Dear friends and users,
we are happy to announce the release of SUMO version 1.28.0.
The download links are at https://sumo.dlr.de/Download.
If you cite the new release in your publications, please use the DOI
10.5281/zenodo.23219601 (https://doi.org/10.5281/zenodo.23219601)
This release brings many enhancements and fixes to mesoscopic simulations including an improved jamming model (--meso-ltm).
It also extends and adds tools that help with understanding and importing traffic data for traffic generation.
There are many more features and fixes in this release and the most important ones are listed below.
For a full list of changes, as always see https://sumo.dlr.de/Changelog.
## Enhancements
- sumo
- Added carFollowModel NaSch (Nagel Schreckenberg)
- Added option --queue-output.aggregation <TIME> for writing per-edge queue length statistics (max, median, configurable percentile in vehicles and meters)
- meandata values for net, emissions and noise can now be merged into a single line or table row
- Added laneChangeModel parameters lcCooperativeMinSpeed and lcCooperativeHelpThreshold which make cooperative yielding context dependent
- Rerouters now support viaProbReroute for rerouting vehicles but keeping it's destination. This can be combined with closingReroute for fine-grained control over detours
- Stop attribute 'duration' now accepts a distribution string
- netedit
- vTypeDistributions now show up in type-combobox in vehicle mode
- Connection mode now allows to hide connections not originating from the current source edge to reduce clutter
- Connection mode now allows to inspect any connection via clicking
- sumo-gui
- Added new bottom bar icons/labels that track important scenario-health metrics: insertion-delayed vehicle count and percentage of teleports per inserted vehicles
- in meso, the edge context menu now permits to inspect lane attributes (i.e. to check lane permissions)
- The coordinate display in the bottom bar now supports pasting lat,lon coordinates. The view dialog also supports lon and lat entry
- duarouter
- It is now possible to repair routes with stops after removing parts of a network (including busStops) with options --repair --ignore-errors
- meso
- Added option --meso-ltm which activates an improved jam-dynamics model
- Refactored data structure to reduce running time by 5-9%
- Now supports traffic light type 'actuated'
- Added inductionLoop attribute mesoTLS. Detectors with this attribute can be used as custom detectors for actuated traffic lights
- Option --queue-output is now supported
- traci
- Added function vehicle.getShadowLaneID (the other lane if a vehicle occupies two parallel lanes)
- Added functions for getting aggregated traffic data from traci.meandata
- tools
- randomTrips.py: Added option --return-to-origin which can be used with --intermediate to model return trips.
- gtfs2pt.py now interprets extended gtfs route_types. The default interpretation of mode 109 was changed to 'train' as this is by far the more common case. Special cases such as Berlin and Hamburg where 109 is mapped as 'light_rail' in OSM can be handled by setting the new option --sbahn-is-light-rail
- gtfs2pt.py: Added option --write-arrival to include scheduled arrival times in the output
- instantOutToEdgeData.py: added option --poi-output to visualize extra values and aid in debugging
- edgeDataFromFlow.py: added option --turn-output to write edgeRelations for turning-lane detectors
- edgeDataFromFlow.py: added option --skip-incomplete to exclude data from incomplete cross-sections
- flow2POI.poi: added new tool for visualizing detector flow
- flowConservation.py: new tool to sanity check edgeData around junctions
## Bugfixes
- sumo
- Defining a closingReroute with attribute 'until' now interprets the time correctly again (regression in 1.26.0)
- Fixed several bugs related to rerouters that change network permissoins
- Fixed crash/misdetection when trying to detect persons on E3 detector
- Fixed deadlocks and unsafe driving in sublane simulation
- Fixed bug where vehicles were kept in the simulation despite insertion failure for longer than option --max-depart-delay
- Time spent with scheduled stops no longer lowers average trip speed returned in statistic-output
- sumo-gui
- Configuring language settings is working again (regression in 1.27.1)
- meso
- Fixed exaggerated headways when passing short edges (regression in 1.10.0)
- Fixed invalid jam dynamics when running with internal links
- As a consequence, meso defaults to running with internal links if the net.xml contains them networthey are. This improves accuracy at the cost of some simulation speed
- Fixed invalid jam dynamics at short edges
- Fixed right-of-way violation that was causing gridlock in roundabout and exaggerated junction throughput
- netedit
- Fixed invalid error when loading additional object defined at lane end (regression in 1.20.0)
- Fixed invalid connection shapes after computing the network for the first time (regression in 1.22.0)
- vTypeDistributions created by netedit are now loadable by sumo (fixed by giving new capabilities to sumo, regressin in 1.25.0)
- The menu function for saving .sumocfg with save-as is working again (regression in 1.26.0)
- The button for saving the .sumocfg is now functional again after loading a network (regression in 1.26.0)
- Junction context menu operation 'clearing connections' no longer generates an invalid network if the junction had tls-controlled pedestrian crossings (regression in 1.27.0)
- setting "random" as color value for vehicles is now working
- The net "knows" that it needs to be recomputed after changing netedit processing option
- netconvert
- Fixed invalid link index for second part of multi-lane internal edge (regression in 1.22.0)
- Fixed missing connections in networks with dedicated bus lanes (regression in 1.23.0)
- OSM: no longer importing superfluous tls where tram runs on road
- When option --sidewalks.guess is set, high-speed edges without sidewalks are no longer modelled as shared space
- Fixed several issues that caused with options --plain-output-prefix and --output.format csv
- tools
- gtfs2pt.py: fixed several bugs that resulted in invalid outputs
- gtfs2pt.py: routes with the same sequence of stops but different timing are now distinguished
Have fun with the new release,
Pablo, Angelo, Micha, Robert, Jakob and Yun-Pang