Student Work

3D Printed Humanoid Robot Project

Public

Downloadable Content

open in viewer

Research on humanoid robots often involves prohibitively expensive hardware. Using modern manufacturing techniques and off-the-shelf components, a small-scale humanoid robot can be manufactured at a much lower price point while still demonstrating many of the same capabilities as larger robots. Such robots can be used in many different applications including rehabilitation, research, and teaching. This paper presents a 4 kg, 85 cm tall humanoid robot with 27 degrees of freedom based on the open-source Poppy Project. The robot is constructed from DLP/SLA printed components that are easily modified to support alternative motor choices. Resin 3D printing is used to minimize weight via pocketing and allows types of resin to be mixed in ratios based on whether a part needs stiffness or overall strength. Each joint is controlled by one of several types of smart actuators with integrated position control. The Poppy project exclusively uses 25 Dynamixel motors, which make up the bulk of the robot’s cost. We replaced most of these motors with the less expensive HerkuleX DRS-0201 motors that have comparable form factor and power to the Dynamixel MX-28. Physical changes included adaptors for the HerkuleX motors to mount to the preexisting mounting patterns. Pivoting away from the Poppy Project’s exclusive use of the more expensive Dynamixel motors required the codebase to be written from scratch. An onboard Arduino communicates with the different actuators, while a Raspberry Pi performs higher level processing. The Arduino supports multiple actuators using different communication buses. It provides a layer of abstraction allowing the Raspberry Pi to control motors without accounting for the specific actuator or motion limits at each joint. The robot now uses batteries to allow for completely untethered operation. It is able to walk with human assistance but has the capability for future implementation of self-balancing for unassisted walking. Control is based on a primitive system that allows multiple actions to occur simultaneously. This allows either multiple actions using different parts of the body (walking and waving arm), or by averaging positions to combine actions utilizing the same part of the body (a normal walking gait and adjustments to terrain based on sensor feedback). We demonstrate the robot’s functionality using a record/play system that allows a human operator to manually position the robot and then record those positions for later playback along with functions that highlight the capabilities of individual systems of the robot. The paper will describe the implementation, challenges faced, and future work to further added capabilities.

  • This report represents the work of one or more WPI undergraduate students submitted to the faculty as evidence of completion of a degree requirement. WPI routinely publishes these reports on its website without editorial or peer review.
Creator
Publisher
Identifier
  • 65291
  • E-project-042822-123507
Keyword
Advisor
Year
  • 2022
Date created
  • 2022-04-28
Resource type
Major
Rights statement

Relations

In Collection:

Items

Items

Permanent link to this page: https://digital.wpi.edu/show/dr26z168n