Skip to content

Latest commit

 

History

History
108 lines (98 loc) · 4.75 KB

History.md

File metadata and controls

108 lines (98 loc) · 4.75 KB

0.1.1 / 2016-06-04

  • Fixed point cloud sensor data problem with missing odometry topic
  • Added sensors module for simulation point clouds
  • Added (some) parameters to dynamic reconfigure
  • Consolidated relavant dependencies into one place to allow single clone development
  • Many improvements on the interface
  • Updated messages to match status of the SPENCER project
  • Added simple mock scripts for creating repeatable tracks and groups
  • Cleanup many of the dependencies to have a lean setup

0.0.4 / 2015-03-30

  • Updated launch files and config to match rviz plugins
  • cleanup of dependencies
  • adding compatibility modules with indigo
  • fix small bug with missing signal header

0.0.3b / 2014-11-19

  • FIXED slow termination of the main node due to SIGINT and SIGDFL conflicts, thanks to @tlind
  • Fixes #5 removed simulator dependency on spencer messages, now using local pedsim_msgs copies
    • AgentState.msg modified to follow Twist and Pose standards - robot teleop working with keyboard and xbox joystick - rviz file modified for visualization purposes
  • added more message types for self contained compilation
  • added animated markers to rviz config files

0.0.3 / 2014-08-22

  • changes to standing and queueing scenes to generate more realistic data
  • better standing people modeling by tweaking different force weights
  • lots of cleanup, optimizations, speedups...
  • changed social state to a more consistent string notation
  • integrated spencer_social_relation_msgs for publishing simulated social activities
  • changed queue waiting time distribution to erlang
  • fixed spooky behavior in shopping modes
  • added walking people to the simulator and tuned behaviours, thanks to @tlind
  • fixed queue dancing behaviour issues (re-wrote personal space in queue code)
  • added spencer msgs to the simulator, not using standard spencer topics
  • now having people models and more in the simulator

0.0.2 / 2014-05-12

  • added social status to agent status messages
  • minor edits when recording
  • more scenario files
  • new scenerios for collecting training data
  • minor changes in param when making the videos and cleanup
  • removing unnecessary files
  • new scene and launch file for social context experiments
  • polishing config to handle magic numbers
  • more refactoring and cleanup
  • more cleanup, minor fix in documentation
  • adding basic auto documentation
  • added publishing og waypoint markers
  • more cleanup and fixes
  • removing old files not needed anymore
  • continous update of center of mass in groups
  • minor cleanup, removing unnecessary slots
  • fixed standing/slow agents behaviour to simulate elders
  • changed queue service wait time distribution to exponential
  • fixed z visuals size error
  • making simple demo, minor fix on obstacle cells locations
  • fixed planning bug that did not factor obstacles
  • minor fix in slots and group params
  • new behaviour demonstration scene
  • package status update, more cleanup
  • and we have nice teleoperation back!
  • daryl back in visualization, reworking teleop now with waypoint planners
  • add enable/disable switch for groups for easy behaviour switches
  • addign more debug statements
  • more cleanup, removing qt pieces
  • minor fixes and refactoring
  • fixed attraction visualization bug
  • minor fix in attractions
  • adding visuals for attractions
  • working queues, groups and attraction behaviour
  • restored robot command, combined agent status and visual info publishing
  • more cleanup, eliminating qt pieces, more files
  • more cleanup, eliminating qt pieces,
  • finally back to working groups and potentially queues etc
  • more files
  • adding sven files
  • backing up before adding components
  • more tuning, using neighborhood for group additions
  • tweaking parameters for group forces before branching the code to prep for intergration
  • now visualizing groups, force parameters need tuning
  • assigning people to groups...
  • repulsion and coherence
  • connecting forces between pedestrian groups and agents
  • pulling in groups: work in progress... more files
  • pulling in groups: work in progress...
  • moving the area to select agents to join queue to be length dependent
  • cleaning up states
  • adding infrastructure for group forces while making states more clear, additional standing behaviors
  • work in progress: rewritting queue and situation logic
  • congregating around attractions/standing... work in progress
  • adding more variables to config
  • allow explicit re-joining of queues (after a small window that is configurable)
  • reworking waiting points logic
  • randomized service times, queue buffers, queue directions and added flags to directly manipulate waypoints
  • cleaning up the experiment scenes