Damen Naval
The Dynamic Risk-Aware MPPI (DRA-MPPI) framework enables mobile robots to navigate crowded environments by efficiently approximating joint collision probabilities from multi-modal human movement predictions. This method maintains high success rates (98-99%) and low collision probabilities while preserving operational efficiency in both simulations and real-robot deployments.
Researchers at TU Delft developed Dynamic Risk-Aware MPPI (DRA-MPPI), a motion planning framework that enables mobile robots to safely navigate human crowds by efficiently approximating the joint collision probability with multiple dynamic, non-Gaussian obstacles. DRA-MPPI achieved 98-100% success rates in simulations and successful real-world navigation, demonstrating enhanced safety and efficiency over baseline methods.
There are no more papers matching your filters at the moment.