CrowdNav is a simulation based on SUMO and TraCI that implements a custom router that can be configured using kafka messages or local JSON config on the fly while the simulation is running. Also runtime data is send to a kafka queue to allow stream processing and logger locally to CSV.

• Download the CrowdNav code
• Run python setup.py install to download all dependencies
• Install SUMO & set env var SUMO_HOME
  1. Download SUMO from http://prdownloads.sourceforge.net/sumo/sumo-src-0.32.0.zip?download
  2. Extract and note the full path to the directory
  3. Set SUMO_HOME to point to this directory
• Install Kafka (we recommend this Docker image) and set kafkaHost in Config.py
• Run python run.py

A first guide on how to use (i.e. adapt, measure, optimize) CrowdNav with the RTX tool is available at this Wiki page.

• Normal mode (python run.py) with UI to Debug the application. Runs forever.
• Parallel mode (python parallel.py n) to let n processes of SUMO spawn for faster data generation. Stops after 10k ticks and reports values.

• Runtime variables are in the knobs.json file and will only be used if kafkaUpdates = True is set to false in Config.py. Else the tool uses Kafka for value changes.
• To disable the UI in normal mode, change the sumoUseGUI = True value in Config.py to false.

• To let the system stabalize, no message is sent to kafka or CSV in the first 1000 ticks .