Engineering Commons^™

A Study Of Deep Reinforcement Learning In Autonomous Racing Using Deepracer Car, Mukesh Ghimire

Honors Theses

Reinforcement learning is thought to be a promising branch of machine learning that has the potential to help us develop an Artificial General Intelligence (AGI) machine. Among the machine learning algorithms, primarily, supervised, semi supervised, unsupervised and reinforcement learning, reinforcement learning is different in a sense that it explores the environment without prior knowledge, and determines the optimal action. This study attempts to understand the concept behind reinforcement learning, the mathematics behind it and see it in action by deploying the trained model in Amazon's DeepRacer car. DeepRacer, a 1/18^th scaled autonomous car, is the agent which is trained …

Coordinated Autonomy Of Unmanned Aerial Vehicles (Uavs), Paribartan Dhakal

Honors Theses

Unmanned Aerial Vehicles (UAVs) are being extensively used in diverse sectors of the society for various tasks ranging from videography to an extremely sensitive situation such as first responders helping during a disaster. It has been seen that if a fleet of UAVs is deployed, they can perform a task quicker and more efficiently than a single UAV. With an increase in the number of UAVs, a problem arises of handling them with proper control structures. It has been studied that the Behavior Trees (BT) can be a better control architecture to handle the autonomous vehicles, as BTs are more …

Exploring The Behavior Repertoire Of A Wireless Vibrationally Actuated Tensegrity Robot, Zongliang Ji

Honors Theses

Soft robotics is an emerging field of research due to its potential to explore and operate in unstructured, rugged, and dynamic environments. However, the properties that make soft robots compelling also make them difficult to robustly control. Here at Union, we developed the world’s first wireless soft tensegrity robot. The goal of my thesis is to explore effective and efficient methods to explore the diverse behavior our tensegrity robot. We will achieve that by applying state-of-art machine learning technique and a novelty search algorithm.

Autonomous Collision Avoidance In Small Scale Vehicles, Justin T. Sharpe

Honors Theses

The undergraduate research performed in this study focused on autonomous collision avoidance in small scale vehicles. The goal of this study was to find equipment to build a fully autonomous small scale vehicle for use in different applications. Radio frequency communication, ultrasonic sensors, and single board computers were used to create an autonomous vehicle for multiple applications. Different communication protocols and sensors were investigated, and an explanation was specified concerning the hardware choice. The main communication protocol tested was Long Range Wide Area Network, and the main electronics tested and used were ultrasonic sensors, First Person View cameras, and the …

Development Of A Fully Instrumented, Resonant Tensegrity Strut, Kentaro Barhydt

Honors Theses

A tensegrity is a structure composed of a series of rigid members connected in static equilibrium by tensile elements. A vibrating tensegrity robot is an underactuated system in which a set of its struts are vibrated at certain frequency combinations to achieve various locomotive gaits. Evolutionary robotics research lead by Professor John Rieffel focuses on exploiting the complex dynamics of tensegrity structures to control locomotion in vibrating tensegrity robots by finding desired gaits using genetic algorithms. A current hypothesis of interest is that the optimal locomotive gaits of a vibrating tensegrity exist at its resonant frequencies.

In order to observe …

An Internet Of Things (Iot) Robotic Light Switch, Arsal Habib

Honors Theses

The purpose of this project was to design and build an Internet of Things (IoT) Robotic Light Switch. This light switch would act as a complementary device to existing toggle switches and equip them with typical smart device features such as remote control (via Wi-Fi), digital voice assistant control, motion-sensing and integration with other smart home solutions (such as IFTTT). All these features would be supported by a dedicated web page that would also be able to view the usage statistics of the device. This device is aimed at offering users an extremely easy-to-install home automation solution that needs no …

Control System For 3d Printable Robotic Hand, Htoo Wai Htet

Honors Theses

Humanoid robotics is a growing area of research due to its potential applications in orthosis and prosthesis for human beings. With the currently available technologies, the most advanced robotic hands used in prosthetics or robotics can cost from $11,000 to $90,000, making it inaccessible to the general population of amputees and robotics hobbyists. Most of the features provided by these expensive technologies are superfluous to many users, creating a great gap in cost and services between users and technology. Using the emerging 3D printing technology, my project is to construct a 3D printed robotic hand that can reproduce as many …

Follow Me Robot, Victoria M. Edwards

Honors Theses

The idea of a personal robotic assistant, while once made up the tales of science fiction, is now moving closer and closer to reality. It is easy for two humans to walk together where one person knows the directions and the other person follows along. However, it is not a trivial problem for a robot to recognize a human, and follow side by side along with them. There are three key parts to this problem: 1.) Detection of a human, 2.) Tracking the human, and 3.) the robot's Motion strategy. Using baseline assumptions, I can simplify the detection process allowing …

Evolving Soft Robots With Vibration Based Movement, Andrew Danise

Honors Theses

Creating effective designs for soft robots is extremely difficult due to the large number of different possibilities for shape, material properties, and movement mechanisms. Due to the lack of methods to design soft robots, previous research has used evolutionary algorithms to tackle this problem of overwhelming options. A popular technique is to use generative encodings to create designs using evolutionary algorithms because of their modularity and ability to induce large scale coordinated change. The main drawback of generative encodings is that it is difficult to know where along the ontogenic trajectory resides the phenotype with the highest fitness. The two …

The Design Of A Maneuverable Rolling Robot, David Carabis

Honors Theses

The purpose of this project was to design, fabricate, and test a maneuverable rolling robot. Although some other rolling robots were researched for this project, a novel approach was taken to design a unique, cheap robot that could turn and was fully enclosed by a rotating outer shell. The design and research phase of this project included the evaluation of several designs, the development of a mathematical model detailing forward motion of the robot, and the derivation of several design equations. Of the possible designs, an interior counterweight was chosen to provide a torque to the outside shell and move …

I Am The Robot. Are You?, Rachel Michelle Feldman

Honors Theses

I was originally inspired by a previous project in a 3D modeling class. For my final project, I designed two robots; male and female; who’s naked bodied reflected the presence of technological communications in society and lack of face-to-face interactions. I displayed the robots in a variety of sexual positions, which appear cute or sweet to the viewer because of the toon-shader effect. The toon-shader effect makes the 3D models look like cartoons. I loved combining the idea of technology, robotics, and sex into one. To this day, the poster of the robots in my room always gets a response. …

Union Advanced Educational Robot, Erik Skorina

Honors Theses

Our project was the design and construction of a medium sized (approximately 40 kg when fully loaded) mobile robot for educational and research use. This robot was designed to provide a strong baseline chassis with sensors, actuators and necessary software to allow for easy integration with existing robotic equipment, software and curricula. In particular, it was designed for use in intermediate and advanced undergraduate robotics courses and undergraduate research projects in robotics.

Scalable Co-Evolution Of Soft Robot Properties And Gaits, Davis K. Knox

Honors Theses

The field of soft robotics is very promising; applications in-clude urban search and rescue and covert surveillance, but these projects are not yet realized, partly because of the difficulties in soft robot shape and locomotion design. Be-cause of this, traditional design methods do not prove to be effective. This project attempts to come up with solu-tions to this soft robot design problem; utilizing a genetic algorithm, a computer simulation of Darwin’s “Survival of the Fittest,” this project attempts to make soft bodies move. This genetic algorithm evaluates each solution in simulation, and assigns each one a fitness based on distance …