Technology

RRT*

An asymptotically optimal motion-planning algorithm that refines random search trees into the shortest path through high-dimensional configuration spaces.

Sertac Karaman and Emilio Frazzoli introduced RRT* in 2011 to solve the sub-optimality of standard sampling-based planners. The algorithm adds a rewiring step: it checks if connecting a new point through nearby neighbors creates a shorter path than the initial assignment. This ensures the solution converges to the global optimum (the shortest distance) as the sample size grows. It is the industry standard for high-dimensional tasks (like 7-axis robotic arms) and autonomous navigation in cluttered environments. By maintaining a cost-to-come metric for every node, RRT* delivers a path that is both collision-free and mathematically efficient.