如何让SUMO车辆持续留在基于OSM路段的NS2-VANET仿真场景中？

Hey there! Let’s work through this problem where your SUMO vehicles are bailing on your VANET simulation too early—especially when using an OSM road network. I’ve run into this exact scenario before, so here are practical, actionable ways to keep your vehicles looping around or re-entering the scene for the full simulation duration:

The simplest fix is making your vehicles repeat their paths instead of exiting once they reach the end of a route. Here’s how:

• Modify Route Files with Repeat Logic:
  If you used duarouter to generate your .route.xml file, edit the <route> tags to add a repeat attribute. Set it to -1 for infinite loops, or a specific number if you want a set number of repeats. For example:

  <route id="looping_route" edges="osm_edge_1 osm_edge_2 osm_edge_3 osm_edge_1" repeat="-1"/>
  

  Just make sure the route forms a valid loop (the last edge connects back to the first, or the path is continuous enough for SUMO to handle the repeat).
• Use SUMO’s Command-Line Flag for Automatic Repeats:
  Skip manual XML edits by adding this flag when starting SUMO:

  sumo -c your_simulation_config.sumocfg --end-of-route-repeat
  

  This tells SUMO to automatically restart a vehicle’s route as soon as it finishes, so it never leaves the network.

If you’re spawning all vehicles at once, they’ll eventually all exit. Instead, use continuous flows to keep new vehicles entering the scene:

• Replace Individual Vehicles with <flow> Tags:
  In your .rou.xml file, swap out <vehicle> entries with <flow> to spawn vehicles over time. For example:

  <flow id="continuous_flow" begin="0" end="3600" number="150" route="looping_route" departSpeed="max"/>
  

  Here, begin and end match your total simulation time (3600 seconds in this case), and number sets the total number of vehicles to spawn. Combine this with route looping, and you’ll have a steady stream of vehicles that never leave.

OSM maps often have open edges at the boundary, which let vehicles drive off the map. Close those gaps with these tricks:

• Clip the OSM Network to a Closed Area:
  Use netconvert to crop your OSM map into a smaller, bounded region where vehicles can’t escape. Specify a rectangular boundary with coordinates:

  netconvert --osm-file your_osm_map.osm --boundary 40.712,-74.006,40.715,-74.003 --output-file clipped_network.net.xml
  

  This creates a self-contained network where it’s easier to design looping routes.
• Add "Ghost" Edges for Redirection:
  Use SUMO’s NetEdit tool to manually add short "ghost" edges at the network’s open boundaries. Connect these edges back to entry points in the map, then update your routes to use these edges—so when a vehicle reaches the boundary, it’s redirected back into the scene.

Since you’re pairing SUMO with NS2, make sure the two tools are aligned to preserve vehicle behavior:

• Match Simulation Time and Step Lengths:
  Ensure SUMO and NS2 use the same step length (e.g., 0.1 seconds) and total simulation duration. This prevents desync where SUMO thinks a vehicle is still looping but NS2 registers it as exited.
• Use TraCI to Control Vehicles Dynamically:
  If you’re using TraCI to link SUMO and NS2, write a small script to detect when vehicles approach the network boundary and re-route them. Here’s a quick Python example snippet:

  import traci
  
  traci.start(["sumo", "-c", "your_config.sumocfg"])
  while traci.simulation.getMinExpectedNumber() > 0:
      traci.simulationStep()
      # Check each vehicle's position
      for veh_id in traci.vehicle.getIDList():
          current_edge = traci.vehicle.getRoadID(veh_id)
          # List of boundary edges to watch for
          if current_edge in ["osm_boundary_edge_1", "osm_boundary_edge_2"]:
              # Re-route the vehicle back to a starting path
              traci.vehicle.setRoute(veh_id, ["entry_edge_1", "entry_edge_2", "core_edges..."])
  traci.close()
  

  This script actively monitors vehicles and redirects them before they can exit the scene.

内容的提问来源于stack exchange，提问作者Houston61