Skip to content

StevenAskwith/deepracer-tools

Repository files navigation

Deep Racer Heatmap

Add logging to model

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

Python Modules

Install matplotlib

pip3 install matplotlib

Usage

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

About

Various tools for working with AWS Deep Racer

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published