Skip to content
pluteski edited this page Apr 23, 2017 · 30 revisions

Welcome to the wiki for this project.

Stage3 paths found for northeast peg.

Origin Story

This began as an exploratory motion planning problem posed by Valkyrie Robotics, a FIRST Robotics team, for the 2017 competition.

See for example this robot cam view of a gear handling robot.

The challenge is to automatically find a path to goal from an arbitrary position and heading within striking distance of the goal.

This could be used to provide driver assist by moving the robot towards the goal along an efficient trajectory unless overridden by the driver.

Findings

See Background for background, summary results, and pointers to additional visualizations and simulation results.

See RelatedApproaches for approaches that are suitable for real-time motion-planning.

See Planning-Stages for a description of the motion planning algorithm.

  • A "quickturn" maneuver is a rotation about the robot vertical axis, changing the heading without changing the (x,y) location.

End Result

This approach is fully tested and ready for integration. It requires a state estimation system to predict the position and heading of the robot on the field 50 milliseconds in advance based on data obtained from several sensors, including vision, gyro, accelerometer, and wheel encoders. The team plans to work on it as an off-season project in preparation for the 2018 season.

The outcome was extremely insightful as to what could be achieved with circular arcs and quickturns*, which was useful knowledge for the driver team to have, especially because the motor control system of the robot was designed around circular arcs and quickturns.

MORE:

Background

Planning Algorithm

Welcome to Pluteski's Repo. Welcome to Pluteski's Repo (pronounced PLUTE-ski)

Clone this wiki locally