CS297 Proposal
Crowd-sourcing Data for Autonomous Driving and Applying It
Yaoyan Xi (Yaoyan.Xi@sjsu.edu)
Advisor: Dr. Chris Pollett
Description:
Autonomous driving is an artificial intelligence guided transportation mode that is capable sensing the environment and moving with little or no human input. It is a fledgling realm that will significantly transform the way people interact with the vehicles, vehicles interact with each other, and how the roads are built 'smarter'. It will enable more productivities from human drivers, decrease the traffic accidents, and potentially redistribute traffic volume to tackle the traffic congestion.
This project strives to build a mobile phone application that collects real-time road scenery and steering wheel movement data, both direction and steering wheel arc degrees, via sensors. It is a fact that every road has already been driven on by some one at some point and will likely be driven on by some one again in the future. It is also a fact that mobile phones are ubiquitous. So it seems promising to collect car driving data using a mobile application and then use that data to train autonomous vehicle AI systems. We will in this project develop applications to actually do this.
Our application will collect and integrate data such as lane, travel time/traffic volume, season/weather, speed limit, and GPS coordinates into the machine learning model to decide how to guide the steering wheel of autonomous vehicle.
We will use the Convolutional Neural Network (CNN) model and the training will be done through Python sklearn library.
Schedule:
Week 1:
Jan. 27 - Feb. 2 | Kickoff meeting and proposal draft |
Week 2:
Feb. 3 - Feb. 9 | Proposal refinement and finalization |
Week 3:
Feb. 10 - Feb. 16 | Read Menze paper |
Week 4:
Feb. 17 - Feb. 23 | Start to build mobile application |
Week 5:
Feb. 24 - Mar. 2 | Continue to build mobile application |
Week 6:
Mar. 3 - Mar. 9 | Finish building App [Deliverable 1] |
Week 7:
Mar. 10 - Mar. 16 | Get the simulated dataset and preprocessing [Deliverable 2] |
Week 8:
Mar. 17 - Mar. 23 | Read Pomerleau paper |
Week 9:
Mar. 24 - Mar. 30 | Start to build CNN model |
Week 10:
Mar. 31 - Apr. 6 | Continue with CNN model [Deliverable 3] |
Week 11:
Apr. 7 - Apr. 13 | Start to validate CNN model built |
Week 12:
Apr. 14 - Apr. 20 | Finish validating CNN model built |
Week 13:
Apr. 21 - Apr. 27 | Start to integrate Google Map API and routines to control vehicles |
Week 14:
Apr. 28 - May 4 | Finish integrating Google Map API and routines to control vehicles [Deliverable 4] |
Week 15:
May 5 - May 11 | Start to write proposal report |
Week 16:
May 12 - May 18 | Finish writing proposal report [Deliverable 5] |
Deliverables:
The full project will be done when CS298 is completed. The following will
be done by the end of CS297:
1. Understand state-of-the-art knowledge in autonomous/driverless vehicle and learn to build mobile application to sense vehicle's location
2. Secure the dataset suitable for this project from open sources and preprocess data via feature engineering
3. Build and train the CNN model
4. Validate the CNN model and integrate Google Map API and all routines for overall controls
5. CS297 proposal report writeup
References:
[Long2007] A Review of Intelligent Systems Software for Autonomous Vehicles. Long,L., Hanford,S., Janrathitikarn, O., Sinsley, G., Miller, J. Proceedings of the 2007 IEEE Symposium on Computational Intelligence in Security and Defense Applications. 2007.
[Menze2015] Object Scene Flow for Autonomous Vehicles. Menze,M.,Geiger,A. Computer Vision Foundation. 2015.
[Pomerleau1991] Efficient Training of Artificial Neural Networks for Autonomous. Navigation. Pomerleau, D.A. In Neural Computation 3:1 pp. 88-97. 1991. |