All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog and this project adheres to Semantic Versioning.
- the possibility to publish odom and tf from the robot odom estimator
- wolf_controller_utils
- use desired contact forces for the odom estimation, this seems to solve the z drift problem
- Postural task name
- height control when going un slopes using the ground_truth ( or any other absolute reference for the base )
- now visualization of the capture point info is done wrt the terrain for the z coordinate
- imu accelerometer filter and rqt interface
- controller state message
- des_contact ins ContactForces msg
- odometry with wolf_estimation inside the controller
- contact force minimization tasks, qddot minimization task
- function naming inside state estimator to reflect the same functions in wolf_estimation
- increase to 1khz the publishing frequency of the joint state publisher
- bug about imu angular accelerations in wolf_gazebo_interface
- reset input command if inactive
- a bug about the command velocities, they were not as the same generated by the foothold planner
- clean trajectory waypoints if control mode is changed
- clean waypoints if control mode is changed
- wolf_msgs
- robot_semantic_description changed to robot_description_semantic to be compliant with MoveIt!
- ramp up the desired joint position instead of the desired torques when standing up
- support polygon scaling inside push recovery
- make rt_gui dependency optional in package.xml
- isFirstStep and isLastStep in gait_generator
- export activate_com_z, activate_angular_momentum, activate_postural, activate_joint_position_limits, regularization
- add the possibility to activate the postural, com height control, angular momentum and joint position limits
- joint position limits
- trajectory interpolation and waypoints for the cartesian tasks
- update odomPublisher to use tf2 instead of tf
- reduce torque limits to 0.9 in the solver for safety
- avoid publishing duplicated transforms in the odomPublisher, this solves the "Warning: TF_REPEATED_DATA" issue
- fix state_estimator estimation_t deduction from string and vice-versa
- avoid segfault when activating cost
- fix reference frame for the foot tasks
- fix regularization for point contact model
- Set default_step_reflex_max_retraction value to set the maximum step retraction when the step reflex is active
- Mux with priorities
- Spacemouse support
- License
- activate push recovery and step reflex in the param files
- plot capture point information with RVIZ
- add option to integrate floating base velocities in the state estimator
- send input device commands only if the robot is in the active state
- push recovery based on capture point
- integrated other services in the keyboard node
- angular velocities bug
- arm startup problem
- base_footprint REP-120
- add controller test with EIGEN MALLOC checks
- step reflex
- separate sliders for the velocities
- reorganized foot trajectory files
- perform init() only one time
- new ROS services (reset_base, set_swing_frequency, etc...)
- new ROS services to the keyboard node
- working on issue #2
- clean the ros topics
- reduce xbot logger verbosity
- cleaned up various branches
- bug fix in the foothold reset about the terrain estimation
- submodules
- first implementation of the push recovery
- support for ubuntu 20.04
- support for ubuntu 18.04
- control stack
- init
- arm support