adidas is a global company rooted in sport with a primary focus in scalable 3D printed footwear. adidas strives to understand the relationship between mechanical and material properties and experiential and emotive metrics in 3D printed, lattice-based footwear components. The students worked to form the foundation of “Sensory Standards,” a library that helps to identify the relationship between the mechanical and materials properties of a footwear component and the emotive response it inspires.

Advanced Navigation and Positioning Corporation (ANPC) makes transponder-based aircraft navigation systems which allow for landing in remote environments and places without traditional airport infrastructure. The Transponder Landing System (TLS) has recently been scaled down from a shipping container form factor to the Small Footprint Precision Approach Landing Capability (SF-PALC) design which makes the whole package able to fit on a 463L pallet- a MIL-STD pallet which is used for transport of items of a variety of shapes and sizes on many different military vehicles and aircraft. The student team worked to design a way to airdrop the TLS from an aircraft in flight, while ensuring the system lands in a target area undamaged, upright, and ready for mission performance.

AeroVironment is a global leader in unmanned aircraft systems, tactical missile systems and electric vehicle charging and test systems. The company serves militaries and government agencies in the U.S. and more than 35 allied nations, as well as businesses and consumers. The student team worked to design a very small multi-rotor UAV and deployment mechanism that fits within the secondary payload bay of an AV Puma LE to help infantry personnel to scout and survey a variety of targets.

LOPAs are a standard way to represent the layout of passenger accommodations within an airplane, such as seats and lavatories. They are required by the Federal Aviation Administration to certify an airplane. The student team worked to create a web-based application that provides a seamless user interface for creating LOPAs – with a cloud-based LOPA storage and permission management system – to improve project creation, multitasking, and editing. The student team also worked to integrate part information from all fleets into an easily accessible display system.

Cold chain logistics refers to the transportation and handling of temperature controlled products such as groceries, pharmaceuticals, and chemical goods. Amazon’s current Freight Inbound (AFI) Network for cold chain delivery is dependent on third party shippers and managed by suppliers. This is not cost effective and prevents Amazon from managing their inbound cold supply chain. The student team created an automated optimization tool for inbound cold chain transportation network that processes the input variable and generates the most cost efficient execution plan for inbound deliveries based on constraints creating a more transparent, informed, timely, accurate and cost effective transportation process.

One of Applewhite Aero’s unmanned aerial, precision-delivery systems, the Baton, has two basic functions of payload delivery for critical supplies (e.g. blood plasma or anti-venom) and landing a sensor or sensors (e.g. a matrix of sensors such as temperature, sound, or video monitors). The student team worked to refine the design and increase the reliability, functionality, and utility of the system the Applewhite Baton, to help improve delivering supplies during search and rescue.

In many current smartphone retail purchase applications, retailers lack foot traffic, exposure to customers, and have difficulties reducing inventory, while consumers lack access to time-limited sale items. The student team worked to design and implement a web interface/retailer dashboard that allows retailers to manually send notifications to customers, and a corresponding customer application that allows customers to place orders.

Within the aerospace industry, transport erectors are used to maneuver a shuttle or rocket from a horizontal position into a vertical one. These highly loaded critical structures are often operated in coastal environments, and are therefore prone to corrosion. The student team focused on the bearings found within these complex systems, and developed custom test methods to observe the materials’ behavior when subjected to a corrosive environment. By studying the corrosion and wear response of metals that are commonly used in bearing systems, there is the potential for increased cost saving opportunities, not only in the manufacturing of the part but also in the maintenance performed throughout its lifetime. The student team simulated in-service conditions by carefully selected metals and subjecting them to heat treatment, corrosion, hardness, and tribology testing. From this data, in-service inspection intervals were developed to be used in the maintenance of a bearing part found on a transport erector.

Voice recognition is widely utilized for automatic fraud detection. The student team developed a cloud-based voice recognition security system that helps the bank staff verify new and existing members of BECU.

Boeing is interested in assessing the suitability of lattice structures to replace traditionally machined fittings for use in primary structures (fixed leading and trailing edge fittings). The student team worked to create a titanium structural fitting using structural optimization and lattice structures to replace a traditionally manufactured fitting. The team characterized 3D-printed titanium behavior and design, and analyzed and tested creative lattice-type and hybrid geometries to optimize performance.

Boeing Product Development needs a solution for the trimming of a green stringer charge. The stringer is in the cross-section of a T with the blade (aka web) pointing up. The forming operation leaves excess on both the base flange and the blade of the T which must be net-trimmed prior to the next operation. The student team designed a method to stabilize the flexible blade during the trimming operation in order to ensure a high-quality cut and high rate capability. The team also did a failure modes analysis for the design.

The student team worked with Boeing Product Development to explore a linear flow line for handling long, narrow beams that are positioned along the length of an airplane wing or fuselage (called stringers). Carbon fiber pre-preg material is laid up, trimmed, rotated, and transported on a universal handling tool (pallet) that provides indexing to each process cell. The team created a design for the pallet and indexing system, built a prototype that can accommodate material up to 6”x24’’, and demonstrated the ability of the handling tool to automatically connect to a vacuum system and secure and release carbon fiber material of a variety of sizes.

In order to help coordinate a response, provide the necessary situational awareness, and maintain open and reliable communications to a large wildfire or hurricane disaster area, a rapidly deployable fleet of autonomous unmanned aerial vehicles (aUAV) is necessary. The fleet of UAVs should communicate with each other and allow coordinated coverage of a large area, while allowing units to swap out when they need refueling. The student team worked to design a medium to high altitude, high endurance UAV, carrying a surveillance payload to provide the necessary information and assist with communications and observations.

Currently at Boeing camera footage of manufacturing processes are manually filtered through human supervision and require inspection for quality control. The student team worked to help Boeing apply computer vision techniques to automate defect detection and assist with quality control, and created an easy-to-use application to deploy these techniques onto local machines.

Thermoplastic matrix composites have gained attention as an alternative to thermosets due to their lack of shelf life, recyclability, and their ability to melt and resolidify. Aerospace companies such as Boeing are interested in repairing these parts using epoxy and other thermoset materials, but additional surface treatment is required due to their lower surface energy. Energetic surface treatment methodologies like atmospheric plasma, laser treatment, and UV light have shown promise as a robust method to increase the surface energy before structural bonding. Currently, these surface treatments are controlled by bulky robotic arm systems which are often impractical in terms of costs and accessibility to parts on aircrafts. The student team worked to design and build a prototype mechanical fixture/guidance system that is less expensive and able to attach to the aircraft in any orientation.

Recent advances in titanium powder bed fusion technology offer attractive aerospace design solutions. However, to offer additive parts as competitive alternatives to traditional casting and machined parts, materials and processes resulting in performance variations must be well understood for wider commercial usage. Research and development is needed to understand the influence of key process parameters on material properties. From Phase 1 study conducted by a capstone team last year, students performed in-depth literature research on powder recyclability and effects on performance. In Phase 2 this year, the new student team focused on understanding performance variations resulting from different material supply (same grade of the material but different powder fabrication method and different vendors). The student team built multiple sets of tension coupons from various titanium material suppliers and performed characterization and static testing. They characterized metal properties with respect to the material supply, the processing variations caused by the differences in supply, and root causes for variations in properties. Results will drive recommendations for material supply and process control parameters that optimize performance/quality.

Tooling at Boeing refers to unique structures and equipment built for very specific production needs (these are individualized and do not refer to traditional tools). Tooling fulfillment is currently inconsistent and fails to meet projected schedules in the 767 Line, resulting in an increasingly large backlog. The student team worked to improve communication, accountability, and order visibility in the Tooling Value System (TVS) to prevent and address tooling order delays.

The student team investigated the static properties of thermoplastic chopped carbon fiber composites and predicted how these properties change with variations in the orientations of the fibers. The results of this project give early information to Boeing to help streamline design processes for complex parts. With this data, Boeing can more effectively map out the alignment of complex carbon fiber parts to handle the unique stress conditions of these parts, and ultimately eliminate the need to test individual parts. Additionally, utilizing thermoplastic carbon fiber parts gives Boeing more opportunities for diverse, lightweight, and strong materials to manufacture quality airplanes.

Two years ago an interdisciplinary capstone team designed and built a three foot symmetric wing, then tested the wing in the UW low-speed wind tunnel. Last year a ME capstone team designed, but did not build, a 3ft asymmetric wing for use on a Formula SAE car. This team did extensive research on feasible welding techniques to join thermoplastic (TP) composite internal structure with TP composite skin. Building on the past two years of work, this year’s student team developed a feasible welding technique that the students can implement and characterize the strength of the weld. The results will contribute to building a welded thermoplastic wing for a Formula SAE car in the future.

The student team worked to establish a process for topology optimization (TopOp) and 3D printing of the upright and other parts of an electric formula-style car for the UW Formula Society of Automotive Engineers (FSAE). The upright rigidly transfers loads from the tires and brake calipers to the chassis. A design for the upright was finalized by a capstone team last year that achieved a 24% weight reduction, and a 300% increase in stiffness. This year’s team printed and processed the final design, validated loads, boundary conditions and material properties, and conducted testing to correlate finite element modeling (FEM) results.

University of Washington Human Powered Submarine is a student organization that races single-occupant submarines around timed courses. The student team worked to provide proportional, integral, derivative (PID) control of the submarine control fins to mitigate unintended roll at high velocities in races.

Autonomous Wheelchairs increase freedom and ease of mobility for the most vulnerable people in society. However, their autonomous navigation suffers from occasional loss of loop closure and localization. This can be disastrous in the real world either in a hallway, roadway, or sidewalk. The student team worked to develop a system that uses camera vision to check for lateral drift of the wheelchair from the desired path.

Left Ventricular Opacification (LVO) provides a means to better visualize the left ventricular cavity with contrast-enhanced ultrasound (CEUS) imaging (using microbubble contrast agents). Point of Care (POC) users are typically interested in delineating the left ventricle border for qualitative assessment of the ejection fraction and for visualization of the significant anatomical features of the heart in emergency or critical care situations, and would use LVO on patients whose body habitus makes this difficult with echo imaging alone. The student team worked to evaluate the FFSS LVO feature on the transthoracic echo (TTE) transducer for POC applications and suggested technical specifications for an LVO mode as well as an improved prototype setting.

Seattle Goodwill’s e-commerce has been expanding rapidly with increased growth expected. The student team worked to design a streamlined facility, improve existing processes, and create a flexible design that allows for expansion and future improvements to optimize the effectiveness of Seattle Goodwill’s e-commerce sector. The team created a process-based layout with a CAD drawn floorplan that significantly increases the items that are processed per month, improved the flow of items, and allows for future expansion.

The student team worked to develop an open source command-line utility and Application Programming Interface (API) to support single-command application deployment and service catalogs (chart repositories) in Nomad, HashiCorp's workload orchestration and scheduler.

The student team worked to explore behaviors and trends of web-based usage and specific user hardware attributes, to identify trends in user behavior that might be helpful for future CPU development.

Therapeutic proteins, an important class of drugs for many disease areas, are typically produced in bioreactor cell culture systems. Bioreactor outlet streams contain many non-product components of the cellular machinery, so therapeutic proteins must be isolated with downstream purification steps. Proteins are typically purified by liquid chromatography, where packed bed columns selectively remove the target protein or impurities from the complex bioreactor feed stream. Purification process development (PD) requires extensive experimentation for every new drug development campaign because the optimal chromatography process depends intimately on the surface chemistry of the target protein. The time and cost of purification PD could be reduced by replacing some or all of this experimentation with computational modeling. The student team worked to build and characterize inferential models to predict protein chromatography behavior from molecular properties.

Rod springs are widely used in aircraft door mechanisms to secure component and mechanism positions. Conventional manufacturing technologies use a spring and guiding mechanism, assembled in three parts. Last year a student capstone team demonstrated that it is possible to design a spring rod in one single element (without assembly) using 3D printing. This year’s student team manufactured, tested, and iterated the design, validated against FEA analysis, and developed a parametric design tool to aid in the design of family of similar springs. The testing covered compression testing, stiffness characterization, fatigue, humidity effects, and the effects of internal friction, hysteresis, and strain relaxation.

Currently aircraft primarily use flexible steel cables and coax cables for control systems; however, the data demand of these systems is increasing exponentially. Optical communication systems utilize light, transmitted over fiber-optic cables, to deliver signals between computers and offer many advantages over electrical systems, such as low power consumption and high data transfer rates. Single-mode fiber (SMF) is well-characterized but transmits only a single signal and is sensitive to vibration and thermal instability, so multimode fiber (MMF) would be more advantageous. The student team worked to characterize a SMF signal with a physical test bench that can determine linewidth, or a way to quantify the received signal’s power and quality in order to determine the respective qualities of both SMF and MMF signals and determine whether MMF is appropriate for aircraft applications.

Spherical bearings are used in mechanical design to address misalignment. These bearings are costly for a small component and have a complicated assembly process. Replacing a spherical bearing with a plain bearing means the capability to self-align is lost resulting in a poor contact distribution in the bearing (i.e., line or point contact). The student team designed a one piece bearing that is additive manufactured and has a contact surface that can accommodate the required loads under misalignments of up to 3° while still providing the required radial positioning. The team tested strength, bending, stiffness, and friction.

The student team worked to make state-of-the-art machine learning and artificial intelligence techniques more accessible to engineering domains by providing suitable benchmark datasets with interactive tutorials for advancing prognostics.

By some estimates, building operations and construction account for more than 50% of the carbon footprint of the United States. For decades, great attention has been placed on reduced energy use in buildings. However, in the last few years, awareness has grown that the carbon created in the construction of buildings (known as embodied carbon) is of critical importance. The Carbon Leadership Forum at UW has been a center of thought leadership in reinventing buildings structures and systems to have substantially lower embodied carbon. Mechanical, electrical, and plumbing (MEP) components of new construction buildings can result up to 10-15% of the total embodied carbon of a building design. The student team set out to find ways to reduce the carbon footprint of building MEP systems.

Water hardness plagues many commercial and industrial users; there are numerous disparate technologies to address the challenge. However, the performance of these technologies are rarely directly compared and never compared on a CAPEX/OPEX & levelized cost basis. The student team worked to investigate the performance of various water purification technologies for the purpose of determining which technology can best address water hardness issues, and worked to establish a baseline for four technologies: Reverse Osmosis, Chemical Treatment, Ion Exchange Resins & Electrodialysis Reversal.

Industrial equipment degrades with time, making it hard to predict when a potentially costly system failure will happen. In certain cases, equipment’s acoustic signatures change as a result of degradation. The student team worked to demonstrate the ability of Convolutional Neural Networks (CNNs) to detect various acoustic anomalies, help improve performance metrics, and perform inference.

Minecraft's game engine is built on an ECS architecture, but currently its network and replication system is outside of the ECS. The student team worked to build a sandbox multiplayer game with an ECS engine using EnTT game programming library. The team also built a replication system around that game and recommend a design pattern for an ECS replication system for Minecraft that assumes a reliable data transfer protocol.

Minecraft has over 131 million active users each running more than 300 mob entities per game. Mob entities compete for resources and produce high load on the Minecraft server. The student team worked to optimize mob behavior by implementing AI algorithms to improve game performance and reduce the server cost usage.

In the world of Minecraft, there are many different types of mobs with distinct mob sizes and moving abilities. The current pathfinder in Minecraft has a very high time complexity in helping a mob with sizes larger than 1x1 find the optimal path to its target. In addition, Minecraft is limited in the types of blocks they can add due to limitations in the pathfinder. For example, vertical slabs create unique challenges that the existing blocks do not have. The current pathfinder in Minecraft cannot find the optimal path if the optimal path involves partial blocks. The student team worked to run pathfinder with various mob sizes with partial blocks, determining that Big O runtime is not affected when all search notes are generated before pathfinding.

Minecraft Bedrock edition is currently utilizing “LevelDB” to save worlds data. This works well, however Minecraft is pushing the technology to the limits. The student team worked to develop a testing framework in C++ that can be used to test and evaluate key/value store implementations in order to allow future Minecraft developers to easily experiment with “LevelDB” replacements.

MilliporeSigma currently uses two lab technicians to perform sampling, which involves holding the bags, sampling the media and returning the bag to the holding bin. The student team worked to develop a solution to improve the long-term efficiency and reduce the number of lab technicians by assisting them in the bag handling process without causing cross contamination. The team built and tested a final prototype featuring an ambidextrous handle and ratcheting clamp that is used by a single lab technician, supports media filled bags upright without spilling, and makes media sampling easier.

The student team worked to help the UW Mobility Innovation Center (MIC) and Sound Transit create a cost-effective electronic system to gather data about the utilization of parking spaces and to share more accurate information about parking availability.

Moonbeam uses virtual reality and artificial intelligence to transform the future of remote creative collaboration. They build tools to help corporate innovation groups work more effectively with startups without the constraints of geography. The student team worked to develop features to support remote collaboration through diagramming within VR.

MWD Technology & Innovation Center LLC & CalRamic Technologies LLC specialize in processing low and high voltage ceramic capacitors. These companies seek to push the limits of current class 2 dielectrics by processing dielectrics operable above 225˚C with greater than 10 M-ohms of resistance. Such devices would benefit aerospace systems, satellite systems, military systems, geothermal detection devices, etc. The student team worked with current MWD BaTiO3 dielectrics, tested the effects of compositional changes and modified sintering conditions on the device resistivity using electrical impedance spectroscopy with a goal to raise operating temperature for BaTiO3 dielectrics in capacitors without reducing charge storing ability and to build a framework for future efforts.

The Cooperative Autonomous Distributed Robotic Exploration (CADRE) project led by NASA is an initiative focused on the creation of a network of shoe-box-sized mobile robots that could enable future autonomous robotic exploration of the moon, Mars, and beyond. The student team worked to create a software stack to facilitate joint navigation between multiple robotic operators. Additionally, the team created design proposals for improved mobility and charging by creating an improved robotic tail, creating a charging platform, and creating a circuit designed to characterize solar cell efficiency, taking into account dust and debris accumulation on the cells.

Chronic illness, defined as any disease that requires at least one year of ongoing medical attention, affects half of all American adults and at least 7% of American children, and is particularly devastating in underserved communities in the US where they are contracted at a higher rate than the national average. The economic burden of chronic illness, such as diabetes, obesity, and cardiovascular disease, in the US is unprecedented. Chronic illness also imposes great psychological distress and reduces quality of life for diagnosed individuals and their loved ones. There is growing evidence showing positive outcomes in self-management of chronic diseases with behavioral interventions. However, many of these solutions require active monitoring and burden the patient and caregiver with the need for constant monitoring. The student team tackled the needs of young Type 1 Diabetes (T1D) patients to help manage and monitor symptoms to reduce daily stress and burden associated with T1D. Beta Watch is a medical storage accessory that can attached to a mobile phone and safely and discretely houses all daily essentials of T1D care. Paired with an educational, T1D tracking mobile application, Beta Watch helps T1D patients manage the burdens associated with the chronic illness.

Hypertension and diabetes mellitus are both common conditions that have a high prevalence of co-existence, and are also highly associated with obesity. These conditions are considered risk factors for many diseases, including coronary artery disease, cerebrovascular disease, renal failure, and congestive heart failure, so treatment of these conditions is essential. Monitoring blood pressure is a key factor in managing the co-existence of these conditions. Blood pressure is currently measured using inflatable sphygmomanometer; however these measurements need to be made by a health care worker or at home using an automated tool. Blood pressure can vary greatly over time due to various influences such as diet, stress, body position, temperature, and a wide variety of other factors. For this reason, it is difficult to obtain an accurate baseline blood pressure reading and understand how various interventions and lifestyle choices impact blood pressure. The student team worked to develop a method for unobtrusively measuring blood pressure several times a day to provide accurate information to clinicians and researchers aiming to reduce the hypertension and improve outcomes for chronic illnesses.

Nutanix is a cloud computing company that offers a variety of services for on-premises and hybrid cloud deployments. Nutanix offers a disaster recovery service, Xi Leap, which utilizes Nutanix’s own data centers to provide reliable storage for disaster recovery. Presently, the Nutanix data centers do not have a central solution for power management. To perform even a trivial power cycle operation, they must file an IT ticket, which is costly in terms of time and human labor. The student team worked to design a system as a microservices type application, developing on local Kubernetes cluster, to provide workload isolation and scalability, including processing incoming requests, securely communicating with devices in the Nutanix data center, and returning a response to the user.

Reliable visual perception plays a critical role in enabling autonomous vehicles to safely navigate unseen, unstructured environments. In order to anticipate and avoid obstacles, such a perception system needs to detect, classify, localize, and track dynamic objects within range of the vehicle. Many perception systems in state-of-the-art autonomous vehicles rely on LiDAR (light detection and ranging) to produce an accurate geometric representation of the vehicle’s environment; however, such systems can be costly to acquire and maintain. The student team worked to create a system to that helps with 3D object classification, detection, and tracking of dynamic objects to improve motion planning and navigation in autonomous driving.

Many autonomous vehicle systems in development today rely on traditional color video in addition to LiDAR and/or Radar information to identify and localize objects surrounding the vehicle by utilizing Deep Neural Networks. Unfortunately, some classes of objects such as pedestrians and cyclists can be difficult to identify. The student team worked to analyze the possibility of using thermal imaging data instead of camera data, and developed a Neural Network to optimize performance.

To accurately model the dynamics of a tractor semi-trailer, it is important to accurately model joint stiffness within the chassis and attached components. There are several ways in which joints can be modeled in simulation, but it is difficult to say which is most correct for stiffness without physical testing. The student team planned and conducted physical tests of some typical joints found in a tractor chassis to measure their stiffness, created finite element models to replicate the tests, and compared simulation and test results to determine the most accurate way to model the joints. They ultimately analyzed test results to choose the best joint model(s).

The student team worked to select a new composite material for the roof structure of the Peterbilt 579 Ultraloft truck manufactured through PACCAR. The goal of the project was to select a composite roofing material that will provide better mechanical properties than the current randomly oriented glass fiber sheet molded composite (SMC) while not spending more than $3 per pound of weight saved in overall structure. The team conducted mechanical testing of selected carbon fiber composites to compare to the original SMC material, performed a cost analysis on the manufacturing process for the new composite, and conducted a roof crush test simulation through finite element analysis (FEA) to demonstrate how the new composite will meet and exceed the properties of the old material.

Residual stress can lead to premature failure and reduced durability of bonded joints. Most test methods for predicting long-term performance of bonded joints decouple aging and environmental conditioning from constant- or cyclic stresses. Other test methods that apply stress during environmental conditioning do not record stress during tests and are susceptible to changes in stress as the fixture and load-train thermally expand and contract. The student team worked to develop a test fixture capable of applying, measuring, and adjusting stress in bonded joints during exposures simulating thermal-shock, high-temperature processing, and environmental aging.

With the onset of COVID-19, user testing for PACCAR trucks significantly decreased. The student team worked to improve the user testing process, meet safety guidelines, and practice social distancing to allow truck testing to resume remotely. Their solution was required to be completely remote, flexible, maintain WiFi connection at all times, and not distract the driver.

Currently robots’ movements are not very maneuverable for traveling in multiple directions. The student team worked on building a prototype for a self-balancing spherical robot that can be more flexible, agile, and overall, more maneuverable.

To help aid effective decision-making following a meeting, as well as extensive note taking during a meeting, the student team worked to increase the accuracy of ngMeeting, a meeting transcription system that records in-person and telephone meeting conversations to make spoken communications close to real-time, accurate, and easily searchable and actionable.

The student team worked to write a transportation electrification plan including solar generation for UW Transportation and for Recology, as well as develop a software tool to schedule and allocate charging for an EV fleet that meets operational requirements while optimizing the demand for solar grid capacity. The software tool and transportation electrification designs will help UW transportation achieve its goal of reducing the university’s carbon footprint to 45% in 2030 and help Recology electrify its fleet efficiently in the near future.

Sironix Renewables is a growing start-up company exploring manufacturing routes to a new line of plant-based surfactants. These high performance green surfactants include personal care, industrial, and agricultural applications, among others. Sironix is exploring multiple synthetic routes to produce materials in an economically-viable and environmentally-responsible manner. In one iteration, a patented series of reactions is carried out to produce the surfactant precursors, followed by a series of separations and purifications, recycling of solvents and unused reactants, and treatment of waste streams. The advantage of this specific process is a facile synthesis of the desired materials, while drawbacks include the use of multiple solvents and the production of several waste streams that must be mitigated or disposed of. While the technology and practice of the reactions for this route are well established by the company, a thorough design and analysis of the downstream separations required for an industrial scale production based on this technology needs to be conducted. The student team worked to evaluate separation methods to identify one or more suitable purification methods for each of the separation challenges in this process, optimized the separation train(s) using a process simulator, conducted an economic analysis to determine the most beneficial separation method, and incorporated the economic analysis with provided technoeconomic data for the overall process, to determine the financial viability of the process as a whole.

Time and motion studies are the basis of determining how much labor is required to staff Starbucks stores to support customers and inform business decisions on new beverage, food and equipment programs. While important, time and motion studies are resource intensive, and can be unreliable and prone to error. The student team worked to prove the feasibility of a system that could accurately recognize one key object manipulation activity in the Starbucks store and time how long it took to complete by utilizing transfer learning with machine learning models to classify human actions with video data.

Synchronized cardioversion (sync cardio) is a frequently used therapy for the treatment of potentially life-threatening arrhythmias such as atrial fibrillation and pulsatile ventricular tachycardia. Sync cardio is a therapeutic electrical shock delivered at a precisely controlled time point in the cardiac cycle. Incorrect shock timing can lead to ventricular fibrillation (VF)—a lethal arrhythmia. Devices that perform sync cardio can also deliver unsynchronized defibrillation shocks and have different operational modes for defibrillation and sync cardio. Being in the wrong device mode for the intended therapy can have major adverse consequences leading to patient death. For example, applying a defibrillation shock to a patient who should have received sync cardio can lead to VF. On the other hand, being in sync cardio mode when a patient is in VF can prevent the device from delivering a lifesaving defibrillation shock. Decades of post-market device surveillance indicates that such adverse events are rare but continue to occur. The student team worked to develop a novel monitor/defibrillator feature that intelligently selects the appropriate device operating mode, and either advises the operator or automatically switches to that mode. The greatest lifesaving impact is expected where healthcare providers are either inexperienced or need to use a variety of device models.

Capnography is the waveform that shows how much CO2 is present at each phase of the respiratory cycle. End-Tidal CO2 (ETCO2) is the partial pressure of CO2 detected at the end of exhalation. The value is normally 35-45 mmHg. Existing capnography devices utilize side stream technology, which is inconvenient, messy, and bulky, and does not provide Emergency Medical Technicians (EMTs) with the most helpful, real-time feedback. The student team worked to create a portable capnography device that is capable of measuring End-Tidal CO2 from the exhaled breath of patient, as well as inspiratory and expiratory flow rate and airway system pressure so that EMTs can know, real-time, whether they are ventilating a patient properly.

The student team worked towards designing a wireless pulse oximetry measurement device capable of measuring SpO2, SpCO and SpMET. The project is a continuation from a project last year in 2019/20. This year’s team iterated on last year’s device by refactoring and simplifying the system design to make the device more compact and suitable for use in real medical settings.

Achieving customized functionality on NB-IoT (Narrowband Internet of Things) devices can be time and resource intensive. Creating a single, customizable device would essentially eliminate the development process and give the user great flexibility. The student team worked to create an end-to-end NB-IoT device which communicates over the T-Mobile narrowband network. Deliverables included a frontend web application, a backend for data storage, an IoT portion, and interconnections that execute automated functions.

SIM card usage in drones is increasingly popular because it provides longer range (cellular) connection than Wi-Fi or Bluetooth. The problem with this trend is that base station antennas are optimized for terrestrial coverage and drone SIM usage can cause interference as well as expose networks to security issues. Technology exists to accommodate for SIM usage in a drone, i.e., telecom providers offer drone-specific data plans. The student team worked to use a logistic regression model to help T-Mobile find a way to identify customers using SIM cards on cellular data plans to connect those clients to the appropriate technology for their needs.

Talking Rain is a fast-growing beverage company known for their popular brand, Sparkling Ice. They are interested in reducing costs in their supply chain by scheduling fewer unnecessary refrigeration trucks and implementing a centralized kitting system. The student team built an optimization model that finds the optimal city to locate the kitting facility to minimize total transportation costs.

The Talking Rain Beverage Company is the creator of the beverage Sparkling Ice. Currently, Sparkling Ice is created directly from raw material at both the Talking Rain facilities and copackers. This process could be improved using a concentrate kit which would decrease cost, processing time, error rate, and required storage space. The student team worked to calculate the required storage space for the current demand, find the feasibility of implementing the concentrate kit system, and give concentrate implementation recommendations.

The TE Connectivity Automation Manufacturing Technology (AMT) team is looking for innovative designs for a rotary insertion machine. An insertion machine is used to insert metal contacts into a plastic housing to form a connector part. Most insertion machines today have horizontal or vertical motions, which involves a change in direction after every insertion. A rotary insertion machine would not have to change directions, increasing the speed of the process. This rotary design would enable potentially higher speeds of the insertion process than what is achievable through linear motions. The student team designed prototypes, analyzed them using simulations, and created a final design model in CAD of the rotary insertion machine.

TE Connectivity designs and manufactures connectivity and sensor parts. The Automation Manufacturing Technology team is looking to minimize total inventory and production costs on their Universal Mate-n-Lock manufacturing processes. Currently, connector assembly machine schedules do not include low volume parts resulting in a sub-optimal process. The student team built a tool to improve demand forecasts based on historical values, recommend optimized sequences for scheduling, and recommend minimum order quantities for low volume parts.

People use their voices for in-car navigation systems. However, the system might have trouble processing exactly what the user says, depending on noise from surroundings and accents. The system will come up with a few strings for what the user could have said. For example, if the user said “university,” then the system’s voice processor might hear two possibilities: “university” and “universe city.” The student team worked to rank these possible queries in order of likelihood of what the user actually meant in order to help increase in-car navigation system accuracy.

Toray is a cutting edge carbon fiber composites manufacturing company continuously developing high performance materials for companies like Boeing whose aircraft have increasingly utilized composites for their high strength to weight ratio. The composite development Toray is involved in generates vast amounts of data that is currently only being stored in Windows files system, with cross-experiment analysis only possible through Excel. The student team worked to create a data management tool to improve data accessibility and streamline cross-experiment analysis. The solution is capable of storing raw data, recording extracted values from data sets, and analyzing sample sets based on designated parameters, all through a user-friendly UI. The database was designed following relational database normalization forms to reduce redundancy and ensure data integrity. This improved data management has the potential to be the groundwork for advanced data science technologies at Toray which are increasingly becoming essential in the research and development of modern materials.

Tupl, Inc. is a technology company that provides technical services and support to a variety of companies. Tupl sought to design an intelligent chatbot that can accurately and efficiently address the needs of a phone company’s customers. While the chatbot is usually powered by natural language processing and machine learning, the student team worked to have the chatbot identify user intent from text input, as well as classify a user’s problem, respond with logical answers, and take correct actions according to different customers’ situations.

The University of Washington Applied Physics Lab (APL) and NASA Jet Propulsion lab have designed melt probes to study water under thick layers of ice on “Ocean Worlds” of the solar system such as Europa. The problem with the current probe is that there is no way of determining the melt jacket thickness around the probe, while it is deep under the ice. This data is required to prevent the probe from being trapped due to water freezing around it. The student team worked to design a hardware and software solution that would measure the melt jacket thickness, using piezoelectric transducers.

Currently, the College of Engineering lacks a modern tool to manage requests for administrative services, specifically financial transactions. The student team worked to design and develop a web application that serves as an advanced ticketing tool for UW College of Engineering departments to manage fiscal related tasks (reimbursement, travel, purchase, etc.) from users' request submission through fiscal staffs' approval or denial.

In the immediate aftermath of a mega-quake, all disaster response agencies and personnel will be completely overwhelmed. External help (primarily from the government) is unlikely to come immediately after a disaster. If neighborhoods rely mostly on such help, as is currently the status quo, it is likely there will be more fatalities and greater economic loss. As a result, neighborhoods are urged to prepare for community-based survival for up to three weeks. This means that residents must be able to share useful information, carry out essential activities (e.g., staying cool/warm in summer/winter, securing food), and use effective socially integrated technological solutions to enhance their ability for survival and real-time response. The student team worked to develop technologies that enable real time information gathering and sharing (while safeguarding privacy), and solutions for efficient resource matching by leveraging social ties.

The Particle Cyclotron creates proton and neutron beams to remove tumor cells in cancer patients. As more cyclotrons have moved towards using proton beam radiation therapy, the cyclotron at UW Medicine is the only one left in the US that uses neutrons to kill cancerous cells. Having been developed in the 1980s, the cyclotron has many components that are now outdated and that require redesign consideration. The cyclotron’s Beam Current Amplifier (BCA) System works by receiving feedback from devices that narrow beams and particles throughout the system, known as collimators. The BCA System receives the collimators’ current when the particle beam is misaligned and represents it using a digital dial. Operators use this information to tune the trajectory of the beam. The student team worked to redesign the BCA System, now called the Beam Current Signal Conditioner Card, from an analog system to a digital system. This modernization and redesign helped with display output feedback precision, machine bulkiness, spacing and safety replacement issues, and also reduced the number of parts required to run the system.

Ransomware attacks are becoming more prevalent in the world. Some ransomware is designed to avoid computers with certain configurations. The student team worked to find common configurations that these ransomwares check for and to exploit them as a technique to obstruct ransomware encryption. They also worked to expedite the testing process to allow for rapid testing in the future.

The hydraulic infrastructure or assets maintained by the state of Washington Department of Transportation (WSDOT) such as sewers, pipes, and culverts, need frequent monitoring for upkeep. The student team, building upon a platform from last year’s team, worked to improve a remote controlled monitoring device that is able to access difficult to reach or tight spaces.

Whooshh Innovations, Inc. designs novel approaches to transport fish over natural and man-made barriers. Whooshh has successfully used a portable, modular flume fishway that mimics a highly turbulent stream channel with three sections that, when linked, entices salmon and other strong swimmers to swim up an elevation of 4-6 feet and places them at the entrance of the fish passage system. The student team worked to design a new portable, modular flume fishway that is passable by fish with differing swim behaviors and abilities (non-jumping, weak swimmer), to expand the use and application of Whooshh’s technology. The team analyzed several possible designs using 2D and 3D computational fluid dynamics and studied the effects of several design parameters on the flow conditions. They also analyzed the manufacturability of the designs and tested flow conditions with a scaled model.

Similar to the way Harry Potter Marauder’s map functions in the popular and well-known literature series, this student team worked to provide a multi-target multi-camera tracking system to help improve the security of a mapped space as well as the tracking of an individual within a mapped space. Some examples of how this technology can be used include tracking at risk individuals in a long-term care homes or detecting unauthorized intrusions.

Package theft is a problem that negatively impacts the public especially during a pandemic situation where the demand for online shopping and package delivery increase tremendously. The student team worked to build a machine learning based software system that can identify package theft and improve home security levels for families and companies.