Finite State Machine library #418
Conversation
|
@julianoes I forgot to tag you earlier, please take a look around (particularly the |
make sure the vehicle raises to the loiter altitude before taking a decision
|
Let me know if/when you need more testing capacity and we can talk about moving from Travis-ci to Jenkins (http://ci.px4.io:8080/blue/organizations/jenkins/PX4%2FFirmware/detail/master/1082/pipeline). At times it can be more annoying than travis, but already has drastically more resources (50+ slaves) and can easily dynamically provision anything else from ec2 as needed. |
| if(TARGET ${PROJECT_NAME}-test-roscore) | ||
| target_link_libraries(${PROJECT_NAME}-test-roscore ${PROJECT_NAME} | ||
| if(TARGET ${PROJECT_NAME}-test-roscore) | ||
| target_link_libraries(${PROJECT_NAME}-test-roscore ${PROJECT_NAME} | ||
| ${catkin_LIBRARIES} | ||
| ${YAML_CPP_LIBRARIES}) | ||
| endif() |
There was a problem hiding this comment.
Did these whitespace changes get in through autoformat? My clang-format doesn't check cmake files, which is why I'm curious. In terms of keeping this huge PR somewhat reviewable, consider getting this fluff out
There was a problem hiding this comment.
It was just annoying me when I was looking into adding roscore-style tests to the safe_landing_planner. They didn't end up being added, but I felt that the reformatting I did would still be useful.
safe_landing_planner/include/safe_landing_planner/waypoint_generator.hpp
Outdated
Show resolved
Hide resolved
|
|
||
| protected: | ||
| // config | ||
| float yaw_setpoint_ = NAN; |
There was a problem hiding this comment.
Since I just spend an enormous amount of time debugging different conventions for yaw, can we specify (at least in comments here, preferably in the variable name) the frame and the units? (I'm guessing here it's fcu frame and degrees, so maybe yaw_setpoint_fcu_deg_?
There was a problem hiding this comment.
It's in the local frame in radians, it seems we don't have the same complications here that we would in the local_planner because we don't have a histogram.
There was a problem hiding this comment.
sure, but units might make sense still
| Eigen::Vector3f velocity_setpoint_ = Eigen::Vector3f(NAN, NAN, NAN); | ||
| Eigen::Vector3f loiter_position_ = Eigen::Vector3f(NAN, NAN, NAN); | ||
| Eigen::Vector3f exploration_anchor_ = Eigen::Vector3f(NAN, NAN, NAN); | ||
| Eigen::Vector3f vel_sp = Eigen::Vector3f(NAN, NAN, NAN); |
There was a problem hiding this comment.
There's a vel_sp and a velocity_setpoint_... Smells like debug code that wasn't properly cleaned up
safe_landing_planner/include/safe_landing_planner/waypoint_generator.hpp
Show resolved
Hide resolved
safe_landing_planner/include/safe_landing_planner/waypoint_generator.hpp
Outdated
Show resolved
Hide resolved
This PR introduces an example implementation of a Finite State Machine library. It has a few desirable features:
.dotfile for each FSM with the help of a simple Python script.This is generated from the FSM library's own unit tests:
high res version
The second part of the PR was porting the
waypoint_generator_nodein thesafe_landing_plannerto use this, as a real-world example.high res version
This still needs further SITL testing to make sure nothing has broken, and then additional unit tests for the
safe_landing_plannerwaypoint_generator_node, which wasn't tested until now.FYI @dagar @Jaeyoung-Lim
TODO: