2023 Robotics Engineering MQPs
|
Abstract: Occasionally causing power outages for up to 10,000 customers, ravens continue to damage equipment at Eversource Energy's Rimmon Substation in Goffstown, New Hampshire. To address this issue, a project team previously created a bird deterrent prototype that utilized an ultrasonic rangefinder to detect birds and flashing lights and sound to deter them. In an effort to improve upon the previous prototype, we have added a bird detection model using artificial intelligence (AI) to our system. Our AI detection model can recognize the presence of a raven, notify the system, and activate a set of deterring stimuli. Additionally, we have incorporated new sensors for bird detection, including a laser-based "Time of Flight" (ToF) sensor that detects the bird's proximity. We have also created a mobile application for Eversource employees to remotely monitor and control the robot while it is deployed on the wire. To ensure durability, we have re-designed the electrical housings to include weatherproofing and mechanical stability, and we have introduced a docking charging station that wirelessly charges the robot using solar energy. Our improvements are expected to enhance bird detection accuracy, system robustness, and user experience. Team Members: Brandon Chong; Jack Leserman; Jake Mercier; Jeremy Trembley; Jolie Walts Advisor: Professors Greg Lewin; William Michalson; Jing Xiao |
|
CLARA – Continuum Locomotive Alternative for Robotic Adaptive-exploration Abstract: This project follows in the footsteps of its predecessor project submitted in 2022, and aims to implement various mechanical and functional improvements in the robot. The previous year’s project created a salamander-inspired soft robot for use in the inspection and exploration of pipe networks capable of maneuvering in a variety of pipe shapes, sizes, and configurations. The robot features an origami body segment following a Yoshimura crease pattern. This pattern allows the body segment to deform and bend in a manner controllable by three cable winch mechanisms. The robot locomotion is achieved with a three-segment wheeled mechanism on a suspension-based linkage. These segments can expand and contract to provide variable force on the interior surface of the pipe, allowing for variable grip. This iteration of the project improved upon multiple systems of the robot to provide better functionality, robustness, and overall quality. New wheels were fabricated out of silicone for better grip. A new casing structure was introduced to improve rigidity and protection of vital robotic components, as well as reduce the overall robot profile. The new system uses an ESP32-Cam module to provide live footage to the operator. Relocation of actuators allowed for a maximum bending angle of approximately 140 degrees, an approximate 60% increase over previous iterations. This project expands upon the previous project’s foundations, providing further support and functionality for future research and applications in pipe network exploration and inspection, and possible payload integration. Team Members: Derik Pignone; Blaise Schroeder Advisor: Cagdas Onal |
|
Abstract: In this project, we worked interdisciplinarily with Interactive Media and Game Design and Computer Science students to create an engaging robot escape room. Participants join the game via a website, where they are informed that they are in an asteroid field on a spaceship and that they must fix the ship in order to escape the field and save the crew. The player must control the robot using a keyboard while the camera feed is shown on the website and navigate through a series of puzzles. This not only tasks the player to think critically, but also demands versatility from the robot and the room in order to make this a fun playing experience. The website links the room and robot together to create a seamless experience for the player, with the room keeping track of the player's progress in regards to the puzzles and the robot acting as the physical manipulator of the room. The project examines the combination of real-world robotics engineering and the fantastical world of game design. Team Members: Olivia Bell; Nathan Clune; Lillie DeHaemer; Liang Lu; Grace O'Reilly; Zachary Sarrett Advisor: Professors Berk Calli; Melissa Kagen; Gillian Smith |
|
Symbiotic Multi-Agent Construction 3.0 (SMAC 3.0) Abstract: Swarm robotic solutions have the potential to solve many of the dangers and issues that current construction practices have today. By having many robots working on a single project, construction can be completed in a fast and efficient manner. The overall intent for this project is to design a versatile robotic swarm system capable of autonomously constructing complex structures by means of inchworm-like linkage robots and standardized interconnected “smart” blocks. This project introduces a new backbone linkage system in the blocks and improved stigmergic control that builds upon the groundwork established by past iterations of the project such that the system is capable of constructing higher-complexity structures. This system will come with a complete algorithm capable of detecting if any structure is stable during and after construction while building the blueprints to be propagated among the blocks. Team Members: Jacob Leavitt; Kayla Lepping; Jose Rivera; Brian Shin Advisor: Professors Greg Lewin; Carlo Pinciroli |
|
Abstract: Humanoid robots can potentially replace humans in hazardous environments. If these robots can imitate humans with precision, they can also be utilized in a variety of other fields where they can collaborate with humans. The objective of this initiative was to create a novel torso mechanism that could be utilized in a humanoid robot. The newly created mechanism can move in flexion/extension, lateral flexion, and rotation, much like a human torso. It also includes a unique imitation spine that lessens the force on the torso actuators while contributing to the robot's anthropomorphism. Team Members: Jacob Bernard; Ryan Kievra Advisor: Professor Mohammad Mahdi Agheli Hajiabadi |
|
Abstract: Mobile robotics is a growing field that focuses on developing platforms that can move throughout an environment to accomplish several tasks. Traditionally, mobile robotics consists of a payload attached to a wheeled base, like what can be seen on anything from iRobot’s Roomba robotic vacuum to NASA’s Mars Rovers. However, wheeled robots are largely restricted to smooth and flat surfaces. In many cases, it is useful for robots to be able to climb obstacles like stairs or jump up to higher elevations to improve their mobility. Recently, there has been a push to recreate the adaptability of humans and animals’ legs to make a robot that can truly go anywhere we can. Robots like Boston Dynamics’ Spot and Atlas, Agility Robotics’ Digit, and the hundreds of quadruped robots across the world push the boundaries of where robots can go. The limitless applications of legged robotics make them extremely attractive for working alongside humans since, in theory, there is no obstacle that we could traverse that they cannot. However, legged robots are incredibly inefficient in comparison to wheeled robots, and in general, complex obstacles that require legs to traverse only make up a small fraction of the total terrain that a mobile robot can expect to encounter. For example, if a robot is attempting to navigate a warehouse, it will likely spend most of its time driving around a smooth floor and only occasionally need to climb a staircase or traverse rough terrain. A legged robot would accomplish this navigation, but would not be the most efficient design. A wheeled robot would be highly efficient for traditional floor navigation, but unable to traverse complicated obstacles. This project explores the effectiveness of combining wheels and legs to create a highly robust mobile platform that can accomplish many types of locomotion as efficiently as possible. Team Members: Brian Boxell Advisor: Professors Mohammad Mahdi Agheli Hajiabadi; William Michalson |
|
Robot Ecosystem for Monitoring Climate Change Abstract: The effects of climate change are far-reaching and difficult to accurately predict. The Intergovernmental Panel on Climate Change estimates 0.26 – 0.77m of average sea level rise by the year 2100. The difference in these estimates will impact the lives of more than 10 million people. We designed an underactuated underwater glider with a closed-loop fluidic controller; soft hydrostatic pressure sensor; and custom electronic sensor board that measures temperature, pressure, and orientation. Our proof-of-concept designs are incorporated into a low-cost, power efficient glider that is easy to customize and fabricate. When deployed, the glider will be able to monitor the thickness of an ice sheet while collecting data on essential ocean variables. Our approach seeks to provide physical oceanographers with inexpensive tools that allow for collecting spatially distributed data sets with the end-goal of predicting sea levels more accurately than currently possible, and ultimately suggest timelines for installing important countermeasures to combat floods. Team Members: Maya Angeles; Kalina Bonofiglio; Matthew Haahr; Joshua Palmer; Brandon Simpson; Lauryn Whiteside Advisor: Professor Markus Nemitz |
|
Abstract: Soft prosthetic hands have difficulty replicating the motion and feel of hands, due to their lack of firmness. Our soft robotic hand achieves segmented finger motion due to the presence of PLA bones and a firm, human-like grip. A hybrid of positive-pressure pneumatic actuation and cable pulley actuation gives each finger 2 degrees of freedom. Silicone is used throughout the robot for human-like feel, a compliant grip, and to facilitate the pneumatic actuation. In addition, the hand has three degrees of freedom at the wrist, provided by three linear actuators forming a tripod parallel mechanism. Overall, this project represents an approach to prosthetic hand development that prioritizes recreating key characteristics of biological human hands via a novel bone-in design. Team Members: Brian Fennell; Mason Figler; Ndenda Mutsaku Fierro; Joelynn Petrie; Ethan Turett Advisor: Professors Mohammad Mahdi Agheli Hajiabadi; Markus Nemitz |
|
Humanoid Nursing Robot Research Platform (Gopher) Abstract: This Major Qualifying Project aims to integrate a bimanual mobile manipulator robot for nursing assistance. To this end, we developed a motorized supporting structure to integrate seven-degrees-of-freedom (DOFs) manipulator robots (Kinova Gen3) with a nonholonomic mobile base (Freight research platform). The motorized supporting structure can be controlled as an additional DOF to significantly improve the robot’s reachability and manipulability to when mobile manipulation tasks in cluttered patient rooms. We further integrated virtual reality (VR) human-robot interfaces (Meta Quest, with head-mounted display and hand-held controllers) to support the intuitive robot control and visual display. The project involves the efforts for: 1) hardware design, manufacturing and assembly; 2) developing Unity- and ROS-based software architecture; 3) integration of VR human-robot interfaces, and 4) pilot user studies to test the system usability for mobile manipulation. Team Members: Maanav Iyengar; Yveder Joseph; Hoang Nguyen Advisor: Professors Jane Li; Bashima Islam |
|
Super-elastic Continuum Robot for Endoscopic Articulation and Manipulation (SCREAM 5.0) Abstract: The objective of this project is to develop a robotic system for in-office laser surgery of the vocal folds. Laryngeal tumors affect approximately 2.5% of the adult population, and office surgery is emerging as a convenient treatment option for these tumors, as opposed to traditional (and more expensive) surgical treatment in the operating room. Unfortunately, not every patient qualifies for office surgery: recent studies indicate that 1 in 3 patients have tumors in locations that physicians cannot reach with the limited instrumentation available in the clinic. To tackle these issues, we explore a new set of needle-thin, highly dexterous robotic instruments, with the overarching goal of amplifying a physician’s reach during office procedures. We articulate our proposed instruments using a concentric agonist-antagonist mechanism, which enables tight bending in a small form factor. We propose a novel intuitive control system and an experimental procedure to verify its ability to reduce the difficulty of completing the in-office procedure. Our experiments verify the mechanical capabilities of the device and demonstrate that our device increases the portion of the voice box accessible to the physician during an endoscopic procedure. Team Members: Christopher DeMaio; Julia Farnum; Ananya Gopalan; Jacquelyn Lopez Advisor: Professors Loris Fichera; Yuxiang Liu |
|
Real-Time Magnetic Sensing for Soft 3-DoF Continuum Robots Abstract: Designing accurate and real time sensing systems for continuum robots is difficult because of their high degrees of freedom. This project proposes a magnetic field sensing solution for 3-DoF pose estimation of soft continuum robots. We investigated our sensor inside an origami continuum robot, capable of 3-DoF constant curvature motion. Using motion capture, we validate our sensing solution in this system, and we measure sensing errors of less than 1.6mm. Our experiments also demonstrate the sensor’s accuracy during external object contact and unmodeled shape deformation. This application uses a custom high speed particle filter that runs in excess of 2 kHz, allowing this system to be applicable for real time feedback control of continuum robots. Team Member: Mason Mitchell Advisor: Professors Loris Fichera; Cagdas Onal |
|
Assessing Path-Planning Algorithms for Collaborative Multi-Robot Exploration with Memory Constraints Abstract: Simultaneous localization and mapping, or SLAM, is a problem in robotics in which a robot must track its own location while building a map of the environment at the same time. The use of multiple robots adds to the complexity of this problem, but has proven useful in many applications such as forest fires and disaster response. Previous research in Collaborative SLAM (C-SLAM) focused on coordinating the robot navigation and map merging for maximum efficiency in mapping an area. However, little research exists that takes into account minimalistic robots with limited onboard memory. This project analyzes multiple algorithms for their performance with regard to the C-SLAM problem with a robot memory capacity constraint, such that no one robot could hold the entire map. To achieve this, the team performed a literature review to identify algorithms from similar problems. These algorithms were then adapted to a series of sub-problems. Each sub-problem isolated a factor of the larger C-SLAM problem and allowed for a progression of complexity. In the sub-problem simulations, each algorithm’s performance was evaluated to find which one was best suited to the C-SLAM problem. The team took the best algorithm and implemented it into a multi-robot system to collaboratively map an environment. Team Members: Nicholas Grumski; Abigail Hyde; Advisor: Professor Carlo Pinciroli |
|
Creating a Multi-Robot, Multi-Human, Multi-Platform System Abstract: The study of multi-robot systems is a growing field in robotics, as it enables large-scale and parallel operations, with applications including search and rescue. Robot swarms are often imagined as fully autonomous systems, but when integrating robots into real-world applications, human supervision is still needed to diagnose issues and provide direction. However, as the number of robots in the system grows, it quickly becomes challenging for humans to control actions, understand intentions, and troubleshoot errors. With this project, we propose a novel approach in which multiple humans assume roles organized into a hierarchy. Our approach includes two roles; a proximal worker and a remote supervisor. A key insight in our work is that different roles require different interfaces. For this reason, we explore the design of an interface based on augmented reality goggles that offers the proximal worker visualization and direct control of the robots, while the remote supervisor uses a 2D interface to allocate resources and manage robot status information. We hypothesize that combining the two roles allows for improved mental workload while interfacing with multi-robot systems. Team Members: Samantha Braun; Alyssa Magaha Advisor: Professors Jane Li; Carlo Pinciroli |
|
Khepera IV Gripper Module for Multi-Robot Self-Assembly and Collective Transport Abstract: Swarm robotics takes inspiration from biological collective systems to create decentralized solutions to the coordination of large groups of robots. An important, yet understudied class of problems in swarm robotics deals with coordinating robots with physical connections to each other that can accomplish tasks that exceed the physical capabilities of any individual robot. However, hardware that allows for physical connections is not commercially available, and research is typically conducted in simulations or with custom hardware designed for specific research applications. The aim of this project is to supply a commercially available mobile robot, the Khepera IV, with a module that enables research in collective transport, self-assembly, and collaborative navigation. To this end, we identified three basic functionalities our module must offer: (i) a gripper capable of latching onto other robots or other objects; (ii) the possibility to indicate latching slots for other robots; and (iii) the ability of the gripper to rotate independently from the body of the robot, to allow for gripping while moving in arbitrary directions. To meet these requirements, we created a custom rotating turret module. We designed a ring compatible with the module’s gripper which allows the Khepera IV to attach to other robots or objects for transport. A string of LEDs adorns the ring which can be used to convey latching points. The turret body is also capable of rotating 360° continuously, independently of the chassis. Team Members: Yasmine Aoua; Chandler Garcia; Julian Poindexter; Zachary Rivernider; William Stanley Advisor: Professors Carlo Pinciroli; Gregory Lewin; Stephen Bitar |
|
Small-Size Soccer Robots Abstract: The Small-Size Soccer Robots MQP is an interdisciplinary first-year project that aims to design, fabricate, and test a multi-robot system for the international RoboCup Soccer League, targeting the Small Size League competitions. This project unites Robotics Engineering, Computer Science, Mechanical Engineering, and Electrical and Computer Engineering teams to develop a team of small autonomous robots adept at playing soccer with a golf ball. The Small Size League highlights intelligent multi-robot/agent collaboration and control within a dynamic environment, employing a hybrid centralized/distributed system. The project encompasses various tasks, such as designing, fabricating, and integrating the robot's structural and electromechanical components, including the chassis, ball control, and drive systems. The team also designs, assembles, and implements the robot's electrical circuits, featuring the processor, motor controllers, solenoids, and power distribution, while developing corresponding firmware for seamless integration. Additionally, the team crafts software to govern robot movement and execute strategic game tactics, ensuring a competitive performance in the RoboCup Small Size League. Team Members: Spencer Belleville; Conner Christensen; Ashley Espeland; Logan Rinaldi; Nathan Rogers; Benjamin Schwantes; Evan Vadeboncoeur; Yifei Zhao Advisor: Professors Siavash Farzan; Stephen Bitar; Alireza Ebadi; Joshua Cuneo |
|
Beach Swarm – Phase IV Abstract: This project aims to provide a sustainable approach to keeping beaches clean. Beaches are the habitat of a multitude of organisms that are vital to a network of ecosystems. The integrity of these ecosystems is threatened by waste and pollutants that litter streets, waterways, and wetlands across the world. Allowing pollution to disrupt the environment endangers both wildlife and humans. This project aims to help clean beaches using an autonomous swarm robotic system. The robot will be capable of collecting trash by sifting out sand. There are currently beach-cleaning robots, however, most are not autonomous. Our robot system will utilize computer vision to autonomously identify trash to clean up and organic materials to avoid. Team Members: Lucia Bernard; Sophia Cheng; Cooper Ducharme; Benjamin Watkin Advisor: Professors Nicholas Bertozzi; Haichong Zhang |
|
DigSafe Autonomous Cable Detection Abstract: Underground power cables must be marked by utility providers before any groundbreaking work can be performed on a work site. Currently, technicians mark cables by hand, which can be time-consuming, monotonous, and dangerous. The Dig Safe Major Qualifying Project aims to develop a prototype robot to perform autonomous underground electrical cable detection and marking. The robot will navigate a work site, detect buried cables, and accurately mark them in compliance with Massachusetts Law. Based upon the previous work executed by the two teams from years prior, our goals were to strengthen the subsystems - cable detecting, following, and marking. Within these subsystems, we sought to identify the areas that need improvements, determine how they can be improved, and conclude how these changes can affect the efficiency of the whole system. To this extent, several key discoveries were made such as the motion controller sending inconsistent signals, spray arm breaking due to its weight, and the camera being unable to retrieve data from the wand. Through our development stages - we mitigated the sensor interferences, refined the wand data retrieval, reconstructed the spray arm and its motion planning, and developed a new cable prediction algorithm. Additionally, we provided further developments to the navigation stack, allowing for further testing to be executed autonomously. Team Members: Victoria Heffern; Grace Holden; Zach Jester; Leo Morris Advisor: Professors Greg Lewin; Jing Xiao |
|
Electric Conversion of a Triumph Spitfire MKIV (1972) Abstract: With an objective of increasing reliability and longevity, our team converted a student-owned 1972 Triumph Spitfire (MK IV) from a gasoline internal combustion engine to a fully electric system. Our team removed the engine, transmission, and OEM supporting equipment. We then designed, manufactured, and integrated a manual-electric drivetrain, 24 kWh battery unit, embedded sensor suite, and cloud connected user interface. Our system is capable of a theoretical ~ 90hp, and 100 miles of range. Development, fabrication, and testing are still underway, including early development of level 2 autonomy using computer vision with the goals of lane-following and adaptive cruise control capabilities. Initial test drives are extremely promising. Our team hopes that by completing this conversion, we can improve the reputation of EVs and renewable energy through our implementation of cutting edge technology in a vintage car. Team Members: Shane Donahue; Patrick Flanigan; Grace Magnotta; Sean McMillan; Blaise Pingree; Wynn Roberts; Rachael Smith; Bradley Sprunger Advisor: Professors Nicholas Bertozzi; Donald Brown; Joshua Cuneo; Craig Putnam |
|
Modular Package for Autonomous Driving (mPAD) Abstract: The mPAD project aimed to develop an autonomous driving solution for scaled cars. The modular package consists of a Raspberry Pi and an Arduino Uno, with sensors such as a webcam, ultrasonic distance sensors, TFluna and an inertial measurement unit. The system was developed using the Donkey-car platform, which is a neural network-based system. Sufficient training data and various pre- and post-processing activities allowed the mPAD system to achieve a linear regression model with an cross-entropy loss of 0.05% for throttle and steering predictions. Object detection techniques were also integrated to enhance the safety and reliability of the autonomous driving system. The system was tested on ten different student-built cars that were refurbished to working condition and integrated with mounting fixtures for electronics. The mPAD system also includes a web-based dashboard to control the car as well as to present data from the various sensors on the car. Team Members: Rohan Anand; Martin Bleakley; Allison Colon-Heyliger; Ian Khung; Ishan Rathi Advisor: Professors Kaveh Pahlavan; Pradeep Radhakrishnan |
|
NASA Lunabotics Competition 2022-23 Abstract: The goal of this MQP was to fully automate the functions of the previously-built rover designed to compete in NASA’s annual Regolith Excavation Challenge. This was accomplished by incorporating encoder monitoring and hall effect sensors to measure the position of each subsystem relative to starting configuration. In addition, the task of full autonomy meant that the robot would need to perceive and interpret the competition field around it. This was done by integrating a D455i camera for visual and stereo-depth processing. Additional work was done to improve some existing mechanical systems. These included redesigning last year's team's storage and deposit subsystem to overcome the clearance issue. An additional redesign of the excavator subsystem was conducted so that it would require less time to reconfigure from the storage position and the fully extended digging configuration. Team Members: Alexander Brattstrom; Carter Bullock; Roopsa Ghosh; Patrick Hagearty; Helen Le; Joshua Moy Advisor: Professors Kenneth Stafford; Joshua Cuneo; Hektor Kashuri |
|
Abstract: The goal of this project was to further develop WPI’s current Sailbot, an autonomous robotic sailboat created to compete in the Sailbot International Robotic Sailing Regatta. In preparation for this competition, we further developed the mechanical, electrical, and software components of the robot to improve navigation, function, and design. Team Members: Samuel Alden; Douglas Moore; Cameron Pelletier; Owen Pfannestiehl; Joshua Unger Advisor: Professors William Michalson; Kenneth Stafford |
|
Bimodal Quadruped Robot (BiQu) Abstract: The focus of this project is on the development of a quadruped robot named Solo 12, which is designed to switch seamlessly between quadrupedal and bipedal modes of locomotion. This adaptability makes it ideal for navigating unstructured environments. The project delves into the technical aspects of the robot's development, including hardware architecture, electronics installation, and software stack with an emphasis on locomotion controls. Additionally, the project explores the use of computer vision technology for person tracking. Team Members: Hushmand Esmaeili; Yuen Lam Leung; Freud Oulon; Aadhya Puttur Advisor: Professors Mohammad Mahdi Agheli Hajiabadi; William Michalson; Andre Rosendo |
|
Development of 3D Printed Humanoid Robots, Koalby and Ava, as Lab Assistants Abstract: 3D-printed humanoid robots, Koalby and Ava, as versatile lab assistants with a focus on lifting objects, pushing a cart, and walking. Static and dynamic analyses were carried out to guide a series of redesigns to improve strength and integrate new components. The redesigns included a new grip mechanism to lift objects, sensors to aid the walking component, and component updates to integrate new lower cost motors. The grip was an under-actuated, 3-point finger grip with an electromagnet attached to the base of the forearm. A new spine was created, attached at the chest and pelvis to assist with stability for standing and walking. The chest and feet were redesigned to include sensors to assist in walking. To make the robots more cost effective, all of the Dynamixel motors were replaced with HerkuleX DRS motors allowing for uniformity in design and programming. Team Members: Aashish Singh Alag; Zeñia Alarcon; Emily Austin; Tessa Lytle; James Van Milligen Advisor: Professors Pradeep Radhakrishnan; Kaveh Pahlavan |
|
Getting 3D Printed Humanoid Robots, Koalby and Ava, to Stand and Walk! Abstract: Koalby and Ava are two toddler-sized 3D Printed humanoid robots intended for human interaction and lab assistance. The capabilities planned for integration during this year’s project include standing without support, walking while pushing a cart, and lifting objects. The first task involved updating the wiring of motors and electronics to improve performance and ensure consistent functionality. Kinematics and trajectory planning were implemented to replicate human movements and simplify motion control. This was followed by the integration of sensors such as TF Luna LiDAR, Husky Lens, and IMUs to provide data required for walking trajectories and feedback control. Additional batteries and power circuits were incorporated to account for new motors and sensors. In order to test all of these aspects, a simulation model was also developed in CoppeliaSim. Finally, a user-friendly interface was created to view all the sensor data and control the humanoid robots. Team Members: Joshua Fernandez; Erin Lee; Tessa Lytle; Finbar O'Sullivan; Casey Snow Advisor: Professors Pradeep Radhakrishnan; Kaveh Pahlavan |
|
HURON: Full-Sized Humanoid Robot (Lower Body) Abstract: This project aims to develop a self-balancing bipedal robot, HURON, to replace human rescuers in hazardous natural disasters. This project's goal is to design, manufacture, and control HURON, which can react to forces on any part of its body and exhibit a human-like walking gait. The HURON Major Qualifying Project's first year focused on developing a disaster relief humanoid robot with three central systems: designing, sensing, and controlling. The team created a CAD model and machined the robot's structure from aluminum and steel components. The sensing team used a circuit design and foot force stability margin theory to measure the robot's balance and send a response to the motors to react to the imbalance. The control team used inverse kinematics to move the robot and simulate a walking gait based on the positions and angles of each joint. The team successfully manufactured the robot and connected it to the drive system to perform a walking gait, replicating a human's response to a push from behind and standard walking patterns. Team Members: Rachel Dancy; Jonathan Gong; Aislin Hanscom; Cameron Huneke; Peter Lam; Curtis Lee; John Marcotte; Rahil Parihkh; Angelo Ruggeri; Brendyn Sang; Kyle Staubi Advisor: Mohammad Mahdi Agheli Hajiabadi; William Michalson; Markus Nemitz; Cagdas Onal; Karen Troy |
|
PRIMO (mobile printer): 6-axis 3D Printing Construction Robot Abstract: Construction sites are typically associated with the need for caution, adherence to protocols, and manual labor. This project explores the development of the PRIMO (mobile printer) - a 6-axis 3D printing construction robot - with the potential to revolutionize the construction industry by enabling printing in remote locations and improving these aspects of construction. This six-legged (hexapod) robot can walk, and its main body has six degrees of freedom, facilitating nonplanar printing on an almost limitless print bed size. In addition, the robot is equipped with a custom concrete extruder and brick placement mechanism, enabling the printing of entire structures with minimal human intervention, one after another. Team Members: Aaron Longo; Christian Stilwagen Advisor: Professors Mohammad Mahdi Agheli Hajibadi; Markus Nemitz |
|
Automated Quality Analysis of Disc Golf Discs Abstract: Customers of disc golf discs desire individual data about the disc they are purchasing. A successful online individual disc retailer needs an efficient system to take this data for every disc in their inventory, but there are no machines that currently exist to solve this problem. The team is designing and constructing such a machine to allow Maple Hill Disc Golf to begin an online store selling individual discs. Team Members: Vanessa Cardaropoli; Ayden Duncan; Greg Marshall; John Robinson Advisor: Professors Greg Lewin; Jing Xiao |
|
CNC Alarm Resolution and Work Cell Monitoring with a Robotic Arm Abstract: Computer Numeric Control (CNC) revolutionized the manufacturing industry by enabling the creation of highly accurate and complex parts at high speeds. However, CNC systems have been heavily reliant on human operators to oversee and control the manufacturing process since their creation. Despite recent advancements in automating CNC milling, existing approaches have overlooked the need for a methodology to handle errors and failures in the system. To address this gap, we developed a set of Robot Operating System (ROS) nodes that monitor and autonomously manage a CNC while reacting to alarms in real-time with the hope of reducing downtime during lights our manufacturing. At any point during normal operation a CNC could be stopped by an alarm halting production. Our system is designed to reduce this downtime by resolving the present alarms. Our approach additionally tracks a variety of performance metrics, allowing for later evaluation and continuous improvement of the system via the Flexxbotics environment. The methodology we created is scalable and adaptable to a wide range of CNC systems and can be customized and configured to fit the specific needs of different manufacturers. Because the framework is adaptable to other CNC systems, our methodology is a valuable tool for manufacturers looking to improve their efficiency and productivity. Team Members: Brian Francis; Mayank Govilla; Niko Neather Advisor: Professors Siavash Farzan; Greg Lewin; Jing Xiao |
|
Machines Building Machines Abstract: Advancement in the field of soft robotics depends on the ability to consistently manufacture soft components. This process is difficult since Fused Deposition Modeling (FDM) printers can print materials with a minimum shore hardness of 50A, silicone injection models do not allow for rapid prototyping, and hand-poured silicone molds are subject to human error. Silicone elastomers come in a wide variety of hardnesses, however, printing systems cost thousands of dollars, making them inaccessible for most labs. In contrast, our newly developed printhead is capable of FDM printing silicone parts by integrating with the existing E3D ToolChanger system for under $250. This print system was purpose-built in software, design, and wiring to only require an hour or two of setup and basic 3D printer knowledge. This is possible because the pump design of the printhead has one stepper motor which can be commanded in the same manner as the extrusion stepper on a traditional FDM printer. Despite a 10% flow rate inconsistency in our pump system, the printhead produced successful prints with similar properties to traditionally molded silicone, with a shore hardness of 00 35. Team Members: Elena Bachman; Jared Minnich; Ronald Pfisterer Advisor: Markus Nemitz |