexapod robot consists of 6 legs with 4 degree of freedom on each leg to allow for better mobility and exceptional range of movement. The robot will be controlled using the popular Arduino Uno programming board and lynx motion servo controller. To achieve autonomous functionalities, we decide to use infrared sensors or ultrasonic sensors for range detection and utilize touch sensors for close obstacle detection. Finite state diagrams will be developed for obstacle detections.


Group 8: USB Light Pen

The problem is to make a USB light pen to read information from screens and to write on the screens working as generic mice. Since the USB light pen works on any LCD screens having USB connection, it requires special color mapping to detect positions of the curser. It also needs a small CCD to convert images to data code and vice versa on the tip of the pen.

A USB light pen needs to be built and the corresponding software needs to be implemented.

The pen consists of a core of multimode fiber at its tip that transmits light to a photo-detector. The photo-detector translates colors into different voltage, which is then converted to serial data (A/D). The serial data is transferred to the computer to be processed. The software uses a color map that covers the entire screen to match the location of the pen with the color contained within the serial data. The cursor is moved to this location. The color map is only generated within the area of the cursor when the pen is in contact with the screen. Otherwise, the entire color map is rendered, although it only appears at intervals (such as once every 60 frames) to allow for the user to see their screen rather than the color map.


Group 9: Design of UVIC EcoCAR's Energy Storage and General High Voltage Systems

The UVic EcoCAR project team requires a group of students to design the electrical energy storage system (ESS) and the overall high voltage (HV) system of the EcoCAR. For this project we will be designing a battery pack for the ESS based on specific sets of battery modules and control electronics provided by the EcoCAR team. This includes the electrical component specifications, thermal analysis, EMI, and safety analysis.

In a similar way to the design of ESS, we will also be designing the HV System. This will include the design and the specifications of the overall HV distribution hardware that connects the ESS, electric motors and the safety systems.

During the last semester, a group of students worked on the EcoCAR and completed part of this project. Our task is to continue and complete their work.


Group 10: Stabilized Interferometer

Interference Lithography is a method of producing a standing wave through the use of two lasers which stem from the same source. The concept behind interference lithography is to use a laser (in our case a Uniphase 1.5mW HeNe laser) which is directed through a beam splitter. As the name suggests, the beam splitter will split the laser into two separate optical paths where one is directed onto a mirror while the second is directed onto a piezoelectric adjusted mirror. Each path is then directed onto a substrate where the standing wave can be observed.

An uncompensated interferometer is vulnerable to vibrations that stem from acoustic sources and thermal expansion within various components. The effect of vibrations on the interferometer is erratic fringing in the observed standing wave pattern.

The goal of this project is to determine and implement a quick detection method and design a control system that will be used to adjust the placement of the piezoelectric adjusted mirror in order to compensate for acoustic and thermal vibrations which would otherwise cause the erratic fringing; as described above.

When working with an optical system (laser) on a platform, vibrations can affect the signal output greatly. Our project is a proposed solution to this problem by creating a stabilized Interferometer. A difference in location can be observed by splitting a laser (optical signal) into two equivalent parts and reflecting them to a common point. Using the difference between the two reflected laser patterns, a control system can be used to align the reflected lasers and stabilize (align in a preferred way) its platform thereby removing the vibrational effect on the optical signal. In our case we want to improve upon currently used techniques by creating a quick response while maintaining a high precision.


Group 11: Modular lithium Polyme Battery

The telecommunications industry is a major user of modular battery backup systems. Currently these systems are composed of three main components:

  • Rectifiers (convert Mains AC into 48VDC)
  • Lead acid batteries
  • System controllers

Significant strides have been made to improve the efficiency of the rectifiers, both in terms of power density and energy efficiency. However, little has changed with regards to batteries. Lead acid batteries have a number of disadvantages when compared to more modern battery technologies. They tend to be significantly heavier, more bulky, and less energy efficient. They do have a number of advantages, mainly in system cost and ease of integration. As requirements for energy efficiency become more and more stringent and the desire to offer more distributed energy storage increases (for the use in peak load shaving and distributed generation systems) modern battery technologies becomes more appealing.

Currently most Lithium Polymer Battery Management Systems (BMS) work with low level module controllers on each battery module with a connection to a central controller. Previous attempts by ATL to utilize these types of systems have resulted in less than ideal solutions. The complexity of integrating a third party controller into the very stable and mature Cordex Controller has proven to be difficult at best and unreliable at worst; a better solution must be found. We have formed a team to study the problem and build a prototype Lithium Polymer battery module which would provide all the internal protection to make the module intrinsically safe. Modules will be connected in parallel to form large battery banks and any external communication between the modules and the Cordex controller will be through CANbus. The batteries will operate at a nominal 48V with over-voltage and under-voltage thresholds of approximately 54V and 42V respectively. The modules will provide all required protections to make the cells intrinsically safe.

The use of more advanced battery chemistry's in the telecommunications industry is being hampered by the increased system integration cost and the difficulties in making large battery packs safe. With the drive for higher efficiencies, power densities, grid modularity, and 'green' products, the telecommunications industry will transition away from the current lead-acid batteries. While numerous companies are creating battery packs for the electric vehicles market, a viable solution specifically tailored for the telecommunications industry does not currently exist.


Group 12: ECOSat Power Subsystem

The UVic ECOSat team is designing a miniature satellite for the Canadian Satellite Design Competition. Our team, HELIOS, will design, build, and test the power subsystems required bythe ECOSat team. This includes optimized power generation from the solar cells, energy storage provided by lithium-ion batteries, and power regulation for the distribution within the satellite.

The ECOSat team will provide specifications for their anticipated solar panels, battery cells, and power requirements. Our team will design a fully-functioning and complete solution to the power generation and distribution for the satellite, satisfying ECOSat's requirements. The final solution will include all electrical circuitry required for the power generation and regulation, as well as the control and software code for correct operation and the desired communication protocol. Satellite-specific tasks, such as choosing radiation-hardened components, performing thermal analysis and optimization, and reducing design size to fit the satellite's physical constrains will only be investigated if sufficient time remains after creating a functioning prototype. The group will attempt to design for redundancy and fail-safe operation. However, given the time constraints of the project, the final prototype may not include complete fail-safe operation.