The International Journal of Robotics Research · 2014 · 113 citations · 49 references
ComputingComputer ScienceRobotics EngineeringHuman-robot InteractionMachine LearningRoboticsCollision DetectionTrajectory AnalysisTrajectory PredictionEnsemble Kalman FilterEnsemble Kalman FilteringHuman-robot NavigationMotion PredictionMotion Prediction For NavigationPedestrian DynamicsPedestrian Trajectory PredictionVelocity Obstacles
We introduce an online method that predicts pedestrian trajectories using agent‑based velocity‑space reasoning to enhance human–robot interaction and enable collision‑free navigation. The method models each pedestrian with velocity obstacles, adaptively learns parameters from sensor data, and computes agent‑specific motion models via ensemble Kalman filtering and maximum‑likelihood estimation. The approach learns individual motion parameters at interactive rates, outperforms prior techniques in real‑world crowded scenarios, and enables collision‑free robot navigation with improved trajectory accuracy.
We introduce a novel, online method to predict pedestrian trajectories using agent-based velocity-space reasoning for improved human–robot interaction and collision-free navigation. Our formulation uses velocity obstacles to model the trajectory of each moving pedestrian in a robot’s environment and improves the motion model by adaptively learning relevant parameters based on sensor data. The resulting motion model for each agent is computed using statistical inferencing techniques, including a combination of ensemble Kalman filters and a maximum-likelihood estimation algorithm. This allows a robot to learn individual motion parameters for every agent in the scene at interactive rates. We highlight the performance of our motion prediction method in real-world crowded scenarios, compare its performance with prior techniques, and demonstrate the improved accuracy of the predicted trajectories. We also adapt our approach for collision-free robot navigation among pedestrians based on noisy data and highlight the results in our simulator.
49
The EM Algorithm and Extensions
Debashis Kushary, Geoffrey J. McLachlan, Thriyambakam Krishnan · Technometrics · 1998 · 5.1K citations
The unscented Kalman filter for nonlinear estimation
Eric A. Wan, Rudolph van der Merwe · 2002 · 3.9K citations