2018/2019 Electrical & Computer Engineering Industry-Sponsored Student Capstone Projects

Accolade is a personalized health and benefits solution providing personalized advocacy. The student team developed a system to enhance user experience and interaction with physicians by sharing users’ health data from Apple Health Kit with the Accolade platform.

Hospitality venues around the world are switching form staffed cashiers to Self Service Kiosks. Students worked to develop a software plugin that utilizes network voice recognition services to provide voice-to-text and text analysis for guest contact with hotel employees.

Axon develops technology for law enforcement, self-defense, and the military, with the goal of making communities stronger, safer, and more connected. The student team developed two neural network systems for text-independent speaker identification of law enforcement officers and other authorized speakers.

Hearing loss is the third most common physical condition affecting people of all ages. Booz Allen Hamilton is creating Bluetooth devices as hearing aids. The team worked to design, build, and test a system using a single host serving multiple Bluetooth audio devices which stream the same audio.

Collins Aerospace is looking for a solution to detect level in an aircraft water/waste tank in real time that can inform flight attendants. The student team designed, built, and tested a non-intrusive continuous level sensor and communication system for aircraft waste systems.

The aerospace industry is increasingly looking into electric power options. The student team developed a digitally-controlled synchronous battery charger for avionic applications. The converter acts as both a battery charger and bus-sustainer upon loss of power.

Conventional Visual SLAM based Teach & Repeat Self-Navigation relies on expensive LiDAR for operation in large scale environments. Cyberworks has developed “core” code for two novel methodologies of Teach & Repeat that use only a cheap webcam, allowing cost sensitive products like autonomous wheelchairs to become accessible to the masses and marking a shift in self-driving technology. The student team developed and integrated a robust visual odometry package with sensors, an onboard processor, and a user interface to assist autonomous wheelchair motion.

Eagleview is a technology provider of aerial imagery, data analytics, and geographic information system solutions. The student team developed a system that uses deep neural networks and optical character recognition techniques to detect and read oblique signage in aerial imagery.

Fluke is a manufacturer of industrial test, measurement, and diagnostic equipment. The student team developed and tested a system for accurate identification and monitoring of different electrical loads.

Glympse delivers last-mile solutions for delivery, curbside/click-and-collect, BOPIS, and roadside assistance creating a smooth customer experience when people, products, and services are in motion. The student team developed a system using GPS location data to detect and categorize dangerous driving maneuvers and assignment of a driver safety score, with the goal of improving delivery driver safety.

Truck manufacturer, Kenworth, constantly seeks to improve truck safety. The student team worked to design, build, and test a system to cancel turn signals automatically for semi-trucks based on CAN (Controller Area Network) bus signals and image processing.

Kernel Labs’ expertise is in machine learning, computer vision, and security. The student team was tasked to design, build, and test a cost-efficient general purpose rover for farm use. The rover is remotely operated and built with a machine learning framework for implementing future applications.

Lightning Network has created the next generation decentralized global payment platform. The student team developed a payment app utilizing the Lighting Network to increase the rate of completion of cryptocurrencies to meet consumer needs.

Magic AI is building a platform through machine learning and computer vision that will bring ubiquitous computing to industries and implementations previously excluded, specifically focused on nonverbal forms of data and communication related to animals and humans. The student team developed a data pre-processing pipeline and classification neural net to individually identify horses.

Microsoft knows that tomorrow’s workforce will require both domain-specific knowledge and coding skills, but there is currently a lack of learning opportunities and hands-on experience necessary to inspire young learners, particularly those in underrepresented groups. The student team worked to develop a hands-on data science curriculum which aims to make K-12 learners data literate and introduce them to data science topics such as data visualization, statistics, and machine learning.

Mindshare Medical creates evidence-based clinical decision support technology providing personalized diagnostics and guidance on the effectiveness of follow-up procedures and treatments with the goal of improving quality of patient care and reducing financial liability for healthcare organizations. The student team designed and built a neural network system for localizing and characterizing breast cancer lesions in 3D tomosynthesis scans with significantly reduced false positives.

NASA PUFFER (Pop-Up Flat-Folding Explorer Robots) are small, low cost, expendable rovers that can transport key sensors and instruments to locations considered too risky for the primary lander, rover, or astronaut. Because of the high communication latencies of deep space missions these rovers must minimize their dependence on ground control and be able to operate primarily autonomously. The student team worked on a multi-robot navigation and search for the PUFFER, including computer vision, LIDAR distance sensing, and SLAM, to cooperatively find a goal object.

The Puget Sound is seeing a declining population trend in Puget Sound species of zooplantkton. NOAA seeks to analyze the population by using an underwater camera at night that attracts zooplankton into funnels with light. The student team worked to develop a smart light trap to capture video of zooplankton and their surrounding environment, and developed a software package to analyze the video, giving scientists a view of the composition of species and a count of specimens in a given trap.

Radar is one of the most powerful sensor technologies available for automotive perception systems. As part of an automated driving system, radar is partially strong in all-weather operation, low cost, proven reliability and the ability of provide fast velocity measurements. A challenge in using radar is noisy data due to multipath returns, as a result a lot of effort is needed in digital signal processing to improve the reliability of the detections. Both false positives and false negatives are a concern for automotive applications. The student team worked to apply commercial radar technology to a heavy duty commercial vehicle for forward based object detection and tracking, and to create maps of the environment surrounding the vehicle, including moving and stationary objects.

Mosquitos carry deadly viruses including malaria, Zika, dengue, and West Nile virus. Agricultural pest ruin crops, both outdoors and in greenhouses. Current prevention techniques for both problems are costly and marginally effective. Photonic Sentry has created a laser based insect monitoring and eradication system called Photonic Fence. The student team developed a system to detect irregularities in object detection and target tracking of bugs in the Photonic Fence.

Pollen Systems harvests data for agricultural growth, helping farmers manage their crop growth and vitality, identify irrigation and moisture issues, and note pests and disease problems using drones, data analytics, and proactive alerts. The student team developed a semi-autonomous ground vehicle which surveys vineyards and other high-value crop fields, with a focus on collecting pictures to evaluate expected yield and plant health. This ground vehicle will augment Pollen Systems’ current services.

Seattle Art Museum is seeking an art base that can support sculptures and prevent damage during earthquakes and other environmental vibrations. The student team developed a system to actively stabilize a sculpture subject to seismic forces, using sensors and motors.

Sound Transit seeks accurate traffic counts, duration of stay, and other metrics in Link Light Rail stations. The student team developed a machine learning system which combines image object detection and tracking to produce aggregated mobility data and traffic counts in Link Light Rail stations using station security camera footage.

Modern smart phones are highly sophisticated devices with multi-core CPU and GPU, multiple sensors and transceivers that can be used in the field to provide situational awareness using signals of opportunity. The student team developed a machine learning Android application that finds the best driving path between two locations based on optimal weather conditions and other user requirements.

Steth IO has built a smartphone stethoscope to improve patient experience and outcomes and make exams more objective. The student team designed and implemented a single lead electrocardiogram (ECG) system for integration with Steth IO’s digital stethoscope, and machine learning classification of ECG data.

Sweet Tea Cancer Connections works to forge connections and build relationships among families affected by pediatric cancer using smart technology, national partnerships, and focused research. The student team designed and built an app to connect parents who have a child with pediatric cancer with their child’s caregivers.

Billions of dollars in damage is caused every year in the US from fires originating in the electric grid, many of these caused by growing vegetation causing a short. The student team developed an NB-IoT-enabled device to detect obstructions near power lines and report them to a remote web server, assisting in efforts to prevent wildfires and save lives.

Globally, there are limited commercially available power chair platforms that can exhibit socially responsive automated behavior, such as autonomy in negotiating through crowds or facilitating appropriate navigation behavior when a having a conversation with a friend. The student team developed a smart sensor suite to be used in an autonomous wheelchair, enabling context awareness and path reconstruction.

Telenav is a wireless location-based services corporation with a goal to build the largest network of connected cars on the road. The student team designed a machine learning model to predict a user’s destination continuously during a trip, with the flexibility to adopt time/distance transitions and spatio-temporal data.

Tupl develops software with state-of-the-art big data technologies. The student team worked to improve prediction and visualization of customer and network issues with deep learning using call detail records, key performance indicators, and Twitter datasets.

The University of Washington Medical Cyclotron Facility is a one-of-a-kind cancer treatment and research facility. The student team worked to design, build, and test a motion control system that tunes the Cyclotron radio-frequency system. Tuning to different frequencies allows the Cyclotron to be used for cancer treatment, isotope creation, and advanced materials testing.

Verathon’s current BladderScan devices are used by healthcare providers at the point of care to perform automated measurments of bladder volume. Clinicians use bladder volume measurements to determine whether to catheterize a patient. The student team developed a wearable device to periodically monitor bladder volume and send phone alerts when detecting high volume readings.

WatchGuard products are designed to protect computer networks from outside threats such as malware and ransomware. The student team designed and built a module for storage and computation of network data to be integrated with existing technology to provide user with more information on network traffic.

Witekio specializes in embedded software with a system level approach, and engineers and integrates intelligent systems software for any device from hardware to the cloud. The student team worked to design and build a secure over-the-air firmware update system for ARM Cortex-M4 microcontrollers that enables device maintenance and management.

Zetron is a critical communications technology company that provides integrated command and control solutions and information services for public safety, transportation, utilities, natural resources, healthcare, and academic institutions. The student team was tasked with development of an Internet of Things environmental sensing and communication system for fire rescue operations.