RRT path planning: exploration of FSG'18 trackdrive circuit.
This video shows application of RRT-based path planning approach (RRT with multiple remote goals) for exploration of trackdrive circuit used in Formula Student Germany Driverless competition in 2018. Additionally, a concept of adaptive exploration speed depending on RRT performance is introduced to reduce the track exploration time.
After successful track exploration the control system is switched into more aggressive mode, which is used to achieve better racing performance in next laps. Although the implemented approach may look like a Model Predictive Control (MPC), it is a much simpler control strategy, which based on the analysis of reference path curvature in front of the car. As a result, this control technique provides good racing performance similar to MPC approaches, but does not require large computer power for implementing.
This simulation was done in scope of my Master's Thesis at Technical University Hamburg (TUHH). The full video of my Master's Thesis presentation can be found here: • RRT-based path planning and model pre...
The source code of my path planning algorithm is available here: https://github.com/MaxMagazin/ma_rrt_...
Formula Student Germany: https://www.formulastudent.de/
e-gnition Hamburg e.V.: https://egnition-hamburg.de/