Fall 2022 MAE 245 Halloween Robots

Fall 2022 MAE 245 Halloween Robots

These Halloween robots were delivering candy at UCI on October 31, 2022. For Fall 2022, we were back to in person project-based learning and the students responded enthusiastically. Over the summer we improved the joint design as well as the drive train, and we focussed on two legs to simplify the construction. The result was an improvement in overall performance that lays the groundwork for the future. Please take a look.

Kinematics Lecture:  Curvature Theory and Walking Robots

Kinematics Lecture: Curvature Theory and Walking Robots

This lecture describes my approach to project-based learning for Kinematic synthesis, which I developed while on sabbatical at Stanford University. I presented this lecture at the 2019 Kinematics Summer School organized by Anurag Purwar. Since that time, we have gotten better at making these walking robots easier to build and more reliable.

Kinematics Lecture:  Design of a Linkage System to a Draw Curve

Kinematics Lecture: Design of a Linkage System to Draw a Curve

Kempe’s Universality Theorem states that every algebraic curve has an associated linkage that draws the curve. A popular phrasing of this theorem is “a linkage exists that can sign your name”. This lecture summarizes our efforts to find simplified versions of curve-drawing linkages. A remarkable outcome is the ability to mechanically draw Bézier curves, which yields a linkage system that can sign your name, and even write cursive Chinese. I was honored to return to the University of Pennsylvania’s Grasp Lab after over 30 years to make this presentation. This highlights the excellent work Yang Liu.

Kinematics Lecture:  Six-bar and Eight-bar Linkage Synthesis

Kinematics Lecture: Six-bar and Eight-bar Linkage Synthesis

This is my lecture from the 2014 Kinematics Summer School organized by Anurag Purwar and is my best summary of the status of kinematic theory for the synthesis of planar six-bar and eight-bar linkages. It highlights the excellent work of Mark Plecnik, Kaustube Sonawale, Brian Parrish and Brandon Tsuge.

Design of Mechanical Walking Robots

Now Available on Amazon

In this book, we present the detailed design of mechanical walking robots that are driven by a single motor.  These walkers rely on specially designed leg mechanisms coordinated by gear trains in order to walk, rather than multiple computer controlled motors per leg.  The result is a simplified walking robot that provides a platform for other mechanical and electronic functions.

Two, four and six legged walkers are presented that implement different types of leg mechanisms and power trains.  In each case, we provide drawings for a laser cut wood or acrylic chassis, 3D printed parts and a complete parts list.  Several of the designs implement an additional motor for steering as well as electronic components and software for speed control.

Our goal is to provide enthusiasts of all backgrounds what they need to build a walking robot at home, to explore new design ideas, and, perhaps, to enjoy the operation of one of these robots as it moves across the ground.   

The paperback version is available from Amazon.

Design of Mechanical Walking Robots
Design of Mechanical Walking Robots

21st Century Kinematics: NSF Workshop

I am pleased to provide the presentations from the 2012 National Science Foundation Workshop on 21st Century Kinematics. These presentations provide insight to the challenges and opportunities for research in mechanical systems and robotics.

The NSF Workshop on 21st Century Kinematics at the 2012 ASME IDETC Conference in Chicago, IL on August 11-12, 2012 consisted of a series of presentations and a book of supporting material prepared by the workshop contributors.

The book is available at amazon.com: 21st Century Kinematics–The 2012 NSF Workshop.

And here are the presentations given at the workshop.

  1. Computer-Aided Invention of Mechanisms and Robots. J. Michael McCarthy, Professor, University of California, Irvine.
  2. Mechanism Synthesis for Modeling Human Movement Vincenzo Parenti-Castelli, Professor, University of Bologna.
  3. Algebraic Geometry and Kinematic Synthesis. Manfred Husty, Professor, University of Innsbruck.
  4. Kinematic Synthesis of Compliant Mechanisms. Larry Howell, Professor, Brigham Young University.
  5. Kinematics and Numerical Algebraic Geometry. Charles Wampler, Technical Fellow, General Motors Research and Development.
  6. Kinematic Analysis of Cable Robotic Systems. Vijay Kumar, Professor, University of Pennsylvania.
  7. Protein Kinematics. Kazem Kazerounian, Professor, University of Connecticut.
  8. Development of Fast Pick and Place Robots. Jorge Angeles, Professor, McGill University.
  9. Kinestatic Analysis of Mechanisms with Compliant Elements. Carl Crane, Professor, University of Florida.

It seems time to consider another similar workshop for 2022.

The Bored Robot: Controlling Two Drive Motors for a Walking Machine

Brandon Tsuge describes how to assemble the controller for two motors to drive the right and left sides of a walking machine using an RC transmitter and controller. See The Bored Robot: Using a DC Brushed Motor with a Rotary Encoder.

How to Fix SW Motion Analysis Error: Too Many Redundant Constraints

Kevin Chen,  J. Michael McCarthy,  Shaun Bentley

The design and assembly of our four-legged mechanical walkers can yield single degree-of-freedom systems with so many redundant mates that it stalls SolidWorks’ Motion Analysis.  For example, the walker shown in Figure 1 had 782 redundant mates.   The procedure outlined below reduced the number of redundant mates to 114, and Motion Analysis executed efficiently.

Four-legged Walker Assembly
Figure 1.  A four-legged mechanical walker consisting of a body, drive train, and four-leg mechanisms.

Our walker consists of a body, drive train, and four legs.  The legs mechanisms are identical but assembled as front-to-back mirror images.  The component parts of this walker mates were assembled using mates to align and coordinate various subassemblies, resulting in a large number of redundant mates.

In order to reduce the number of redundant mates, we dissolve the subassemblies, combine rigid elements, and mate new subassemblies as follows.

Step 1

Dissolve all of the subassemblies in the walker. To do this, hover over each assembly and select the menu item Dissolve Assembly. See Figure 1.

New Sub Assemblies
Figure 2.  Selected parts for new subassembly.

Step 2

Form new subassemblies for each leg, the drive train, and the body. See Figure 2. To do this, first, hover over the part, press “tab” to hide the part in order to identify it easily; and then, select all of the hidden parts, and right-click to open menu and select Form New Subassembly.

Form New Assemblies
Figure 3.  Within each new subassembly form subassemblies of parts that do not move relative to each other.

Step 3

Within each new subassembly combine parts that do not move relative to each other.  See Figure 3. The tree structure should consist of separate assemblies of rigid elements with the remaining mates between the assemblies. See Figure 4. 

Rigid Elements
 Figure 4.  The assembly should consist of subassemblies that move as rigid elements relative to each other.

Step 4

 Repeat Step 3 for all of the new subassemblies.  The result is shown in Figure 5.

Rigid subassemblies
 Figure 5.  The subassemblies that define the mechanical walker.  Notice that the tree structure consists of subassemblies and no individual parts.

Step 5

Delete the mates in the main assembly.  Introduce the mates required for movement using hinge mates, rather than coincident or concentric mates, where possible.

Step 6

Make the subassemblies at the top-level flexible.  Right-click on the assembly and select the flexible assembly icon  pastedGraphic.png.

The result of this procedure is a system with 114 redundant mates that Motion Analysis can process effectively.  The result is that animation shown below.

Construction of a Leg Mechanism

This is a series of four videos that show how to:

  1. Specify three positions for the foot of a leg consisting of a hip and knee joint;
  2. Use three position synthesis to design a four-bar function generator to guide the hip joint;
  3. Then use three position synthesis to design a second four-bar function generator to guide the knee joint;
  4. And finally assemble the linkage to determine the trajectory of the foot. Adjusting the lengths of the leg segments, the position of the hip, the specified positions of the input cranks, and the position of the coupler attachments to the input cranks vary the resulting foot trajectory. An example leg mechanism is shown at the end of this video.

Part 1:4 Setting up the design

Part 2:4 Synthesis of the hip function generator.

Part 3:4 Synthesis of the knee function generator.

Part 4:4 Assembly of the leg mechanism, exploration of design variations, and an example final leg design.

Walking Robot Student Project

Walking Machine Class Projects: Ohio State ME 5751

Prof Haijun Su at Ohio State University had his students design walking machines for their final project in ME 5751. Here are videos of four project teams from that event.

Team A:

Team B:

Team C:

Team D: