Tuesday, October 28, 2014

P.U.R.E Midterm Report

A slide show of 4-5 slides detailing our progress needs to be prepared for P.U.R.E.

The slides should include:

1) The research statement/Motivation (Why is this topic worth investigating?)
2) Basic information about the research
3) What you have learned and done so far
4) What needs to be done next

Friday, October 24, 2014

Interesting Links for Finding Hardware

Robot Shop
Advantage Hobby



Tendon Driven Robotic Arms

Tendon Driven Robotic Arms

This design allows for 3 axis movement using a tendon driven system. This should help us generate a design for the joints.




Tuesday, October 14, 2014

Inner Tendon Design


Placing the tendon on the inside of the joint removes the need for a bridge connection above, cutting down weight and bulk.  To achieve adequate tension, re-looping the tendon at the base acts like a pulley and requires longer draw but with lower peak tension.  Ample torque is generated for supporting the leg and some portion of the chassis.




Monday, October 13, 2014

Outer Tendon Design


As we continue to develop a design, we began prototyping leg designs using K'NEX. The purpose of the joints are to allow the servo to bend each leg segment using a tendon like string. The joints need to reduce the force the servos apply in order to prevent the servos from burning out. Also, the joints need to limit leg movement to two-axis. Otherwise, the leg would flex unpredictably and would be far more difficult to control. To accomplish these task, we built several prototypes of increasing complexity.

First Prototype: Pulley System



For this prototype, we focused on creating a system to reduce the amount of force needed to operate the joint. On the underside of the joint an elastic band is attached. This is what restores the joint to its natural, closed state when no force is applied to the servos. The string tendon is attached to the top of the joint in a pulley system. This system still requires the same amount of work to operate, but replaces a high force load (yet shorter amount of string) with a low force (but a lot more string to wind in). The optimal ratio that we found for this setup was 4:1. After that, the additional strands resulted in too much additional friction and prevented the joint from properly restoring itself (if different materials are used, it is likely that this result will vary). The joint in  operation can be viewed below.

2nd Prototype:Double Joint


As the full leg will have a minimum of two joints on it, a system had to be devised so that each joint can move independent of the others. To achieve this, we relied on the principles of torque. The joint works because as the string tightens it exerts a force on the two struts at each joint. The force is multiplied based on the distance it is from the joint (the farther the distance, the greater the multiplication factor). This, along with the pulley system, reduces the amount of force that servo needs to apply in order to manipulate the leg. The downside of this  is that any small force that is applied to the struts is then amplified and placed on the joint. Therefore, if we ran multiple lines through the strut, each line would greatly affect the joint, making the system dependent to multiple inputs. In order to counter this, the second line can be run through the center of the joint. As this is 0 distance from the joint, it applies 0 torque, leaving the joint unaffected as additional strands are tightened or loosened. This systems allows us to control each joint separately, greatly increasing our control over the leg. This system can be seen in action below.

Complete Leg


The completed leg features longer struts in order to better mimic the anatomy of the opiliones. As this added more weight, the joints had to be reinforced to prevent them from collapsing. Also, a third joint
was added at the bottom to mimic a foot that adapts to the slope of the ground surface with respect to the leg. This provides the leg with greater traction with the surface, helping to prevent slipping. The finished design appears in action below.

Friday, October 3, 2014


Now this is what we need!!
(adaptation to unknown terrain with pressure sensitive feet)

Sunday, September 28, 2014

Project Opiliones


Members:

Advisor

Stephen E. Levinson, selevins

Mentor

Luke A. Wendt, wendt1

Mentees

Shivani Iyer, siyer10
Shivam Bharuka, bharuka2
Rebecca Cole, rjcole2
Emily Dixon, edixon4
Sahil Kumar, skumar36
Shanay Jhaveri, sjhaver2
George Moffitt, gmoffit2

Volunteers

Constantine (Dean) Roros, roros2
Amish Ralhan, amish.ralhan.12@ucl.ac.uk
Tara Tripp, ttripp2


Opiliones – The Harvestmen


Lightweight Elastic-Tendon Driven Legs
with Sensor Rich Body


Project Goals:

General

·         Make robotics more accessible to undergraduate research.
·         Emphasize fast build time, e.g., modular hardware solutions (avoid custom hardware)
·         Emphasize fast code time, e.g. user friendly IDEs, APIs and Libraries
·         Emphasize low cost, e.g., Arduino based designs
·         Emphasize low complexity
·         Build sensor rich systems


Specific to Opiliones

·         Implement advanced autonomous choreography, i.e., intelligent walking gaits driven by user input that plan around sensor feedback
·         Build a durable platform for research in embodied cognition http://en.wikipedia.org/wiki/Embodied_cognition

Optional

·         Possibly offering a new course in the ECE curriculum:

ECE ??? – Robot Dynamics & Control with a Lab
There are a handful of robotics and control courses offered to undergrads here at UIUC such as ECE 470 – Introduction to Robotics (which focuses primarily on coordinate systems and configuration space), GE 423  Mechatronics (which focuses primarily on wheeled robots), and ECE 486  Control Systems (which introduces most of the necessary tools for linear control).  However, there is no course (with the exception of maybe senior design) that then combines this material into a focused application.  In particular, there is no coursework covering the nonlinear modeling and control aspect of robotic systems, e.g., Chapters 6-10 of Robot Modeling and Control by M. Spong, S. Hutchinson and M. Vidyasagar.  If an additional course were taught, what would the accompanying lab look like?  Being able to rapidly build a skeletal structure and control it with low cost motors, sensors and microcontrollers, the student could be provided with a kit and given the task of building a robotic system that implements the subject matter.  The specific application maybe be flexible to the students interests, but would require some novel aspect.

Inspiration Source 1:
Hand Built Robotics

(no machine shop or plastic printers required)

A composite construction that uses thermoplastic (which can be worked by hand), servos, tendons, elastic, nuts and bolts provides a fast, cheap and low complexity solution to rapid prototyping of robots.  Capable of generating complex actuation.





Inspiration Source 2:   
Hexapods

There are many designs readily available.  With 4 legs and up, a stable triangular basis can be present at all times.  With 6 legs, walking solutions involving alternating triangular patterns become very easy to implement. 



Inspiration Source 3:   
Octopods

With 8 legs, even more complex choreography becomes possible!