Cornell Electric Vehicles is a Project team at Cornell University focused on building the world's most efficient electric vehicle
We compete in the Shell-Eco Marathon in the Prototype (3-wheeled vehicle) and Urban Challenge (4-wheeled small car) divisions
We are making our vehicles autonomous to improve performance and compete in future autonomous challenges
Currently Motion Planning Lead and work on motion planning, navigation, control, mapping, simulation, and testing with three other members
Started as Motion Planning Software Developer initially and worked on a software pipeline system to validate, filter, and process 100+ vehicle paths to find 4-5 viable paths to a goal point
Competed in Shell Eco-Marathon Autonomous Programming Beta Competition and earned 2nd place
Here is a video from when we were developing the motion planning system for the second Shell Eco-Marathon Autonomous Programming Competition
Goal to create a motion planning system to navigate around a virtual neighborhood and reach goal points with the least amount of energy, time, and distance for the Shell Eco-Marathon Autonomous Programming Virtual Competitions
Created a hybrid potential field motion planner and vehicle control system that took in LIDAR and processed Vision data to plan the vehicle’s motion and used a local and global map to navigate
Designed hybrid potential field method by having the path to each goal broken up into waypoints and then each waypoint targeted with the potential field to create a direction vector
Inter-system communication done through ROS
Designed and implemented central command system that picked and managed navigation targets for the motion planner and calculated shortest path to each goal
Innovated new modular simulation, benchmarking, and interface display system that allowed for configurations and systems to be quickly tested in minutes and validated offline in different maps, saving the team 30+ hours of online testing and development time
Earned 2nd place out of 5 teams in the Shell Eco-Marathon Autonomous Programming Beta Competition utilizing this system
Achieved 3-millisecond runtime for the motion planning system by unit testing and benchmarking systems
Implemented collision detection and pathing recovery as well as 3 Proportional Integral Derivative (PID) controllers for steering, forward throttle, and reverse throttle for driving backwards to get unstuck
Introduced and implemented unit testing frameworks to the team and mentored new members