Projects
|
Autonomous single-axle self-balancing robotOverview: As part of a graduate level Robotics lab, a classmate and I built a single-axis self-balancing robot that we could drive manually with a controller and program to autonomously perform several tasks. The tasks included driving in a 1m by 1m square without driving past the boundary, race down a 10m long track as quickly as possible and stop without falling, and navigate an obstacle course passing through 4 directional gates in order without bumping anything. My teammate and I competed for the best performance among our peers and came out the winners.
Role: I designed a chassis for the robot to mount all the components using SolidWorks and used a laser cutter to cut the design from an acrylic sheet. I then mounted all components to the chassis including servomotors, beams, a battery, an IMU, and a Raspberry Pi board. I then connected all the wiring to the board, calibrated the IMU, and programmed the servomotors to respond to changes in the IMU data. I created a PID controller to keep the robot upright and tested it extensively to ensure the robot could move at high speed while remaining upright. For the final task, I parsed OptiTrack camera data to determine the robot and gates' positions and create a path that passes through each of them in order and in the correct direction without bumping into the gates. |
Autonomous 6 DOF Robotic ArmOverview: As part of a graduate level Robotics lab, two classmates and I built a 6 DOF robotic arm with a gripper to manipulate colored blocks into various patterns autonomously using a Kinect camera to determine the blocks' location relative to the robotic arm. The tasks included mirroring the locations of three blocks along the y-axis, stacking three blocks in color order, destacking seven blocks then lining them up in color order, destacking seven blocks then stacking them into one pile in color order, and creating a pyramid as tall as possible. My team competed for the best performance among our peers and placed in the top 3 in our class.
Role: I calculated the forward and inverse kinematics of the robotic arm and developed a state machine for each task using Python. One of my teammates designed the gripper for the robotic arm and the other teammate calibrated the Kinect camera and isolated the color and position of each block relative to the base of the robotic arm. |
|
|
Autonomous Holonomic Maze-Running RobotOverview: As part of a Mechatronics class, three classmates and I formed a team and were tasked with creating an autonomous robotic vehicle that can independently navigate a maze, identify an object, and move it to a drop off location in the maze. To navigate the maze, we send ultrasound readings from PING sensors on the front and sides of the robot to a Raspberry Pi board that in turn made decisions about the path to take and translated those decisions to the three wheels. To make manipulation of the robot within the maze easier, we decided to make the system holonomic. Once the robot was within line of sight of the object, a Raspberry camera mounted in the front was used to evaluate the distance from the robot to the object. Using these evaluations, the Raspberry Pi made adjustments in order to position the object within range of the robotic arm. Servo motors were used to manipulate the arm to pick up the object before the robot navigated the maze again to find the drop off point. Once the point was in line of sight, a similar process occurred using data from the camera to position the object within the drop off area and release the object. The video to the left shows a preliminary test of the robot's maze navigation.
Role: I took on the role of CAD designer and assembler. Most of the acrylic chassis was made on SolidWorks and cut from a laser cutter by myself. Other team members were in charge of the software and electrical components, while I handled the more physical components such as mounting the motorized wheels and sensors to the chassis, ensuring the arm operated correctly, creating three levels for the chassis to better organize components, and testing navigation within the maze. |
Prosthetic Hand Assessment Method (PHAM)Overview: PHAM is used by occupational therapists to evaluate the wrist motion and handgrips of an amputee's prosthetic arm when moving a pin. When the button is pressed, LEDs light up on the stand to indicate the starting location and the target location of the movement. When the button is pressed again (signaling a completion of the task), the time it took the amputee to complete the task is recorded. Using this data as well as readings from IMU sensors placed along the amputee's prosthetic arm, the quality of movement is determined, allowing occupational therapists to quantify improvements in their patient's prosthetic control. The video to the right shows a volunteer manipulating a prosthetic hand to perform the clip movements on PHAM.
Role: I served as the lead design engineer in this project. I worked with another BME undergraduate to build, wire, and program this module based on feedback from therapists and amputees that tested this system. Together, we designed code in Arduino to signal various hand motions via LEDs, developed a GUI in Python to interface with Arduino code, customized clip handles to deform to four different hand grips, constructed a prototype of PHAM, and tested PHAM with two amputees using Be-Bionic prosthetic hands. |
|
|
Robotic ENT Microsurgery System (REMS)Overview: Microvascular surgeries require a great degree of precision, accuracy, and stability for successful operations. In many cases, the microvascular anastomosis of vessels during free tissue transfer are the most technically challenging and critical portion of these already long procedures. Controlling hand tremor during vein suturing is invaluable, and even skilled surgeons who have spent years perfecting a steady hand still exhibit a slight tremor. The negative effects of hand tremors are magnified at the microsurgical scale. The Robotic Ear Nose and Throat Microsurgery System (REMS) was developed by Dr. Taylor, Kevin Olds, and Marcin Balicki to address this pressing issue by reducing hand tremors. In this project, I redesigned the tool set and attachment mechanism of the REMS for seamless integration of suture needle holders, improving the ease, efficiency, and accuracy of microsurgical procedures. The video to the left highlights REM's tremor reduction.
Role: I served as lead designer of a novel tool set compatible with a steady hand robot to reduce surgeon hand tremors during microsurgical procedures such as vein anastomoses. Some tasks for which I was responsible include: developing CAD files of prototypes using SolidWorks, rapid prototyping and constructing designs in machine shop, and conducting pilot studies on the REMS with 10 participants. More information about the project can be found here. |
Stech CuffOverview: I was part of a team of eight undergraduates working with experts to implement novel cuff designs for endotracheal tubes (ETTs) to decrease tube dislodgment during surgery. Endotracheal tubes (ETTs) are a common conduit for delivering positive pressure ventilation to patients undergoing anesthesia for surgical procedures. These tubes have an inflatable cuff near the distal end to secure the tube in position in the airway tract and prevent leakage of ventilator gases that are flowing in and out of the lung via the endotracheal tube. Single Lung Ventilation (SLV) is a specialized anesthetic technique that allows separate and isolated ventilation of each lung. One problem that ETTs commonly experience especially during SLV is dislodgement; 8 out of 100 pediatric SLV cases experience some form of misplacement or dislodgement of the endotracheal tube. My team designed, prototyped, and tested a novel mechanical, textured endotracheal tube cuff model combined with a biomaterial component, which is designed to attach to an existing ETT. Testing was done on 3 adhesives: polyvinylpyrrolidone (PVP), methyl cellulose, and hydroxyethyl cellulose (HEC), at 3 different concentrations. The biomaterial chosen for the design was hydroxyethyl-cellulose (HEC), a colorless mucosal adhesive gel commonly used in dental glue. Peel and shear testing was performed using a printed, anatomically accurate tracheal model, as well as ex-vivo and in-vivo piglet tracheas. Of the two videos on the right, the top one is an animation of dislodgement using conventional ETTs, while the bottom is an animation of firm placement using the StechCuff.
Role: I served as the lead CAD designer in my team, developing CAD files of ETT cuffs using SolidWorks and AutoCAD and rapid prototyping designs. I sent schematics of our prototype to the patent office and received a provisional patent for my team. I performed additional tasks for the team as well, such as testing and evaluating designs in vitro and in vivo and presenting results at business plan competitions. |
|
|
MRI-Compatible Robotic Cardiac Catheter SystemOverview: Radiofrequency ablation is the gold standard for treating arrhythmia and tachycardia. In the procedure, a surgeon inserts a catheter into a patient's femoral vein and maneuvers it into the heart. Once in the heart, the surgeon locates the area causing the abnormal heart beat and ablates it. X-ray fluoroscopy is used to guide the surgeon while he manipulates the catheter, however this imaging technique is not ideal because an x-ray cannot distiguish soft tissue such as the heart from the rest of the body. As such, surgeons have to rely on their knowledge of that they should be seeing to fill in blanks in the image. In contrast, MR-imaging provides much better images of the heart and could be used to more effectively perform radiofrequency ablations. This project attempts to create a robotic catheter system to allow surgeons to maneuver catheters within the tight confines of an MRI machine, while maintaining a safe distance away from any radiation. To do this, I modified the Amigo catheter system to make it MRI compatible. The video to the left shows the ablation procedure.
Role: I was in charge of modeling the system in CAD and creating structural changes that would facilitate MRI-compatibility. For instance, I replaced steel and aluminum parts with stainless steel ones and replaced electric motors with hydraulic and pneumatic ones. Currently, no testing has been done with the new system. |