2021 Robotics Engineering Virtual MQPs

MQP 5 - Robotic Waste Sorting V2

The U.S. recycling industry depends on material recovery facilities to process materials collected through single stream recycling. A growing global population and recently imposed higher material purity standards require more efficient waste-sorting methods. This project contributes to research on automated solutions to this problem. A modular cartesian robot with a closed-loop linkage arm and 3-pronged gripper was designed and built to assist in sorting cardboard. Additionally, a conveyor belt for testing was designed and built, and a preliminary machine vision program to detect cardboard was developed. This work provides a basis for further research and development of waste sorting technologies.

MQP 6 - Vineyard Robot

The purpose of this project was to create a prototype of an automated pest deterrence system for Wachusett Vineyards. The following report details how birds can greatly affect a vineyard’s crop yields, how robots are used to solve agricultural problems, and how we planned and designed a cyber-physical system to deter birds from the grape crops. We used a Raspberry Pi and camera to run motion detection software and automatically scare away these birds via predatory audio. We also created a desktop application to monitor the live stream footage, manually activate the sound deterrence, and view information about previously captured motion detections.

MQP 7 - swARm

Methods of robotic swarm control are becoming increasingly prevalent in academic literature. Recent approaches have studied how augmented reality applications can facilitate control of robotic swarms. However, studies suggest that robotic swarm systems can overload the operator due to the large number of moving parts. Therefore, we developed a novel multi-user augmented reality application, deployed on the Magic Leap One, allowing operators to distribute responsibilities among each other when controlling robotic swarms in an immersive and intuitive way. Furthermore, by creating a modular architecture, new methods of multi-user swarm control can be introduced and evaluated quickly.

MQP 8 - Flexible Robotic Origami Gripper (FROG)

Origami robots have been introduced as a new soft robotic technology that can be easily constructed from planar sheets of material to create rigid or deformable linkages. With this Major Qualifying Project, the team investigated the capabilities of a modular soft robotic gripper with triangular beam and Yoshimura origami finger designs folded from PET plastic. Due to its origami design, the gripper is lightweight, flexible, and durable. A vision system working in conjunction with an impedance controller aided the gripper to determine and provide three stable grasp patterns to pick up an assortment of objects. The implementation of a switching control system has demonstrated the need for more adaptive control for this compliant gripper. This project provides a foundation for future research into origami grippers.

MQP 9 - Interactive Bio-Inspired Dino

The goal of this MQP was to design, fabricate, and implement an interactive robot model of the head and neck of a Deinonychus that creates the impression of a living dinosaur. The neck contains 6 actuated degrees of freedom and is composed of custom-made vertebrae. The vertebral column is modeled as a continuum arm and controlled by a network of Arduinos. The project makes use of Popovic Lab’s Hydro Muscles, which are elastic linear actuators that replicate muscular motion. These Hydro Muscles are powered by a custom closed-loop hydraulic system and control the dinosaur neck by pulling on wire ‘tendons’ that are attached to key points on the neck. The team focused on elements that mimicked anatomical structures in order to develop a system to mimic lifelike behavior. The dinosaur is capable of neck and jaw movements, and continuous random eye motion. Future work will focus on developing sensor-based intractability.

MQP 10 - Variable Impedance Actuators: Robotic Drumming Applications

The goal of this project was to create a robotic drumming system which could play a variety of percussive instruments, produce basic and complex strikes, and demonstrate aural and visual expressivity. To accomplish this a robotic arm was designed to approximate a human arm, with a variable impedance actuator (VIA) utilized at the wrist. The VIA can be tuned to simulate the dynamics of a drum roll. Due to circumstances resulting from the COVID-19 pandemic, thee team produced simulations to verify the design of the robot. Despite the arm being unable to fully reproduce the range of human playing ability, this simulation demonstrated the capability of the robotic arm to replicate, with moderate fidelity, the aural and visual characteristics of human players.

MQP 11 - AFIA – Aid for Investigating Accidents

In Ghana and other sub-Saharan countries access to robotics is extremely limited by the lack of materials and the cost of imports. The AFIA team developed a robot made mostly out of materials readily available in Ghana in the hopes that future designs like this can help make robotics more accessible. The robot is a multipurpose robotic platform designed to travel on poorly maintained dirt roads that are common in remote areas around the world and allows for the addition of sensors to fulfill the needs of the user. The goal of this project was to create an affordable, sustainable, multipurpose robotic mobile base to expand the availability of robotic technology.

MQP 12 - Car-Snow Clearing Drone

Manually removing snow from car windshields exposes people to frostbite and can be especially challenging for those with physical disabilities.This Major Qualifying Project (MQP)aims to make improvements to a drone designed one year prior to solve such problems.Unfortunately, COVID regulations restricted our access to campus resources and team collaboration. Despite these hurdles, this iteration generated a MATLAB simulation of the drone,improved the reliability of the spray bar, and began implementation of a Real-Time Kinematics(RTK) system for precise navigation.

MQP 13 - Demining Phase V

Every year at least 7,000 people around the world die from abandoned, unexploded landmines leftover from times of conflict. Current demining methods are expensive, dangerous, and slow. To combat this issue, our team is continuing the project to create an operational autonomous demining system. The system will consist of three parts: a rover, a drone, and a base station. The rover will be able to search a user defined area for unexploded landmines. When the rover finds a landmine, it will record the location. After the search is complete, the drone will then fly to the locations of the landmines and will drop a small payload onto the mines to detonate them. The base station acts as the communication link between the rover and the drone and provides a user interface for the operator to control the system. The goal of our project is to have a functioning, relatively inexpensive system that can increase the safety and efficiency of global humanitarian demining efforts.

MQP 14 - RoboPuppet for Teleoperation

RoboPuppet is a scale-model version of Kinova Robotics’ Gen3 Arm used for intuitive remote control. This platform is ideal for enabling nurses with limited robotics expertise to work remotely with patients in high-risk and contaminated environments. The RoboPuppet arm contains joint angle sensors and motors which allow for gravity compensation and vibration haptic feedback. The project includes a ROS package for controlling the Kinova Arm both in real life and in the Trina2 simulated hospital environment, with a basic real-time GUI for calibration and debugging.

MQP 15 - Assistive Walking Exosuit

This project is focused on building on the work of the Humanoid Walking Robot and developing an application of the CRFC Valve to an assistive walking exosuit that lowers the metabolic cost of walking. Hydro Muscles are placed in the location of the gluteus maximus, gastrocnemius, and quadriceps femoris muscle groups and actuated to enhance the natural muscular movement.

MQP 16 - Autonomous Mural Painting Robot

On average, it takes a person about 40 hours to paint a full color, 10’x10’ mural. The process is time consuming and expensive, as you have to pay the artist for their time. High costs associated with murals make them unrealistic for most places. By automating the process of painting, we can drastically cut the cost of large scale, full color murals, making them more attainable for the average building. The goal for this project was to make a robot capable of painting an 8 color mural on a vertical wall, without human intervention. Our end product is able to decode any image and paint it in a given size.

MQP 17 - SLAM on the Atlas Robot

Simultaneous Localization and Mapping (SLAM) is a powerful method for autonomously exploring new environments, which has never been applied to humanoid robots. This project implemented SLAM on the Atlas humanoid robot, allowing it to interact with and explore novel environments. The Atlas robot was programmed to create 3 dimensional maps of environments and localize itself within them. Due to constraints in time and safety the project was conducted mostly in simulation. Both stereo cameras and LiDAR were used in experiments to identify the robot’s surroundings and avoid obstacles. These sensors paired with the robot’s locomotion allowed the Atlas robot to search and navigate unknown areas.

MQP 18 - Metamorphic Manufacturing

Metamorphic manufacturing is a type of digital fabrication wherein robotic systems are employed to incrementally deform material into a desired shape. This approach has the potential to greatly reduce material waste compared to subtractive methods such as CNC milling, and can achieve material properties superior to what is possible with additive methods like 3D printing. The aim of this project is to develop a prototype metamorphic manufacturing system capable of shaping plasticine clay. Utilizing the 6-axis ABB IRB 1600 robotic arm equipped with a custom end-effector and interchangeable tools, a system was constructed to achieve this objective. A LiDAR camera was employed to capture accurate 3D models of the plasticine workpiece as it is being shaped. Automated tool-changing was also implemented to allow for a fully-automatic workflow.

MQP 19 - Autonomous Drone Pollination

The declining bee population poses a threat to many aspects of life on earth as they are essential pollinators responsible for pollinating 80% of flowering plants. Our team aims to supplement the declining bee population with drones capable of pollinating flowers autonomously. The goal of this project is to create a system capable of locating sunflowers within a region and transferring pollen between them. The proposed solution consists of an off the shelf drone, a flight controller, an on-board processor, an actuated end-effector, and a camera. The team intends to incorporate all of these features together into one system capable of navigating a bounded region, detecting, and approaching sunflowers, and transferring pollen between them using the end effector.

MQP 20 - Multi-Modal Locomotion Robot

A variety of animals such as primates, dogs, and bears switch modes of locomotion between quadrupedalism and bipedalism to better complete certain tasks. However, very few robotic platforms can effectively combine the two forms of locomotion. A multi-modal robotic platform with such capabilities would provide additional adaptability in unstructured environments, broadening its potential applications. Therefore, we extended an existing quadrupedal platform with the capability to transition into a bipedal stance. In this project, we built a physical robot, developed an accompanying software stack with a reinforcement learning pipeline, implemented quadrupedal locomotion, and achieved bipedal standing in simulation. Our integrated hardware and software platform affords future roboticists the opportunity to test and develop more adaptable locomotion strategies and increase the functionality of robots more broadly.

MQP 21 - Lionfish Phase IV

Indigenous to the western Pacific, lionfish are an invasive species which have been wreaking havoc along the coasts of the western Atlantic. The rapid reproductive cycle of these fish combined with the fact that they have no natural predators in the Atlantic Ocean has made them a considerable threat to the local ecology. Current methods of combating the invasive lion fish have seen little to no results. This is the fourth year that students from WPI have collaborated to develop a robotic solution to help curb the exploding lionfish population. This year’s focus was on developing a stereo vision system capable of detecting an object and determining its distance, an improved navigation system which incorporates PID control, as well as a revised design for the harvesters' container.

MQP 22 - Needle Intervention System with Mirror Ultrasound Imaging

Procedures such as Percutaneous Nephrolithotomy (PCNL) often face problems with ineffective needle insertion due to the unknown geometric relationship with the needle, fascia, and ultrasound probe. The goal of the project is to develop a needle insertion device with mirror ultrasound imaging, which provides an intuitive and simple solution to guide the needle insertion path. This is achieved by visualizing the forward-view of the needle insertion by changing the relative angle between the mirror and US probe.

MQP 23 - NASA Lunabotics Competition

The NASA Robotic Mining Challenge is a yearly competition in which university teams build a rover to navigate rough terrain and mine and deliver simulated icy regolith. With strict size constraints, additional challenges are to minimize the rover mass, energy used, and communication bandwidth, as well as achieve full autonomy. Our MQP developed an original design for an all-new rover with a rocker-bogie chassis and a novel, two degree of freedom, backhoe-style excavator. We were only able to build and test the excavator portion of this robot due to COVID and time constraints. We also programmed and tested the state machine and navigation system, creating a code base that can be used by future teams.

MQP 24 - DigSafe Buried Cable Detection

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, and with seasonal fluctuations in the demand for the service technicians cannot always do so efficiently. This project aims to create an autonomous robot and supporting documentation to increase the efficiency of the cable marking process. The robot is intended to navigate a work site, detect buried cables, and accurately mark them in compliance with Massachusetts law.

MQP 25 - Novel Tethered Submersible for Marine Survey Applications

Current underwater inspection vehicles are either extremely slow or high priced despite their plethora of use cases. The Tethered Submersible Surveyor is an underwater, semi-autonomous tethered vehicle designed to glide across the ocean floor for fast close inspection work. The tether provides the craft's primary propulsion as well as communications and power supplied from a boat on the surface. The TSS aims to enable applications such as topology scanning, utility mapping, and other surveying applications.

MQP 26 - Sailbot

The objective of this project is to improve upon prior implementations of the robotic sailboat known as “The Great Awake”. By further developing the mechanical, electrical, and software systems put in place by prior teams, this team plans to increase autonomy and reliability throughout the entirety of the Sailbot. Our main focus for this iteration of the project is to implement a robust, fully autonomous sailing apparatus, and enable future teams to more easily continue improving the robot.

MQP 27 - Swol Kat

As the focus in robotics is shifting from specialized robots in controlled environments to generalized robots in unstructured environments, there is greater demand for legged locomotion research platforms. Four legged robots achieve a balance between speed and stability. In this project the team designed and fabricated a highly maneuverable robot using 3 DOF legs powered by BLDC motors. This quadruped is  an extensible platform for future work in gait development, computer vision, and more.

MQP 28 - Autonomous Combat Robot

In a combat robotics match, two robots armed with high kinetic energy weapons attempt to damage, control, or otherwise disable their opponent. The majority of combat robots are glorified remote controlled cars, implementing little to no autonomous features. One of the most popular weight classes for combat robotics is the 3lb or beetle class. Due to the extreme environment and fast motion of combat competition, automated robots have the potential to out-perform human drivers. This MQP’s autonomous combat robot remains aggressive and deals damage to the opposing robot, which are important elements for success. It is sufficiently robust to survive repeated impacts and accelerations, while challenging human drivers.

MQP 29 - LEGO® Sorting Robot for the Classroom

The objective of this project is to design an automated system which sorts a classroom set of LEGO® Mindstorms robotics kits, a popular K-12 education tool used in classes, summer camps, and after school programs around the world. By automating a repetitive and tedious task, this robot aims to reduce labor costs and increase time available for STEM instruction.

Consumers were interviewed to develop product specifications, including functionality, size, and budget constraints. Then, a three sub-system robot was designed including a serializer, which separates parts to prepare them for identification; an identifier, which uses a camera and convolutional neural network to identify part number and color, and finally; a distributor, which uses a Cartesian robot to place identified parts into sorting trays. Prototypes of these subsystems were then developed and tested.

MQP 30 - BeachBots 2020

Beaches are becoming more littered by trash that is discarded by humans worldwide. Currently, there are methods for cleaning trash from beaches that usually entail large, expensive machines that may not be accessible for all types of beaches. This project is focused on developing a multi-robot solution for cleaning beaches in an effective manner. Using the engineering design process, our team gathered information about beaches to design prototypes, tested designs, and build a functioning proof of concept multi-robot solution.

MQP 31 - Partial Hand Prosthetic

The purpose of this MQP was to develop a partial hand prosthetic for a patient with an amputation at the metacarpophalangeal joint of the index finger and thumb. There are no commercially available prosthetics for this specific injury and so this MQP tackles this gap in prosthetic technology. The resulting solution consists of a two degree of freedom passively actuated index finger, one degree of freedom sensor driven thumb, solenoid locking system, and passive variable compression bladder. Both qualitative and quantitative testing has demonstrated that this device restores the pinching capabilities of the patient.