December 2011

Department of Electrical & Computer

Engineering’s

The goal of this senior design project was to design, develop, and implement a humanoid robot that can locate a ball, move towards it, pick it up, locate a goal shaped like a field goal, and kick the ball towards the field goal. This project was worthwhile because it taught the students involved how to use their knowledge of electrical engineering principles and com-puter programming to imple-ment a robot that can perform this complex task. In order to perform this project, the group had to understand how to build a robot using individual servos in such a

way that provided the correct level of articulation necessary to give the robot human-like movements. Knowledge of the human body and the physics involved in its movements were required to successfully complete this task. Because bipedal motion is essentially falling onto one leg while balancing on the other, a great deal of precision is necessary to maintain stability. The walking motion alone required a great deal of coordination between each of the servos in order to provide sufficient balance such that the robot does not topple over while performing its tasks. The software implementation involved using a programming language similar to C and the group had to consider all of the pos-sible states that the robot could potentially be in, such as moving with the ball and also without. In order for the robot to be able to detect the ball the group had to buy a camera specifically made to communicate with the robot we built. The Havimo 2.0 is the camera we used. It is a computer vision solution for low power microprocessors. It is equipped with a CMOS camera chip and a microcontroller which performs the image processing. The results are then accessed via serial port. Havimo 2.0 is compatible with integrated programs developed in Roboplus. The robot was able to detect the ball using color and area detection. The color detection was calibrated in such a

way that the camera would detect the color of the ball as well as the color of the field goal posts. In order to imple-ment the area detection knowledge of integral calculus was needed to be able to program the camera to calculate different area values of different shapes and only approach the shapes that fall within a particular range that we had predetermined. The group also had to design a custom “head” for the robot in order to mount the camera onto the robot and make the necessary connections for it to communicate with the brain (CM-5) of the robot. The custom

mount was designed with two extra servos that were connected in such a way that it allowed for the robot to pan and tilt so that it could scan the arena for the ball. In conclusion, our group developed a robot which can find a ball shaped object, then find a field goal shaped target, and kick it with a measure of accuracy towards the field goal target. The specific design of a football kicking robot may lack applications outside of recreation, but the field of robot-ics is an emerging technology with robots that can perform human movements being used in numerous applications in the real world.

Field Goal Kicking RobotTeam 1: Field Goal Kicking RobotAdrian Araiza, Chinh Bui, Emil Stavinoha, Chris Watts

The purpose of the Hybrid Power Generator was to provide power to a load with seamless switching between two separate sources. The idea behind the project was to use alternative energy as a power source for as much time as possible. However, when the environmental-friendly energy cannot provide enough power to the load, a more conventional energy source can be used instead. By providing seamless switching between the alternative energy source and the conven-tional fossil fuel source, the load will not be affected when there is a source change. The use of both alter-native energy and fossil fuels as sources produces a hybrid power generator.For this project, the two power sources chosen were

a solar panel for alternative energy and a gasoline powered engine for the fossil fuel source. The solar panel produced about 20 Volts and 3 Amps (60 Watts) when angled for optimum output and was connected in par-allel to a battery in order to con-stantly charge it. The battery and solar panel combination provided the 12 VDC output needed to power the load and fulfilled the alternative energy source element of the project. The fossil fuel

Team 2: Hybrid GeneratorGrant O’Connor, David Reed, John Sokol and Derek Warnecke

Hybrid Generator

source used was a gasoline-powered engine taken from a weed eater. It was connected to an automobile alternator and regulator that produced the desired 12 Volt DC output.Both sources were connected to a 1,100 Watt

inverter that received the 12 VDC input and provided an output of 120 VAC. A relay cir-cuit allowed only one source to be used by the inverter at all times and a 93,000 µF capaci-tor supplied enough charge to allow time for the relay to switch. In order to choose which source to use, a TI Piccolo Microprocessor was programmed to control all monitoring and switching. By using the Analog to Digital Converter the battery level was monitored and when it reached a point that was too low, the fossil fuel source was turned on. After one minute of lag time to allow the gasoline-pow-ered engine to reach a steady state, the relay was switched. The inverter and load were

powered by the fossil fuel source until the battery was charged to a high enough level by the solar panel. Once the battery reached a sufficient charge, the relay was switched again to have the load and inverter pow-ered by the alternative energy source and the gasoline-powered engine was turned off.By using the hybrid power generator, companies

have the ability to help save the environment and cut costs. The green and alternative energy source helps the environment because the fossil fuel source cannot pollute the air if it is running for less time. A company would be able to cut costs by reducing the amount of fuel needed to power their load. Overall, the hybrid power generator is a valuable asset to all people seek-ing to reduce their carbon footprint.

Monitoring brain and heart functions for patients’ wire-lessly through Bluetooth transmission and display it on an Android phone, as well having a backup server saving the data for doctor’s review.

Objectives:• Integrate multiple biomedical monitoring sys-

tems into wireless transmitter• Develop Android application to analyze bio-

medical signals• Provide real-time qualitative feedback to user• Relay data to remote server for later analysis

Deliverables:o 1 channel, 2 electrode EKG Analog Interfaceo 4 channel, 6 electrode EEG Analog Interface with Active Electrodeso Driven Right Leg Common Mode Feedback Circuito Arduino Microcontroller-based ADC & Bluetooth Interfaceo Android Monitoring Applicationo Server Application

System Block Diagram:

A Low-Cost Android-Based Wireless Biomedical Monitoring System Team 3: A Low-Cost Android-Based Wireless Biomedical Monitoring SystemNick Brennan, Kevin Hinson, Chris Patrick, Carlos Silva

Our project is an external transceiver that can be plugged into an iPhone. This transceiver can transmit data coming from the iPhone using RF and receive it on another transceiver connected to another iPhone. The iPhones contained software developed to transmit and receive these digital signals through the serial port of each iPhone. Using this device you will be able to transmit data even though you are out of range of the iPhone network.

ObjectivesBuild our own network that can be

connected to an iPhone for transmitting and receiving, and successfully transmit and receive digitized text.

Iphone Radio Transceiver/ReceiverTeam 4 : Iphone Radio Transceiver/Receiver

DeliverablesiPhone apps and hardware that

transmit and receive text data over the constructed network. The data will be transmitted serially from an iPhone to our transmitter. The receiver will receive this data, re-digitize it and display on the receiving iPhone.

LNADemodulater

IPhone

Rec.Receiver Design

Modulator

IPhone

Trans

Carrier

Transmitter Design

RF AMP

Wheeling ElectricityTeam 5: Wheeling ElectricityMike Lu, Ryan MacGibbon, Sarah Thibodeaux, Sophie Wang

The objective of the project is to convert a bicycle’s mechanical energy to electrical power, connect it to the grid, and compare it with home load consumption. The stationary bicycle is modified to turn an alternator. The low voltage AC power from the alternator is rectified to low voltage DC at about 15-18V, depending on the speed of the bicycle. The DC/DC converter steps up the low voltage DC to about 190-270V DC and provides isolation. The inverter converts the high voltage DC into AC that is synchronized with the grid. A power meter measures the consumption of the house and/or bike power output and displays it graphically on the user interface using an Android OS tablet.

With the smart-phone quickly becoming the mainstay of the mobile communica-tions industry, there are thousands of applications being developed to per-form every function from those that are resourceful such as, monitoring energy usage and con-trolling in-home lighting, to the more frivolous and fun such as, the many gaming applications. One of the many advanced functions available to users of smartphones is the devices’ ability to respond to physical tilting inputs. In addition to pushbuttons, the tilt function allows the user to input to the device by other means when it may be convenient, or in many cases, practical to do so. The objective of this project is to use an Android smartphone to remotely drive and steer an RC car, and to stream front-view video from the car to the smartphone display for out-of-sight control.The design of what is called the Infinite-Range Android

Car (IRAC) features an Android smartphone that transmits data through TCP/IP to a Linux-operated BeagleBoard single-board computer onboard the car. The BeagleBoard processes the data and streams commands to an MSP430 microcontroller, also onboard the car, which produces and outputs pulse-width modulated (PWM) signals to the servos and rear motor of the RC car. A camera mounted on top of the car receives front-view video to the BeagleBoard, and that video is streamed to the display of the IRAC software application user interface. Lastly, an ultrasound sensor is implemented as a failsafe to stop the RC car’s forward motion should there be an approach-ing obstruction.The Android Software Development Kit

(SDK) was used to develop the software ap-plication in Java language. We intercepted the accelerometer and magnetometer signals from the smartphone to generate four degrees of left turn commands and four degrees of right turn commands using the tilt function on the smart-phone. Touch buttons are implemented to gen-erate three forward speed commands, a brake that defaults to neutral, and one reverse speed

Infinite Range Android Car (IRAC)Team 6: Infinite Range Android Car (IRAC)

command. Next, M-JPEG video format was implemented and incorporated into the graphical user interface (GUI) of the IRAC soft-ware application. To communicate between the Bea-gleBoard and the

smartphone, we established a TCP/IP connection hosted on the BeagleBoard to talk over Wi-Fi to the smartphone. A program running on the BeagleBoard parses integer values received from the smartphone, and determines the command that the car needs to follow. After the com-mand is determined, a single character is streamed to the MSP430 microcontroller through the transmit line of the serial port using the RS-232 protocol. The microcontroller is programmed to output two simultaneous PWM signals at 50Hz frequency from two separate capture/compare registers to the car’s servos and rear motor.The realization of the IRAC has the capability to be

utilized in many different areas of engineering, as well as everyday life. Identified areas of real-world applica-tion include: (1) the car can be used in search and rescue services where human sight and intervention is unavailable or impossible, and (2) the software application can be used to control industrial machinery and instrumentation. A rec-ommended future work is to add a voice command control feature to the IRAC.

American Sign Language TranslatorTeam 7: American Sign Language TranslatorAndrew Acosta, Josh Guidry, Greg Kling, and Matthew PriceThis senior design project is a combination

of hardware and software that will translate American Sign Language hand gestures into spoken words. Surface electromyog-raphy (sEMG) sensors are used to detect electrical activity within muscles of the forearm. Each gesture has a specific, char-acteristic electrical signal, thereby allowing the translator device to distinguish differ-ent gestures. The received gesture signal is amplified, filtered (pre-processed), and sent to the internal ADC of the Atmel AT-mega328 processor, utilized by the Arduino platform, for processing. This digitized value is processed utilizing a C-based code that employs signal segmentation, feature extraction, and classification. Reference

digital values, acquired from gesture measurements used to “train” the device, are stored in look-up table fashion for signal comparison. The incoming, processed digital values are compared to the refer-ence values. If a match is detected, the processor will retrieve corresponding audio file from the SD card inserted into the Wave Shield, process the file for playback, and then send it to the audio ampli-fier on the Wave Shield for the translator device speaker. Figure 1 below illustrates operation of the translator device.

Our project utilized the Playsta-tionEye video cameras as the foun-dation of a multiview camera array. A camera array uses multiple cam-eras to function together to achieve a certain functionality. In our project, the functionality was geared towards 3D imaging. The system is scalable to any number of cameras.The first step in this design was

achieving synchronization. This was done through two different ap-proaches: hardware and software.

The hardware approach involved using one camera as a “master” and the others as “slaves” utilizing the vertical synchronization capabilities of the OmniVision chipset inside the PlayStationEye camera.The next level of synchronization was done through software in OpenCV. Running all the cameras

together created issues because some cameras would gain priority over the others. In order to remedy this, threads were used to give each camera its own set of processes. In order to prevent one thread from accelerating past another, software barriers were put in place. Each camera’s thread was unable to move past certain lines of code until all the other cameras had caught up. This dual layer of synchronization yielded an average missynchronization of about 200 microseconds. Other similar approaches achieved missynchronizations of 10 milliseconds.The next stage involved obtaining depth information from the multiview video. Using a Harris detector

and a point matching algorithm, depth information could be extracted from the videos. Exploration into this area also yielded methods for calibration of the cameras. Finally, a modified H.264 compression al-gorithm was used to compress the multiview video into a more efficient and transmittable piece of data. This program yielded high efficiency and great SNR results.

Camera ArrayTeam 8: Camera ArrayDavid Wisecup, Troy Casanova, Aaron West, Matt Covarrubias, Collin Brewer

The project’s goal was to conveniently place all the control of a home theater system into a single Android Device. With the project’s devices and the applications made for this system, the user can choose and stream videos from their home computer to a BeagleBoard which is then connected to your TV through HDMI. The commu-nication between these four devices, the BeagleBoard, Android mobile device, home computer, and IR transceiver, all communicate through a single home router commonly used in a home setting. These programs will find the other devices once initiated and the user can then enjoy using the control from their mobile device as long as they are on their home network.

HomeSync; Wireless Media Streaming with iPhoneTeam 9: HomeSync; Wireless Media Streaming with iPhoneAndrew Vasquez, Michael Elmer, Daniel Bueso and Jeff Ho

Standing on the ground in Caldwell, Texas, our team looked into the heavens while our final electrical design project flew almost 88,000 feet overhead. Two and a half hours after launch, it came falling back to Earth.Our group composed the idea to send a weather balloon

into the atmosphere to capture photos of the curvature of the earth after hearing of a similar endeavor undertaken by Massachusetts Institute of Technology students. Our team designed the balloon to include more measurements than the MIT proj-ect, including pressure and temperature.The project took nearly an entire

semester of planning, and cost us about $170. It also brought in return a series of high resolution photos capturing Earth from the edges of outer space.We’re very satisfied with the results,

and we had a fun time doing it con-sidering this is one of our last projects before graduating. But we’re really relieved it’s over.The launch of the balloon, which fell

back down to Earth with a parachute, took a couple of days to plan as we had to time the launch in accordance with the weather. We also had to run a pre-launch projection to determine where

Balloon Curvature of the EarthTeam 10: Balloon Curvature of the EarthRichard Colunga, Joel Colston, Sonny Nguyen and Johnathan Laake

the balloon would land. Once the balloon peaked at approximately 88,000 feet, the race was on to pick up its signal.Although we planned extensively, the project still

came dangerously close to being derailed.When the balloon was on its way back down,

there was a brief period where we had completely lost the signal. We all kind of stared at each other. We didn’t think we were going to find it.We managed to find the balloon; although it had

landed more than 100 miles away in Spring, Texas.When we uploaded the pictures from the balloon’s

camera, which had been programmed to snap a photo every 20 seconds on the way up, we were astounded with the images we found.It’s really cool to see photographs of the earth

from as far out into the atmosphere as our balloon was. But it’s even cooler to see them when your camera and your work made them possible. We are really happy with the final results.

The work is not quite finished. Our project is slated to enter into a Texas Instruments competition, and a public demonstration of our balloon’s workings took place Tuesday December 6, 2011 at the Zachary Engineering Center.

Demonstration Day Winners

First Place-A Low-Cost Android-Based Wireless Biomedical Monitoring System Nick Brennan, Kevin Hinson, Chris Patrick, Carlos Silva

Second Place-Wheeling ElectricityMike Lu, Ryan MacGibbon, Sarah Thibodeaux, Sophie Wang

Third Place-Team 6: Infinite Range Android Car (IRAC)