Walk Group 3

Goal
To make the robot figure out where the goal is based on the corners that it can see (i.e. corners of the field and the penalty box)

Deliverable (10/31): The robot will be able to know where the goal is even without looking at the post or crossbar.

Current Progress (10/2)
Ran some code on the robot to see how the current attacker works

Familiarization with the code base

Current Progress (10/16)
Switched from working on an attacker to creating a simulator for the robot.

Current Progress (10/28)
Simulator works with static images and detects the ball as well as the goal.

Need to interpret commands from Agent to move the camera and the ball. i.e. Walk, Turn, Kick.

Need to implement realistic motions of the head/camera as the robot moves/kicks/turns.

Goal
To make the default walk more stable by: adjusting stiffness values for the joints, adjusting the six parameters of the default walk, and pausing the walk while the robot is unstable

Deliverable (9/22): The robot will be able to walk a longer distance without falling (and hopefully always prevent itself from falling)

Parameters we are using
speed = 45 LengthStep = 0.035 StepHeight = 0.015 TurnByStep = 0.20 SidestepLength = 0.04 HipHeight = 0.21 TorsoYOrientation = 3.0

Graphs of the X and Y position and acceleration of the robot




Current Progress (9/19)
Goal reached!

 

The robot can walk indefinitely (or at least 400 steps) on a flat floor without falling

The robot can be gently nudged and it will recover

The robot can successfully walk over the hole in the soccer field most of the time

Current Progress (9/16)
Added balance detection

Robot will speak, stop, regain balance, and continue if walk becomes unstable

Got walk code from previous year

Current Progress (9/14)
Have configured our computers to run python scripts on the Nao robots.

Have become familiar with using modules

Have played around with different parameters to improve the Nao straight walk

Have looked into using the Nao's inertial sensors to detect balance