Autonomous Vehicle Simulation

Provide a comprehensive understanding of foundational principles and methodologies in robotics. The course emphasizes hands-on experience, interdisciplinary thinking, and problem-solving skills, covering topics such as algorithms, sensing and actuation, and research methodologies.

Objective

A final challenge composed of two parts. The first being a race where teams compete to see whose car simulation could make it around the course fastest with the least number of collisions using pure pursuit path following. The second being an obstacle course with the goal to complete the course as fast as possible with little to no collisions using the lidar sensor data.

Outcome

Programming in Python
Understanding ROS Concepts and Framework
Robotics Simulation with tools like Gazebo
Robot Kinematics and Dynamics
Sensor Integration (lidar, cameras, IMUs)
Robot Perception (object recognition, localization)
Path Planning and Navigation
Simultaneous Localization and Mapping
Control Systems Design
Machine Learning for Robotics
Simulated Hardware Interface

Hard Skills

Collaboration
Communication Skills
Problem-Solving
Critical Thinking
Project Management
Adaptability
Documentation
Continuous Learning
Attention to Detail
Ethical Considerations

Soft Skills

The Process

MIT 6.141: Robotics, Science and Systems

Spring 2021

I worked on a team of four to complete six labs developing an autonomous system for a robotics simulation. We used python to code a variety of algorithms in ROS. Each of the labs demonstrated something unique about autonomous systems which students could then utilize in the final race where each team needed to create a path following and obstacle avoidance system.

Final Challenge

GitHub

Localization

Monte Carlo localization was used to determine where our robot was in the simulated space.

Path Planning

An occupancy grid and A* search algorithm were used to plan a path. A pure pursuit controller was used to follow the path.