Robot Leg Test Bed Experiment and Strain Data Collection

Voiceover

Presenters/Authors(s)

Katie Sugg

Abstract or Description

The intersection of robotics and neuroscience has allowed innovation in legged robot locomotion. Mimicking biological systems in locomotion requires performing ambulatory tests on different types of robot legs. Currently there is not a user friendly or easily implemented testing system that allows non-engineers (e.g., neuroscientists) to perform robot leg tests. This research goes through the requirements and steps of creating a testing method or setup for running experiments on robot legs. It looks at the ways to simplify complex steps and describes an interface that allows for easy usage by non-engineers. To test the usability of this testing system, users will perform a robot leg stepping scenario. The scenario will be going through the steps of creating a footpath to test on a robot leg and collecting strain data. Specifically the steps would be footpath coordinate generation with Bézier Curves, joint angle kinematics, stepping the leg with Dynamixel servos, and strain data collection. After the scenario is completed there will be a survey that gets usability feedback and ways to improve the system. Preliminary feedback and results show that every participant was able to complete the leg stepping scenario in under 15 minutes and were able to complete each step in the experimental process. Future improvement of the testing system would include streamlining more of the coding environment setup and including a treadmill for the robot leg to step on.