Cozmo Robotics Project: Race Track Pattern Recognition
In this project, we explore how to make Cozmo follow any race track pattern and respond to external signals like traffic lights. The approach includes line following and light detection using advanced techniques. Results show promising progress with room for improvement in line following accuracy and light detection reliability. Future work aims to enhance track customization and decision-making capabilities for multiple Cozmo robots concurrently.
Download Presentation
Please find below an Image/Link to download the presentation.
The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author.If you encounter any issues during the download, it is possible that the publisher has removed the file from their server.
You are allowed to download the files provided on this website for personal or commercial use, subject to the condition that they are used lawfully. All files are the property of their respective owners.
The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author.
E N D
Presentation Transcript
Traffic Cozmo Xinze Deng Bolaji Bankole Cognitive Robotics S2019
How do we make Cozmo follow any race track pattern and recognize when to start, stop, slow down, based on external signals such as traffic lights?
Approach: Line Following 4 sets of training data 800+ images total 4 output nodes from 3 layer neural network Forward, left, right, lost Recovery from lost position Back up until familiar image comes to sight CrossEntropyLoss() function
Approach: Light Detection openCV: filter for red clusters of pixels Based on threshold value Look at top 50% of image Cozmo at set head angle
Results & Observations Line following works 70% of the time Gets distracted by other objects Trained on many different tracks, adjust model to avoid jack of all trades, master of none situation Light Detection Lagging behind due to wireless Need more image filtering Can use ML training for Light detection as well
Future Work Custom tracks with different bend radius More than one cozmo on the course at a time Faster image processing Decision making based on previous actions
