The following are highlights for the week of August 10th, as reported by Dronecode project leads. (Reproduced from the Technical Steering Committee mailing list.)
The T3 event is drawing to a close. http://diydrones.com/profiles/blogs/t3-back-to-the-future?xg_source=acti…
The reason I mention it is we had a few rover entries with interesting results. Most rover configurations are for slow to medium speeds but the T3 is all about going as fast as possible so the setup and running of those rovers is very interesting and has quite a lot of useful information for other Rover users wishing to do this sort of thing.
Rover has had a few changes recently so we are heading towards a release. To this end we have released a BETA v2.2.51 for those Rover users wishing to help out with testing.
A T3 entrant (Mike aka dr0n3droid) uncovered an issue in Rover which turns out had been identified over a year ago by Mike Roberts and logged as Issue #840 . If your Rover is in AUTO mode and you nudge the throttle up past 50% the Rover will increase in speed. This is a way to run auto missions and test out what speeds it’s best suited for. HOWEVER, it turned out when you put the throttle in reverse over 50% exactly the same thing happened! So users in T3 had a Rover go very fast out of control and they instinctively put it into full reverse which made the situation much worse! This has been resolved and now going into reverse won’t change the speed unless the Rover is driving in reverse.
We also did a search through the outstanding Rover issues in Github to see if there were any other previously logged issues that we can resolve. No bugs were found but a couple of enhancement requests were identified so they have been assigned to the Rover maintainer for review.
Through the week a clever user named Marc Merlin was having a lot of trouble configuring Flaperons for his Plane. He eventually figured it out and then took the fantastic step of writing up a document on how to do it! This has subsequently been reviewed, edited and tested (yes we test the documentation) and incorporated into the main Plane wiki – thanks Marc!
Lucas De Marchi has been doing great work on the Inertial Sensors library going through a code review and tidy up.
Tridge has updated the PX4Firmware and PX4NuttX submodules within Ardupilot which is always a time consuming and complex task but does allow Ardupilot to have the latest set of bug fixes and enhancements coming from the PX4 work.
Copter-3.3 beta testing continues. Major blocking issue is improving the EKF’s resistance to high vibration and dealing with Barometer failures. Paul Riseborough has provided a 2nd potential solution which relies on the inflight vibration monitor. We plan to test this solution this week.
Various small fixes including:
- MOT_THR_MIX_MAX parameter added to copter for AC3.3 to allow controlling the attitude vs throttle prioritisation during dynamic flight
- optical flow driver enabled for Linux boards (Victor Mayoral from Erle Robotics)
- Fixed unwanted descent during Spline waypoints if speed too high (Leonard Hall)
- RALLY_INCL_HOME parameter added to allow users to more easily include Home as a safe landing spot when using Rally points (defaults to include home for copter, not included for Plane)
- PR from fhedberg which creates abstraction layers in the AP_MotorsHeli class, similar to the way multirotors are done. This new architecture will enable creation of new helicopter frame classes, such as tandem and compound helicopters.
- Processing this PR requires completely re-write of heli improvements from last week due to heavy merge conflicts.
- Tridge has create a tachometer sensor class, currently being used to gather data from his gas helicopter engine. Needs more work for filtering, second input for rotor speed, etc.
- Tachometer data will be used for closed loop throttle control.
- Tridge’s helicopter which has very high angular vibrations had revealed a problem with aliasing of the IMU rate data.
PX4 Flight Core
- Fixed wing
- Fixes for TECS initialization in-air, prevents climbs / dives on switch (#2588)
- Generics failsafe class evaluating signal quality, timeouts and vibration (#2441)
- Bootloaders for Zubax GNSS and Ben’s ESC have been prepared, which basically concludes preparation for the release, which is planned for Aug 15 (jointly with PX4)
- Recent developments have improved stability, including stability on very low RPM: https://www.youtube.com/watch?v=oGGGSI3pNk4
- Mission planner 1.3.31 was released in the past week which included updates to georeferencing, correct log file creation from rtk gps’s, and updates to support NTRIP data sources for RTK.
- Release v2.7.1 (download) out
- Support for Firmware update on PX4, APM and SiK radios
- Supported airframes are now loaded dynamically from each .px4 file and the user menu of airframes is populated based on the binary
- Improved Flight Mode config with more intuitive UI
- Firmware upgrade fixes for older boards
- Live H.264 video streaming support
This update is a little longer than one week.
We’re working toward building a standard abstract interface for commanding aerial vehicles based on ROS messages. The goal is to define some high level modularity points to allow better interoperability. It’s an ongoing discussion a few links can be found.
- The ongoing discussion in ROS-SIG-MAV: https://groups.google.com/forum/#!topic/ros-sig-mav/f12m3mnqYl4
- In testing new message interface: https://github.com/ethz-asl/mav_comm/pull/30
- A standard set of nomenclature for level of autonomy: https://github.com/ethz-asl/mav_comm/pull/25
Related discussions are for a standard way to define some related concepts for aerial vehicles.
- Coordinate frames for outdoor vehicles: https://github.com/ros-infrastructure/rep/pull/103
- Standard battery messages: https://github.com/ros/common_msgs/pull/74
These are ongoing conversations. Please take some time to review them and add comments. Although this is being developed using ROS nomenclature our goal is to define an abstract interface which can be useable with our without ROS.