Robotics Projects
Embedded Files
Development of autonomous car platform for motion planning and machine learning research
I led the design and development of an autonomous car platform for research in motion planning and machine learning. Below are the three versions of the platform we have gradually developed. Version 1 is the first prototype, and version 3 is the final version.
Version 1
Version 1
Chassis: RC monster truck
Processor: Arduino Pro
Sensor: Hall effect sensors for speed control
Communication: Bluetooth with MATLAB
Version 2
Version 2
Chassis: Tamiya RC electric semi-assembled chassis
Processor: NVIDIA JETSON TK1, Arduino Pro
Sensors: Optical detector for speed control
Communication: WiFi
Version 3
Version 3
Chassis: Tamiya RC electric semi-assembled chassis
Processor: NVIDIA JETSON TK1, Arduino Pro
Sensors: Optical detector for speed control, ZED Sterero Camera, Hokuyo LiDAR, Sparkfun IMU
Communication: WiFi
The variations of the final version for different experiments
Experiments and implementations on the autonomous car platforms
Collision avoidance at intersection: This experiment aims to test the platform's capability of dynamic speed regulation. Collision avoidance is done through speed adjustment.
Third_Person_View.mp4Lane merge: This experiment aims to evaluate a RRT*-based robust motion planning algorithm.
lanemerge.wmvThe escape: This experiment aims to evaluate a RRT*-based adaptive motion planning algorithm for pursuit-evasion game.
the escape.wmvThe dodge: This experiment aims to evaluate a feedback motion planning algorithm. A sudden disturbance (i.e., the manual change of the heading) is applied to see whether the platform is able to adjust in time to avoid collision and reach the goal.
FeedbackRRT CERS.mp4Page updated
Google Sites
Report abuse