Embedded Files
Laboratory for Intelligent Decision and Autonomous Robots (LIDAR)
Agile Locomotion and Manipulation VIP
Project Objective
The objective of the research was to improve the robotic foot’s traversal over various terrains using impedance control techniques. This involved refining the mechanical and control systems to enhance the accuracy of terrain interaction and foot attachment to a dynamic, bipedal robot named Cassie.
This research was part of an interdisciplinary 1-credit VIP (Vertically Integrated Project) in Fall 2023 that brought together faculty, graduate students, and undergraduates to collaborate on advancing robotic systems for dynamic, real-world applications.
Impact of Work
This research is critical because it focuses on improving robotic mobility over complex terrain, which is essential for real-world applications such as search-and-rescue missions, exploration, and prosthetics. Enhancing the accuracy of robotic foot traversal could contribute to the development of more adaptive and resilient robots capable of navigating uneven and unpredictable surfaces, ultimately pushing the boundaries of autonomous robotics and improving human-robot interaction in challenging environments.
Amazon Digit Robot
My Role
My specific role in the project was to assist in developing the mechanical improvements for the robotic foot’s ability to traverse over terrain. This included working on the impedance control aspects, which involved modeling the foot’s attachment mechanisms, improving terrain contact accuracy, and enhancing overall system performance to better navigate complex environments.
Base Foot CAD
My Contribution
Cassie Bipedal Robot
Soft Foot Sensor Design and Terrain Classification for Dynamic Legged Locomotion
Literature Reviews and Research
I conducted thorough literature reviews on impedance control techniques, the integration of inertial measurement units (IMUs), and the design considerations for dynamic bipedal robots. I explored how IMUs could be utilized to improve terrain feedback, enhancing the robot’s ability to adjust to changes in terrain in real time. Additionally, I researched the latest methods in sensor fusion to improve the reliability and precision of the system's feedback loops.
Improved Terrain Traversal Accuracy
I worked on adjusting the impedance control parameters for the robotic foot, focusing on optimizing the interaction between the foot and varying terrain surfaces. This involved researching and reviewing literature on impedance control techniques applied to bipedal robots and terrain interaction. I also performed calculations to fine-tune the impedance control algorithms, ensuring that the robotic foot could adapt its response to dynamic terrain conditions. My work included applying these algorithms to simulate the interaction forces between the foot and the ground, with particular attention to maintaining stability and efficiency during movement.
Experimental Test Bed
Instron-mounted foot impacted cyclically onto various terrain, simulating stepping process of Cassie robot
Impedance Control Calculations
Foot Attachment Mechanism
I helped plan and execute the attachment mechanism for the robotic foot onto a dynamic bipedal robot. The goal was to create a mechanism that would allow the foot to accurately adjust to different terrain conditions while minimizing mechanical stress. I reviewed various attachment designs and materials through extensive literature research to ensure robustness and precision. The final design required testing through iterative prototypes, with a particular focus on ease of assembly and weight distribution to avoid hindering the robot's movement.
Original Tarsal Prototype
Updated Tarsal Prototypes
Simulation and Testing
I conducted simulations using MATLAB and Simulink to model the impedance control system and validate the mechanical design's effectiveness. The simulations involved adjusting the parameters of the impedance control algorithms to ensure they were properly tuned for different terrain types, from flat ground to uneven surfaces. I spent significant time refining the Simulink model, using real-world data to tune the system for optimal response time and stability. The process included running multiple simulations with different terrain inputs, analyzing the results, and adjusting the system's parameters to improve foot placement accuracy and stability during the robot's traversal.
Original Impedance Control Simulink
My Updated Impedance Control Simulink
Prototyping and Testing
My role also involved extensive prototyping using the Invention Studio’s 3D printing facilities. I 3D printed various prototypes for the foot attachment and sensor housings. After multiple iterations, I fabricated a silicon cast to encase the sensors and provide better protection from environmental factors like moisture and dust. This silicone mold was custom-designed to fit the specific sensor dimensions, ensuring durability without compromising the accuracy of the sensor readings.
Tarsal Segments Front View
Tarsal Segments Top View
Sensors before Silicon
Pouring Silicon into Molds
Algorithm Tuning and Optimization
A significant part of my contributions involved the tuning and optimization of the impedance control algorithms. After simulating the robotic foot's interaction with the terrain, I adjusted control parameters such as stiffness and damping to match the expected dynamic responses. The process involved balancing the foot's ability to maintain contact with the terrain while avoiding excessive forces that could destabilize the robot. Through numerous iterations and testing, I was able to fine-tune the system to achieve a responsive and stable movement across varying terrains.
Large Time to Balance (Stabilize)
Significant Reduction in Balancing Time
Challenges & Solutions
Ensuring Robotic Foot Stability Across Varying Terrain
Challenge: The primary challenge was ensuring that the robotic foot could maintain stability and effectively interact with diverse and dynamic terrain surfaces. This required careful tuning of the control parameters to allow the foot to adapt to sudden changes in terrain.
Solution: I refined the impedance control model by adjusting parameters such as damping and stiffness to improve the foot's adaptability. I also collaborated with the team to conduct multiple test trials and fine-tune the control system's response, ensuring it provided a stable and responsive interaction with various terrains.
Sampling Time and Convergence Function
Simulink Model I Created for Adaptability
Implementing a Precise Foot Attachment Mechanism for Load Conditions
Challenge: Designing an attachment mechanism that could withstand varying loads while maintaining accurate foot placement posed significant engineering challenges. The mechanism needed to be both robust and flexible to adjust to the dynamic movements of the robot.
Solution: We utilized SolidWorks to create detailed CAD models and performed simulations to predict the attachment's behavior under load. By testing several prototypes and refining the design through iterative feedback, we developed a reliable attachment mechanism that could endure varying stresses during operation.
Prototype of Foot with Raspberry Pi
Exploded View of Foot
Results
CAD Models: I helped refine CAD models for the robotic foot and attachment mechanism, ensuring compatibility with the existing robot framework.
Software: I contributed to the development of control algorithms using Simulink to simulate impedance control on the robot’s foot. This provided valuable insights into the mechanical performance and the robot’s ability to adapt to various terrain.
Notebook: I documented my research and findings in a detailed notebook, including technical descriptions of the impedance control modifications and test results.
Extension of Learning: This project deepened my understanding of mechanical design, control systems, and simulations, directly enhancing my knowledge of robotics and dynamic systems. It also gave me hands-on experience with impedance control, which is critical in robotics and biomechanical applications.
Block Diagram of Servo Motor and Tarsal Torque Control
Shaping Future Experience
This experience solidified my interest in pursuing a career in robotics, particularly focusing on improving autonomous systems and control methods. It sparked my desire for more individual research, which I then pursued through further involvement in hands-on projects and collaborations with faculty. The close relationships I formed with faculty during this project played a key role in shaping my decision to continue my education in a research-focused graduate program, where I could further explore innovations in space structures and robotics.
Full VIP Notebook
Dylan Winer_VIP_Notebook.pdfGoogle Sites
Report abuse