What’s new in MORSE 1.4?


  • Numpy is now needed for Morse. It is used in several places where computations using mathutils is not precise enough (float vs double precision).
  • Time management has been improved in various way. A morse_sync tool has been introduced to improve precision and timing of high-frequency components (#683). If available in Blender (Blender > 2.77), it is also possible to accelerate or slow-down the simulation (#388). Moreover, Morse now try to compute automatically the right time settings. If you meet any problem related to time, make sure to read the Time and event in MORSE and / or report issue to the Morse project.



  • the Human avatar in MORSE has been entirely rewritten. The new human model is much simpler, yet much nicer (in particular, it features mesh skinning for good looking animations). On the downside, the interactive mode is gone for now. Depending on interest, it can be revived in a future version (possibly through external scripts, for added flexibility).


  • the semantic of the Waypoint and Destination actuators has slightly changed: once the destination is reached, they do not attempt anymore to actively stay at this position. This permits another motion actuator to ‘take over’ the control of the robot. The previous behaviour is still desirable in certain cases (notably for flying robots), and can be re-enabled by setting the property RemainAtDestination to true: motion.properties(RemainAtDestination=True). This option is also added to the Rotorcraft Waypoint motion controller actuator, but it defaults to true (hence, no behaviour change compared to MORSE 1.3).
  • the Orientation Actuator actuator has been enhanced to possibly work more realistically, by limiting the speed of the rotations. The default is still to go directly to the desired orientation.
  • The Keyboard Actuator and Joystick Actuator actuators do not call anymore the robot apply_speed method with values set to zero when no input is received. The previous behaviour prevented them to be used in combination with another motion actuator (they would always overwrite other motion commands with zeros).
  • The Armature Actuator actuator has two new services (rotate_joints and translate_joints) that let the user set the rotations/translations of only a subset of the armature’s joints by providing a custom mapping {joint name: value}.
  • The Rotorcraft attitude motion controller has been extended to be able to control the rotorcraft in yaw rate or in absolute yaw (using the YawRateControl property). If it is the case, you can configure if you want to configure to use normal yaw or north using the property UseAngleAgainstNorth. Last, you can configure the actuator to use a linear or quadratic thrust model using LinearThrust.
  • Introduce the Drag “actuator” which allows to simulate the drag (air resistance) force opposite to the move of the robot. It allows more realistic simulation (if desired).
  • Introduce the External Force/Torque actuator which allows to apply external force (typically force from the environment such as wind) to a robot. It has the same interface than Force/Torque Motion Controller, but apply force in the global frame.
  • Introduce the Quadrotor dynamic controller actuator which allows to control a quadrotor from the speed of its four engine, using simple dynamic equation.


  • longitude, latitude and altitude are not anymore properties of GPS but must be set at the environment level. Moreover, the property angle_against_north allows to configure the angle between the X-axis and the geographic north (must be positive when the Blender X-axis is East of true North, negative if is West of true North).
  • Introduce the new high-level sensor Attitude sensor, allowing to compute the attitude of the system
  • Introduce the sensor Magnetometer, which allows to compute the magnetic field vector of the Earth.
  • Extend the sensor Inertial measurement unit, to return also the magnetic field vector.
  • Fixed the Collision sensor: it now detects collision only when it is actually colliding (before, any object in a 1x1x1m box around the sensor would return a collision). While here, improve the documentation with a complete example.
  • Introduce the sensor Airspeed, which allows to compute the speed of a vehicle relative to the air.


  • Introduce ECEF, Geodetic, Geocentric modifiers, allowing to convert coordinates from Blender world to ECEF-r or Geodetic or Geocentric coordinates (and vice-versa). It should improve interoperability with flight systems in general.
  • Introduce Feet modifier, to convert imperial units to meter buts (and vice-versa)



  • Introduce a binding for the Mavlink protocol, easing the interoperability of Morse with a lot of free autopilots / architectures.


  • Some ROS ‘housekeeping’ has been performed in this release, including removing the need for rospkg (easier installation!), removing ROS interface with the non-standard (and unused?) JointPositions message and removing references roslib.load_manifest(), a memory of rosbuild-era.


  • Handle automatically the case where attributed published by Morse are not owned by it.
  • Allow to specify a stop_time for the simulation (in simulated seconds)
  • Make lookahead configurable for the Morse federate


Builder API

API addition

  • It is now possible to import environment composed of multiples scenes. The user should select which is the main_scene when importing the environment. Moreover, a method Environment.set_background_scene has been added to configure the scene to use in background (#651).
  • The method bpymorse.set_speed, used to changed the frequency of Morse main loop is now deprecated in favor of Environment.simulator_frequency.
  • The method Environmement.set_time_scale allows to accelerate or slow-down the simulation (#388).
  • The new method Environment.use_vsync allows to control the vsync parameter


  • Robots created in loop are handled smartly. They are still usable as previously, but it is also possible to access them using the list foos (if your robot name is foo) (#358).
  • Streams are now created lazily, fixing control with large number of robots / sensors (#626).

Table Of Contents

Previous topic

What is MORSE?

Next topic

What’s new in MORSE 1.3?

This Page