2013/09/03 by Frank Havlak, Mark Campbell, Havlak, Frank +1 · 2 citations
Computer Science · #FOS: Computer and information sciences #Gaussian Processes and Bayesian Inference #Robotics (cs.RO) #Target Tracking and Data Fusion in Sensor Networks #Time Series Analysis and Forecasting
paper · pdf · doi:10.48550/arxiv.1309.0766
openalex publication_date 2013/09/03 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
This paper develops a probabilistic anticipation algorithm for dynamic\nobjects observed by an autonomous robot in an urban environment. Predictive\nGaussian mixture models are used due to their ability to probabilistically\ncapture continuous and discrete obstacle decisions and behaviors; the\npredictive system uses the probabilistic output (state estimate and covariance)\nof a tracking system, and map of the environment to compute the probability\ndistribution over future obstacle states for a specified anticipation horizon.\nA Gaussian splitting method is proposed based on the sigma-point transform and\nthe nonlinear dynamics function, which enables increased accuracy as the number\nof mixands grows. An approach to caching elements of this optimal splitting\nmethod is proposed, in order to enable real-time implementation. Simulation\nresults and evaluations on data from the research community demonstrate that\nthe proposed algorithm can accurately anticipate the probability distributions\nover future states of nonlinear systems.\n