Add the following input parameters to your model if you don't already have them
closest_waypoints = params['closest_waypoints']
track_width = params['track_width']
steering_angle = params['steering_angle']
steps = params['steps']
waypoints = params['waypoints']
Then add the following at the end of your model, it will add the extra logging required to generate the circuit layout on the heatmap
import math
coord0 = waypoints[closest_waypoints[0]]
coord1 = waypoints[closest_waypoints[1]]
myradians = math.atan2(coord1[1]-coord0[1], coord1[0]-coord0[0])
mydegrees = math.degrees(myradians)
print("Waypoint0:{},X:{},Y:{},heading:{},trackwidth:{},steeringangle:{},steps:{}".format(closest_waypoints[0],coord0[0],coord0[1],mydegrees,track_width,steering_angle,steps))
An example model is here
Install matplotlib
pip3 install matplotlib
After training your model with the additional logging added find the name of the log stream that was created for this model, it should be the latest logstream in CloudWatch Logs /aws/robomaker/SimulationJobs e.g. "sim-nd2x4c3ph1d3/2019-06-05T14-39-30.940Z_3f5ddea9-6535-45d0-b1c7-ceb4b054f884/SimulationApplicationLogs"
Now generate a heatmap by running the script
python3 Get-DeepRacerHeatMap.py --profile <yourprofile> --logstreamname sim-nd2x4c3ph1d3/2019-06-05T14-39-30.940Z_3f5ddea9-6535-45d0-b1c7-ceb4b054f884/SimulationApplicationLogs