From Noisebridge
Jump to: navigation, search

Noisebridge is a space to hack and learn how to build/fix/create things. If you're not interested in that path for your project, post a detailed description of your gig and someone will get back to you.


Want someone else to fix your computer screen? Manufacture a part on the 3d printer? Make high precision laser cuts to wood or clear plastic? Weld your steel pieces together? Post on a gig board!

Contact with makers at Noisebridge to get your idea/problems off the ground.

Please include below, and someone might get back to you:

- Your Name
- Email
- Phone Number
- Gig Description
- Skills Required
- Cost Willing to Pay for the Gig

List your gig here[edit]

Building a Prototype for a Parallel/Delta/Pick and Place Robot[edit]

Name: Abhinav Kulkarni

Email: <firstname><lastname>


Phone: Send me an email



I am looking to partner with somebody (or contract the job out) to create a small prototype of a parallel/delta/pick and place robot [1].

We wish to demonstrate the following:

1. We are able to pick and place soft, delicate objects such as fruits and vegetables. We plan to use pneumatic soft grippers for that [2].

2. We are able to handle a variety of shapes - round (tomato), elongated (zucchini), cylindrical (oatmeal box), etc. We will constrain ourselves to lighter loads initially (mostly <1lbs, at max 1-2lbs).

3. We are able to control the robot precisely using an external Computer Vision based system that senses objects using cameras and/or structured light (such as Microsoft Kinect) and/or multi-view geometry. We are a group of engineers and entrepreneurs with a good depth in Robotics, Machine Learning, AI and Vision, so we really need the freedom of programming the controller. For the prototype, the objects will be placed in the robot's vicinity and the robot will be instructed to get closer to and pick them. In a warehouse kind of situation, the objects would be placed on a conveyor belt.

4. We really don't need micro/millimeter level of precision of industry-grade robots. We also don't need high speed such as here [3]. If the robot can pick 1-10 objects per minute, we would be happy.

Here is a sample draft of tech specs for a similar robot [4].


1. Build a parallel robot. You can refer to Instructables listings or to this forum post [5]. The robot needs to have a workspace coverage of approx. 1.5ftx1.5ft. At the completion of the first milestone, the robot should be able to pick and place a round pancake sitting right below the robot using a simple vacuum gripper. It will be supplied with the exact coordinates and/or path by a control and path planning algorithm. It simply needs to demonstrate the ability to execute the path. The robot will be controlled by an Arduino (we can help with Arduino interfacing).

2. Calibrate the robot/achieve a certain level of precision/repeatability/accuracy so that it can be automatically calibrated from time to time. The second milestone should be able to pick and place 2-3 pancakes lying around the robot and stack them at a specified position (we will design the control and path planning algorithm, the goal of this milestone is to test whether the robot is able to execute instructions it is supplied with).


The payments will trigger as the milestones are achieved. The Igus robot [6] retails for about $6200 (without controllers and you have to self-assemble). So, that should give an estimate of the appropriate budget for this project.

Thanks, Abhinav