Many hard of hearing or deaf individuals depend on a translator or lip reading in group settings.

The Smart Voice Recognition System (SVRS) design helps users identify what a speaker is saying via voice-to-text, who the speaker is, and where they are located in a room – all using an iPad graphical user interface (GUI). Wireless transfer of direction information and audio files ensure ease of use and portability.

SVRS incorporates a neural network learning model to accomplish the voice recognition component. A built-in voice-to-text framework provides real-time mapping from speech to text. Filtered audio detected by the microphone array and defined by a GUI arrow identifies directionality. A Raspberry Pi connected to the microphone array houses all of the directionality software. The application that houses the voice recognition and voice-to-text software is loaded onto an iPad. And, the iPad and Raspberry Pi are connected via wireless ad-hoc network using MQTT.

Density and porosity measurements are critical to analyzing asteroid samples, typically found on Earth. The OSIRIS-REx project presents a unique opportunity to analyze carbon-rich samples from an actual asteroid, Bennu. Accurate porosity measurements can indicate trapped moisture during the formation of an astronomical body.

The team applied gas pycnometry theory to design a system that mitigated risk of sample contamination and damage, met material and spatial requirements, and accommodated various sample sizes.

The students used error analysis and modeling to determine optimal volume and number of reference chambers. Without lubricants, non-approved materials or shearing contact points, they designed a novel chamber with a gas-tight seal. They used a pressure transducer and LabVIEW VI for measurements. The team combined pressure measurements, the ideal gas law, and 3D sample measurements to determine porosity. Automation allowed for rapid repetition, minimal change in the sample environment, and precise measurements.

General Dynamics Mission Systems performs periodic preventative and corrective maintenance on all of its Rescue 21 direction finding (DF) radio towers for the U.S. Coast Guard. A new true north offset line of bearing must verify direction during tower maintenance. Previously, the costly process involved sailing a boat along a specific path or driving a vehicle on difficult terrain to take measurements with specialized equipment.

This project presents a lightweight, modular attachment for a drone that effectively and efficiently calibrates DF radio towers.

The design attaches a small module with data recording/measuring capabilities to a DJI Mavic Air Drone. Drone pilots toggle on power to the module, activating the Arduino microcontroller and initiating data recording. As the DF North Offset Drone Test Module (NODTM) transmits very high frequency (VHF) signals at specific time intervals during flight around the tower, it simultaneously records and stores position, time and duration data. The data is exported as a single log file and post-processed to produce a DF tower calibration.

Modern technology is providing new functionality for data processing interactions between general purpose processors (GPPs), field programmable gate arrays (FPGAs) and radio frequency integrated circuits (RFICs). Systems on a chip (SOCs) open up the bottlenecks when GPPs and FPGAs are implemented into different integrated circuits. Software-defined radio (SDR) boards include RFIC radio frequency integrated circuits that reduce size, weight and power while improving performance.

This project involved prototyping FPGA software to evaluate the capabilities of the Rincon Raptor board. The team examined the types of processing performed in SDRs, developed an architecture to exploit the interfaces supported by the SOC and prototyped and evaluated an RFIC/FPGA/GPP implementation on SOC hardware.

The team developed signal processing functions in VHDL to find the limitations of the Raptor board’s data rate. Equipped with a Xilinx UltraScale SOC, featuring a combination of a traditional general-purpose processor and FPGA, the Raptor board acts as an ideal environment for development. This setup allows for quicker signal processing speed.

Medical catheters are tubes inserted in body cavities, vessels and ducts to drain body fluids. The diaphragm membrane is a critical component of the catheter. It bursts at a certain pressure when a blockage occurs to allow fluids to flow through the opening created. One such use for this membrane design would be in a brain shunt to drain cerebrospinal fluid. Because of the body’s sensitivity to pressure, especially in the brain, the membrane must burst at a specific pressure to ensure no damage is done to the brain. These ultrathin membranes are difficult to produce.

This team developed a method to achieve the same burst pressure in the membrane while ensuring a consistent manufacturing process. The approach focused on creating a diaphragm membrane with an intentional slit geometry to act as a weak point and burst at a given pressure, providing an alternative path for fluid drainage. The students analyzed several slit geometries and corresponding manufacturing capabilities.

The slits in the membrane were produced using a laser ablation device. The membrane was made from biocompatible liquid silicone rubber to minimize adverse affects on surrounding tissue. The team pressure-tested its design, and, with a durometer, measured the hardness of the silicone. To ensure the system was reproducible, the team analyzed capability and performance standards using the PPK process.

Mosquitoes transmit deadly diseases and viruses. Entomology researchers place surveillance traps to capture the insects alive and determine what diseases they are carrying. These traps, which are cumbersome, rely on the production of carbon dioxide to attract mosquitoes.

This design combines two modular subsystems: the mosquito lure and environmental sensors. The lure uses a one-pound propane tank with safety mechanisms to produce carbon dioxide at preprogrammed times. Environmental sensors collect data while the propane is releasing carbon dioxide into the environment.

Combustion components include a solenoid valve and electronic piezo igniter controlled by an Arduino. A tilt switch stops combustion if the tank is knocked over or tampered with. Sensor components include a GPS, thermometer, humidity sensor and atmospheric pressure sensor. The data is collected on a removable SD card controlled by the same Arduino Mega 2450 Board coded in C++.

Bench testing is key to the safety and efficacy of medical devices. Models must be extraordinarily similar to the anatomy of the human body to generate valid data before devices can be tested on humans.

This project demonstrates a more realistic model of thrombus adhesion to blood vessels in the human body. The design encompasses multiple components to represent an adhered thrombus, including various sizes of tubing, grooving patterns on the interior wall of the tubing, and collagen coating on the interior tubular interface.

Specifically, the team used three tubing sizes to represent different functions of blood vessels ranging in diameter from 3 millimeters to 10 millimeters. The interior wall of the tubing was scored and coated with a collagen solution to increase thrombus adhesion. The components designed for bench testing closely modeled actual human blood vessels.

Every year, the University of Arizona Baja Society of Automotive Engineers team designs and builds an off-road vehicle to race against other schools in competition. The car has a mechanical continuously variable transmission (CVT). It has pulleys, which use internal flyweights and springs for actuation, that vary their diameters depending on engine RPM and load coming from the gearbox.

This project presents a more efficient method that controls the pulley diameter electronically.

The team modeled and designed a system consisting of two pulleys, two motors, two linear actuators, hydraulics, three microcontrollers, an RPM sensor, and a housing to protect the electronics. Additionally, the team implemented a controls algorithm based on modeled vehicle dynamics to determine the proper gear ratios for optimal performance. The students integrated and tested the system on the club’s 2021 Baja competition car, finding that it displayed superior durability, tunability and control.

Northrop Grumman uses flythrough covers to cloak the shapes of launch vehicles and protect their frontends from environmental conditions. However, these covers are not reusable.

This team tested and modeled a reusable frontend cover and cable and spool retraction system, whereby the cover can be removed without recontacting the vehicle.

The cover has a clamshell-like opening and is made of a soft, durable, inflatable material. It protects the vehicle from high temperatures, wind, precipitation and humidity and can be used two or three times within a year. The easily assembled, operated and transported retraction system can be adapted to different launch sites and vehicle heights.

Being able to ensure correct volumes of staining fluid for specimens allows for more reliable testing. Roche Tissue Diagnostics wishes to test and analyze the viability of using strain gauges to estimate volumes of mostly liquid specimens on slides as they are stained in the BenchMark ULTRA staining system.

Because strain gauges typically estimate mass, or weight, this system must use a known density to reverse-calculate the weight.

In addition to the viability study, the team designed a device to fit within the BenchMark ULTRA for RICK v4.0 testing. Specifications included withstanding temperature and humidity fluctuations as well as the system vibrations during the staining without impeding the process or making it more difficult for technicians. The outcome was a system that employs a microcontroller, load cell with a four-strain gauge system running through a signal amplifier, and various methods of power regulation.