Train a Smartcab to Drive

As part of my Udacity Machine Learning Nanodegree, I applied reinforcement learning to build a simulated navigation agent. This project involved modelling a complex control problem in terms of limited available inputs, and designing a scheme to automatically learn an optimal driving strategy based on rewards and penalties.

See my implementation and report here.

Project Brief

In the not-so-distant future, taxicab companies across the United States no longer employ human drivers to operate their fleet of vehicles. Instead, the taxicabs are operated by self-driving agents — known as smartcabs — to transport people from one location to another within the cities those companies operate. In major metropolitan areas, such as Chicago, New York City, and San Francisco, an increasing number of people have come to rely on smartcabs to get to where they need to go as safely and efficiently as possible. Although smartcabs have become the transport of choice, concerns have arose that a self-driving agent might not be as safe or efficient as human drivers, particularly when considering city traffic lights and other vehicles. To alleviate these concerns, your task as an employee for a national taxicab company is to use reinforcement learning techniques to construct a demonstration of a smartcab operating in real-time to prove that both safety and efficiency can be achieved.

Project Evaluation

My project was evaluated against the Train a SmartCab to Drive project rubric.

Files Submitted

• The agent.py Python file with all code implemented as required in the instructed tasks.
• The /logs/ folder which contains five log files that were produced from simulation and used in the analysis.
• The smartcab.ipynb notebook file with all questions answered and all visualization cells executed and displaying results.
• An HTML export of the project notebook with the name report.html.