A sample Arduino project created by Yevgeniy Yesilevskiy for demonstration in the ME350 course. The player uses a joystick to move a character side to side while they try to catch falling lights. If a light is caught, the bottom flashes white. If a light is missed, the bottom flashes red. The color of the falling lights is chosen by putting a material in front of a color sensor.
RAMone's First Steps and Obstacles
In this video RAMone walks using a modified version of the walking controller proposed by Jerry Pratt. An advantage of this controller is that it is force based so it handles foot disturbances gracefully. We show this disturbance correction by placing hard and soft objects under RAMone's feet as it walks.
A Comparison of Series and Parallel Elasticity in a Monoped Hopper
In this project, Series Elastic Actuation (where the motor moves the location of the top of the spring) and Parallel Elastic Actuation (where the motor contract the entire leg length) are compared by looking at energetically optimal trajectories and parameters.
Control of a Robotic Biped SURE 2015
This is a summary of Kevin Green's work in David Remy's RAMLab. He worked on control of the biped RAMone. RAMone is the RAMlab’s first bipedal robot. It is based on the ScarlETH leg design and used to study efficient locomotion.
Passive Dynamic Quadrupedal Model Explain a Large Variety of Gaits
Six different gaits including trotting, pacing, walking, tolting, bounding, and galloping were identified by the conceptual simplistic quadrupedal model with elastic legs.
Simplistic Model VS Headed Mode
In this video, we show that walking, trotting, and tolting (with : above o) gaits can be created completely passively by using both simplistic and headed quadrupedal models with elastic legs.
Walking in Bilateral Ankle Exoskeletons
Proportional myoelectric control allows us to synchronize the exoskeleton actuation with the user's own movement.
Testing the Smart Braid with a 4 Hz sinusoidal pressure signal.
Saturday Morning Physics Lecture
C. David Remy gives a Saturday Morning Physics Lecture on October 11, 2014. Title: "Dynamic Locomotion in Humans, Animals, and Robots"
ME 543 Final Project
This is an animation of the final project for ME 543 "Analytical and Computational Dynamics I". In this course, students create their own multibody dynamics engine in Matlab. Shown is the forward dynamic simulation of a robot arm that is programmed to follow a straight line, and what would happen if the robot would get tangled in a wire and is constrained to move within a sphere.
This movie shows the four legged robot StarlETH trotting over obstacles placed in its way.
How to Engineer a Robot Dog
This movie illustrates part of the work that was done in the past years during Prof. Remy's time at the Autonomous Systems Lab. Inspired by nature, the team built different types of electrically driven robots that can do various maneuvers from slow and careful climbing to very robust dynamic trotting. This video explains how it is done.
Concept of Limit Cycle Locomotion
This movie illustrates the basic concept of limit cycle gait creation, in which a periodic gait is considered to be a limit cycle in state space, and analyzed within the Poincaré Section. Periodic orbits manifest as fixed points of the Poincaré map. See Remy, C.D., 2011, "Optimal exploitation of natural dynamics in legged locomotion", PhD Thesis
Passive Dynamic Gait Transition
A passive dynamic quadruped walking in a four-beat gait, is slightly disturbed from its periodic orbit. Due to its instability, it passively transits into a stable two-beat gait. See Remy, C.D., Buffinton, K.W., and Siegwart, R., 2010, "Stability analysis of passive dynamic walking of quadrupeds", The International Journal of Robotics Research
AloF, a robot built during Prof. Remy’s time at the Autonomous Systems Lab, showing a walking gait. See Remy, C.D., Baur, O., Latta, M., Lauber, A., Hutter, M., Hoepflinger, M.H., Pradalier, C., and Siegwart, R., 2011, "Walking and crawling with ALoF: a robot for autonomous locomotion on four legs", Industrial Robot: An International Journal (Video speed up 3x)
Due to its large range of motion, AloF is able to perform a crawling gait, in which the robot is alternatingly supported by its shanks and its main body. This gait was employed during the Lunar Robotic Challenge of the European Space Agency. See Belo, F.A.W, Birk, A., Brunskill, C., Kirchner, F., Lappas, V., Remy, C.D., Roccella, S., Rossi, C., Tikanmäki, A., and Visentin, G., 2012, "The ESA lunar robotics challenge: Simulating operations at the lunar south pole", Journal of Field Robotics (Video speed up 2x)
To the best our knowledge, AloF is the first quadruped robot on the world that is able to stand up after falling down. Here it is showing a full flip, getting back on its feet. See Remy, C.D., Hutter, M., Hoepflinger, M., Bloesch, M., Gehring, C., and Siegwart, R. , 2012, "Quadrupedal Robots with Stiff and Compliant Actuation", at-Automatisierungstechnik (Video speed up 5x)
This animation was created by numerically optimizing the electrical cost of transport (including thermal losses and positive mechanical work). Optimization was initialized from standing still. At slow speeds the optimizer converged to a walking gait. See Remy, C.D., Buffinton, K.W., and Siegwart, R., 2011, "A MATLAB framework for efficient gait creation", International Conference on Intelligent RObots and Systems, IROS, San Francisco, CA
Similarly, for higher speeds, the solution converged to a running gait. See Remy, C.D., Buffinton, K.W., and Siegwart, R., 2011, "A MATLAB framework for efficient gait creation", International Conference on Intelligent RObots and Systems, IROS, San Francisco, CA
Simulation of Walking
This forward dynamic simulation of walking is based on actual walking data recorded via motion capturing in a gait lab. Using a Computed Muscle Control (CMC) and Residual Elimination Algorithm (REA), muscle activations where extracted, that are driving the simulation. See Remy, C.D. and Thelen, D.G., 2009, "Optimal estimation of dynamically consistent kinematics and kinetics for forward dynamic simulation of gait", Journal of biomechanical engineering
Moonwalker - It walks almost everywhere
Moonwalker walks everywhere as long as you gives it a slope of degrees from 7 to 10. In the video, the Moonwalker walks on the slope outside EECS building in University of Michigan Campus, the ramp we set up in Mechanical Engineering Assembly Lab, and a random branch we raised up in the duderstadts connector. It's really easy and consistent to let it walk, and it can walk a long distance. Build one and try it out!
Moonwalker - Assembly Process
This video shows the whole assembly process of building a Moonwalker.