Industrial Capstone Senior Design Spring Showcase

Sponsor: University of Massachusetts Lowell

Project Description: This project works to develop a machine learning algorithm to detect birds near wind turbines. The detection capability would serve a twofold purpose. One, to collect data on the frequency and nature of the interactions between birds and wind turbines. Two, to possibly prevent birds colliding with wind turbines; thereby reducing the likelihood of injuring the bird or damaging the windmill. The system should be capable of separating instances of flight paths that do result in the bird nearing the windmill and a flightpath that does not. Once a relevant flightpath is identified, the camera will save the recording of the bird. A stretch goal of this year’s project is to expand the capabilities of the system to detect bats as well as birds.

Mentor: Jean-Francois Millithaler, Electrical and Computer Engineering

Sponsor: Collins Aerospace

Project Description: Collins Aerospace has a machine learning solution to generate bounding boxes around objects of interest in aerial images taken on their DB-110 sensors. They desire a solution that uses semantic segmentation, which identifies objects at the pixel level. This would provide more information about the object such as shape and orientation. We aim to develop a solution that operates in two modes: one which performs pure semantic segmentation based on a pre-trained model, and another which takes the existing bounding box data as input and produces semantically segmented data. The solution is to be developed using the DeepLab deep learning platform and delivered via Docker container.

Mentor: Kavitha Chandra, Electrical and Computer Engineering

Sponsor: General Dynamics – Bath Iron Works

Project Description: Modern shipbuilding is a highly precise and time-consuming process. Delays caused by time-consuming measurements, mistakes, and the required rework are expensive. The client believes these delays can be minimized by Augmented Reality (AR) technology. Shipbuilding requires a component assembly tolerance of 1⁄8”. The client desires a Consumer-Off-The-Shelf (COTS) mobile device that has a compatible application for an end user to provide real-time, AR overlays to increase the accuracy of component assembly within this 1⁄8” tolerance. The objective of this project is to research the state of AR in current industries and improve the accuracy of BIW’s current AR capabilities to have a maximum deviation of 1/8th of an inch by May 2020. The deliverables of this project include research reports which detail the state of AR and potential ways ahead and improvements to BIW’s application, InfoTagger.

Mentor: Michael Geiger, Electrical and Computer Engineering

Sponsor: Security Innovation

Project Description: This project will develop an enhanced version of the cyber security training courses made by Security Innovation. The enhancements will help to make Security Innovation’s courses more engaging and effective. The group will combine their personal ideas for improvement with those of students in relevant fields, who have volunteered to take and evaluate the courses. 

Mentor: Jay Weitzen, Electrical and Computer Engineering

Sponsor: Schneider Electric

Project Description: The Industry 4.0 initiative is the enabling technologies for the factoriesof the future. It includes concepts of Smart Manufacturing/Factory andIndustrial Internet of Things which are possible with the proliferation ofprocessing power and networking at all levels. Two technology areasthat are core to Industry 4.0 are OPC UA protocols and Time SensitiveNetworking (TSN). Together these technologies allow migration from previous generation custom protocols and network technologies to

systems that are entirely based on open standards.

Mentor: Robert Marceau, Computer Science

Sponsor: Red Hat

Project Description: A project to add a Trello-like tool to view and modify content view versions for Foreman/Katello configurations.

Mentor: James Daly, Computer Science

Sponsor: Anokiwave

Project Description: The main objective, using Anokiwave’s own AWMF-108 beamforming integrated chip, is to design and build, on a PCB, a 1x4 phased array antenna and, if time allows, a 1 x 8 phased array antenna. The design will be constructed using Autodesk Eagle. The antenna will operate in the 28 GHz range which can be utilized for up-in-coming 5G communications.

Mentor: Huliang Zhang, Electrical and Computer Engineering

Sponsor: Raytheon

Project Description: Millimeter-wave phased array technology constitutes a high frequency antenna network capable of producing an electrically scannable beam of RF energy.  Constructive interference between the array’s antennas help to form a concentrated beam and built-in phase shifters allow the beam to point in a desired direction.   Phased array prototypes were built and tested to reach an optimized design using 64 individual patch antennas.  Simulation and test data were considered when designing more efficient array antennas and MMIC phase-controlling circuitry.  The project’s long-term goal is to build a fully functional phased array capable of directing a beam of RF energy with I/O capabilities.  

Mentor:  Corey Shemlya, Electrical and Computer Engineering

Sponsor: Skyworks Solutions, Inc.

Project Description: Due to the high frequencies used in 5G communications, between 3.3 and 3.8 GHz, signals become weak over short distances. To combat this, high-gain/low-noise amplifiers (LNAs) need to be designed to boost the signal without losing any data. These amplifiers and receivers are multistage components that also have filters and various LNAs to create systems that are as linear and ideal as possible. They are also built to provide support for the maximum number of carrier aggregation frequency bands. This team will design and test a receiving chain of LNAs to meet these goals.

Mentor: Joel Therrien, Electrical and Computer Engineering