naoth
Table of Contents
- compile source code for new robot with atom
- GermanOpen 2012
- RoboCup 2011
- IMU
- Iran Open 2011
- German Open 2011
- fast walk forward
- Binary for last day [100%]
- Binary for 2nd round in simulation league [100%]
- arm motion of walk
- walk: rotation in side
- call xixi
- better clamp walk request
- Binary for 1st round in simulation league
- test walk parameters, better parameters?
- Test binary for simulation [75%]
- Walk parameters
- step adapt
- stand motion with given speed
- foot soft landing
- adjust just walk for presentation tomorrow
- preview control walk?
- create a wrapper for AL's motion
- try Aldebaran's walk (by python script) - indicates which robot is working
- RoboCup Project Proposal
- Explore the limits of motions (constraints)? PAPER?
- the connection between different motions PAPER?
- kick between simulation and real robot
- Robot Health
- workshop in the friend game against FUmanoid
- Make the old walk and kick usable
- Calibration
- A public letter about qualification
- AUTCUP
- bugs
- improvement
- avoid goal ( get into the field behind the goal)
- potential field
- several points for each robot in potential field
- keep distance to ball while walking to position
- obstacle avoidance: do not slow down so much (move the target point a little far away)
- reduce the distance for no obstacle avoidance…
- the arm! don't move in y direction?
- play against Apollo
- Visa
- Flight –> print
- insurance
- run the walk evolution in SimSpark again
- read the walking code from B-Human
- NaoSimSpark
- Motion Engine
- RoboCup 2010
- Unify the interface for different platform
- Motion Visualizer
- German Open 2009
- binary size
- port the NaoTH source code to the SimSpark [85%]
- looking through the source code
- do some simple test on the real robot
- IMU [100%]
- imporve kick paper
- Nao
ONEDAY compile source code for new robot with atom
[X]check the changes in repository and email from Thomas[X]compile trunk[X]try branch for atom[ ]almost, but some warnings (treated as error)
GermanOpen 2012
kick
test parameters for all robots
left and right foot are different?
- there is an error in our code?
- because the com is not in the center, we reach the limits of joint in right side?
walk
parameters for all robots
- Nao1555
best sometime the right ankle joint disabled after restart
- Nao2190
stable -> move too much?
- Nao2189
shifting to left – height 260 –> better rotation stablizer threshold affect waving right FSR dead
- Nao1670
walk ok, but can not kick with left foot right knee get overheated after ~5min
- Nao1671
walk good, wide step width for turning around ball
check the keep foot stuff
try different stiffness, which one comsum less energe
try spline foot trajectory
DONE protect fall down
stiffness to 0 (not -1)
Behavior
- test error range stuff
- dribble when ball is very close to goal
RoboCup 2011
Simulation 3D league
server enhancement
- CANCELED we need a powerful computer!!!
- DONE single step debugger
- DONE try to disable enableRealTimeMode
- DONE try to disable enableRealTimeMode
- DONE use roboviz visualize our team
motion
- real dribble (control the direction)
- kick
- strong
- fast
- strong
- goalie
- sliding tackle
cognition
- behavior
- formation
- marking
- passing
- formation
- ball model (prediction)
Standard Platform league
fall protection
fast & stable walk!!!
- stability indicator
different types of walk
- balance between speed and stability
- special foot steps
connection between walk and kick
DONE print IK test
DONE Iran Open 2011
Game day 2 [April 8, 2011]
DONE 3D
- time to ball
[ ]ball prediction (ball model)[ ]time to ball (prediction)[ ]update motion speed[ ]face to opp goal
- DONE move around ball!!
- DONE walk parameters!
[X]move the arm less
- init pose
[ ]two middle fielder a little back
- go to strategy position facing to ball all the time!!
SPL
- walk minimum step size
- Team Comm RoboViz
- kick stabilization
- initreadysetplay
Game day 1 [April 7, 2011]
3D
- DONE walk parameters
- DONE turn around ball
- DONE go to position: preview…
SPL
- walk
- DONE parallel moving foot
- stablity indicator
- DONE parallel moving foot
Set-up day 2 [April 6, 2011]
3D
- time to ball
[X]go to ball time (general, static)[X]fallen down state
- DONE walk parameters
- CANCELED test binary, check the schedule
- check self location
- binary for round 1
SPL
- DONE Walk: test parameters for all robots
walk.ZMPOffsetSpeedX = 0; walk.ZMPOffsetSpeedY = -5; walk.ZMPOffsetY = -4; walk.bodyPitchOffset = 0; walk.bodyRollOffset = 0; walk.comFootDiffTime = 0; walk.comHeight = 260; walk.doubleSupportTime = 40; walk.enableDynamicStabilizer = false; walk.footCurveFactor = 10; walk.footOffsetY = -3; walk.footPitchOffset = -1; walk.footYawOffset = 0; walk.leftFootRollOffset = 0; walk.leftHipRollSingleSupFactor = 0.6; walk.maxStepLength = 40; walk.maxStepLengthBack = 30; walk.maxStepTurn = 30; walk.maxStepWidth = 40; walk.maxTurnInner = 10; walk.rightFootRollOffset = 0; walk.rightHipRollSingleSupFactor = 0.6; walk.singleSupportTime = 340; walk.startingPhaseTime = 500; walk.stepHeight = 7; walk.stiffness = 1; walk.useArm = false;
Set-up day 1 [April 5, 2011]
3D
[X]git commit[X]trun around ball[X]robviz test[X]robviz server, collect information from UDP[X]visit the simulation area
SPL
[ ]walk in the field –> soft foot trajectory (8, 6)[X]walk assert line 584: when maxstepturn == 0
DONE German Open 2011
fast walk forward
DONE Binary for last day [100%]
[X]communicate robot position when fallen down makes problem?[X]potential field in front of goal (parameters?)[X]go to position from Thomas![X]use go to relative point instead of gotoballcool in dribble[X]check allowed enter penlaty in formation[X]faster walk –> how?[X]self locator- stand up when fall down in before kick off
[X]defender formation
DONE Binary for 2nd round in simulation league [100%]
[X]defender[X]defender pose –> previewed model[X]better defender position[X]0.5m outside of penalty area
[X]better attack formation[X]fix problem of running around ball –> previewed ball[X]check set pieces[X]own goal kick[X]opp goal kick –> configure file
[X]better walk parameters –> the acceleration affects!!![X]back walk fallen down!!
[X]check self locator[X]ball model in simulation[X]turn around ball with short distance[X]after dribble hit –> bigger distance?
DONE arm motion of walk
DONE walk: rotation in side
DONE call xixi
DONE better clamp walk request
DONE Binary for 1st round in simulation league
[X]self locator crash!! –> intersection problem[X]better walk parameters
DONE test walk parameters, better parameters?
DONE Test binary for simulation [75%]
[X]self locator[X]walk parameters –> bad[-]stand up motion[ ]last key frame[X]put a short stand motion in the end
[X]acc/inertial sensor
sc - [X] test team locally
[X]test team in server
DONE Walk parameters
DONE step adapt
DONE stand motion with given speed
CANCELED foot soft landing
DONE adjust just walk for presentation tomorrow
DONE preview control walk?
DONE try the algorithm in python and plot, analysis the result
DONE try implemented it in c++ and test it in real robot
DONE make the code work in simulation
CANCELED create a wrapper for AL's motion
DONE try Aldebaran's walk (by python script) - indicates which robot is working
the robot walks perfectly!
DONE RoboCup Project Proposal
Robot Health
DONE workshop in the friend game against FUmanoid
DONE Make the old walk and kick usable
CANCELED A public letter about qualification
ideas
- the contribution of team
- scientific (publications)
- the community (code release)
- the historical activities
- distribution of teams in different countries
- the 1/X rule
- results in local events
- different kinds of teams
- new team
- old team which didn't participate at last RoboCup
- old team which participate at last RoboCup
DONE AUTCUP
bugs
ball size
team communication (antenna)
check parameters saving & loading: motion parameter & camera
–> the problem is grasping create an instance of motion parameter!
improvement
avoid goal ( get into the field behind the goal)
potential field
several points for each robot in potential field
keep distance to ball while walking to position
CANCELED obstacle avoidance: do not slow down so much (move the target point a little far away)
CANCELED reduce the distance for no obstacle avoidance…
DONE the arm! don't move in y direction?
play against Apollo
- with rcssmonitor3d 0:2
- with teamcontrol 4:0
DONE Visa
DONE Flight –> print
DONE insurance
DONE run the walk evolution in SimSpark again
DONE check the result
DONE compare the new parameters with old ones
It is too slow!
CANCELED check if everything works well
DONE read the walking code from B-Human
DONE the analytical inverse kinematics
NaoSimSpark
- FSR
- DONE FSR as webots
- DONE FSR as webots
- DONE re-enable the interface
- DONE compile the simspark in windows
- DONE compile the latest simspark with patch
- DONE compile the NaoTH simspark interface
- DONE compile the simspark in windows
- DONE basic communication interface
[X]connect[X]init[X]receive sensor data & parse[X]send joint command
- DONE Camera
[X]apply the patch of camera[X]receive the image from simspark
DONE Motion Engine RESEARCH
- FSR calibration
- stabilizer for walking
- summarize available methods
- CANCELED record the log file while falling down
- try to stabilize the walking
- summarize available methods
- DONE use the same framework as cognition(at least for accessing black board)
- DONE architecture reading
- CANCELED read the walk source code in NaoTH-2008
- DONE inverse kinematics walking [100%]
[X]check out the svn[X]create the InverseKinematicsMotionEngine[X]simple gait generation[X]walking in webots[X]walking in real robot
- DONE slove the problem of hip joints binding
[X]find the relation of feet and chest[X]try "don't control the rotation of chest"
- DONE improve inverse kinematics
- DONE use CCD to slove hip
- DONE slove hip in other method
- DONE check the calculation time of IK
- DONE use CCD to slove hip
- DONE adlebaran walking
- DONE inverse kinematics [100%]
[X]read the source code[X]create the test source: just random and print the error out[X]make the code work[X]optimize the code[X]commit it to the svn
- DONE support polygon
- DONE ZMP calculation
- DONE calculate the COM
- DONE improve the forward kinematics
DONE RoboCup 2010
formation: move faster
calculate is strike
- DONE don't go to ball while teammate occupy the ball
[X]update teammate model by vision
- consider the time (not the distance)
Walk:
- slow down when some opponents near by (sablizer?)
- new parameter: the evolution is running now
- path planning
[ ]debug[X]obstacle avoidance
goalie motion
[ ]ball model –> prediction
CANCELED kick: it should be powerful
[ ]integrate side kick
head motion
[ ]look at ball while standing up[X]search ball: the head doesn't move!!! –> the speed of head joint[X]look at ball: not the center of camera –> it doesn't matter
supporter
[X]goto position
self locator
[X]beam as less as possible
make the stand fast
[ ]first frame of stand up –>[ ]stand firstly in walk –>
Dribble:
- DONE imporve potential field
DONE Search Ball: don't use the one as real robot
DONE communicate one opponent's position
- DONE chose the one that wasn't communicate longest time
Team Behavior Viewer [100%]
[X]basic 3D viewer[X]update frames[X]commnuicate with robots[X]draw debug information
DONE Get up motion: make it stabler
DONE Unify the interface for different platform
- DONE make the abstract interface class
- DONE implement the interface class for Webots
- DONE implement the interface class for Simspark
- DONE implement the interface class for real Nao
- DONE implement the interface class for LogSimulator
DONE Motion Visualizer RESEARCH
- DONE integrate into RobotControl
- DONE check java 3d is installed or not
- DONE draw line
- DONE draw text in 3D
- CANCEL draw text in 2D
- DONE good field
- DONE check java 3d is installed or not
- DONE add time stamp
- DONE use new V3 mesh
- DONE show which foot is support foot
- DONE check the mesh from webots
Lower arm length is 110mm in webots, but 130mm in real Nao
- DONE show FSR
- DONE motion debug request
- DONE add fancy ground, orignal, sky, etc
- DONE test on the real robot
- DONE commit in the first working version
- DONE the problem of communication
- DONE log file support
DONE German Open 2009
- DONE let ALMotion stand up firstly
- DONE height & rotation of stand
- DONE interpolation while switch between differen motion egines
- DONE gotoAngles – set almotionstate to running
- DONE getBodyAnglesError – check it & interpolate
- DONE gotoAngles – set almotionstate to running
DONE binary size
due to debug
DONE port the NaoTH source code to the SimSpark [85%]
[X]send an email about the suggestion and work to the nao-team and professor Burkhard[X]send an email to Joschka, discuss about the vision(image) sensor[X]if everything is ok, then do the work:[X]basic framework[X]joint motor set & get[ ]kinds sensors[X]image sensor
DONE looking through the source code
[X]find the main modules[X]know how the system works[X]the debug system[X]the NaoQi interface[X]the motion system[X]the cognition system[X]the black board content[X]draw sequence diagrams[X]write a mail about notes and questions to nao-team
DONE do some simple test on the real robot
DONE IMU [100%]
[X]patch simspark and NaoController to support real value[X]compare current inertial sensor data and real value- the static situation is okay (use accelerometer)
- but the dynamic situation is not okay (use gyro)
[X]take a look at FU's IMU[X]take a look at Olive's IMU[X]try use filter in IMU[X]experiment and report
DONE imporve kick paper
DONE print & take a look at the comments
DONE add comments to the paper
DONE read comments and mark the suggestion
question about stablization
CANCELED Early Bird Registration
DONE Nao
DONE find the ball
[X]look the code of Aibo looking at ball
DONE qualification
[X]write Team Description Paper[X]publications[X]revise the paper with help of Ms Tan[X]submit to spltech@tzi.de
read some articles and TDPs for image processing [%]
[ ]TDPs in RoboCup 2007[ ]"Digital image processing" Book[ ]some extral articles[ ]wirte some issue about vision in rcssserver3d