docs.md

TheWalkingDead Docs

• Resources
• Setup
• Run Walk Demo
• Run Task 1

Resources

• ROS_tutorial
• Writing a Simple Publisher and Subscriber (Python)
• Writing a Simple Service and Client (Python)

Setup

First, make sure your computer has scipy installed: Install Scipy

Also, install map server package:

sudo apt-get install ros-melodic-map-server

Download 3dcloud map file from our release section, place it under thewalkingdead/map: link

If roscore is NOT runing, run this in another terminal:

roscore

Before using thewalkingdead package, you need to compile source once:

cd { PATH }/ubt_sim_ws

# catkin_make under /ubt_sim_ws
catkin_make
source devel/setup.bash

Run Walk Demo

# Auth to use leg_motion
roslaunch leg_motion walker2_leg.launch account_file:={ Path }/user_account.json

# Run python script
rosrun thewalkingdead walk_demo.py

Use of Inverse Kinematics

# Build service, in ${ubt_sim_ws}
catkin_make

# Run service
source ${ubt_sim_ws_HOME}/devel/setup.bash
roslaunch thewalkingdead solver_server_node.launch urdf_path:=${path_to_walker.urdf}
# [ INFO] [1592677570.580549330]: Ready to Solve.

# IK Service name is "inverse_kinematic_solver"
# FK Service name is "forward_kinematic_solver"
# Message type is "Solver", read more in thewalkingdead/srv/Solver.srv
#   Note that both IK and FK share the same service, see ik_demo.py for 
#   usage of both.

Run Task 1

roslaunch thewalkingdead task1.launch

Run Task 2

roslaunch thewalkingdead task2.launch

Run Task 3

roslaunch thewalkingdead task3.launch

Notes:

• 增加了位移偏差末端补偿系数，提高了抓取 4/5 的准确率
• 有待于做更多测试，有小可能会失败

Run Task 4

roslaunch thewalkingdead task4.launch

Run Task 5

roslaunch thewalkingdead task5.launch

Run Task 6

roslaunch thewalkingdead task6.launch

Run Task 7

roslaunch thewalkingdead task7.launch

Run Task 8

roslaunch thewalkingdead task8.launch

Run Task 9

roslaunch thewalkingdead task9.launch

Run Task 10

roslaunch thewalkingdead task10.launch

优化：

• 最后几步可以增加 阻尼系数，从而快速拉开完成比赛，缩小用时

Run Task 11

roslaunch thewalkingdead task11.launch

Run Task 12

roslaunch thewalkingdead task12.launch

Run Task 13

roslaunch thewalkingdead task13.launch

Run Task 14

roslaunch thewalkingdead task14.launch

优化:

• 可以试着增加阻抗提高重心稳定度，然后可以增大脚步，以减少比赛用时

Python Dependancy:

• from scipy.interpolate import KroghInterpolator

Run Task 15

roslaunch thewalkingdead task15.launch

Notes:

• 6秒左右完成任务
• 没有做优化， 具有不稳定性，有一定可能会摔倒
• 采用倒立摆模型
• 除了Ref中的dynamic时候的步态规划， 还分别增加了起步与止步的步态规划
• 参考:
  • 双足机器人爬楼梯步态规划与参数 吴飞 （文中起码有超出10处错误）
  • 双足机器人步行仿真预实验研究 史耀强

Python Dependancy:

• from scipy.interpolate import KroghInterpolator

Navigation notes:

• 融合Imu参数提高Odometry精度
• 研究曲线走法