2016/01/11 by Allan De Freitas, Lyudmila Mihaylova, De Freitas, Allan +7 · 1 citation
Computer Science · #Anomaly Detection Techniques and Applications #Applications (stat.AP) #FOS: Computer and information sciences #Gaussian Processes and Bayesian Inference #Target Tracking and Data Fusion in Sensor Networks
paper · pdf · doi:10.48550/arxiv.1601.02429
openalex publication_date 2016/01/11 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Autonomous systems such as Unmanned Aerial Vehicles (UAVs) need to be able to\nrecognise and track crowds of people, e.g. for rescuing and surveillance\npurposes. Large groups generate multiple measurements with uncertain origin.\nAdditionally, often the sensor noise characteristics are unknown but\nmeasurements are bounded within certain intervals. In this work we propose two\nsolutions to the crowds tracking problem - with a box particle filtering\napproach and with a convolution particle filtering approach. The developed\nfilters can cope with the measurement origin uncertainty in an elegant way,\ni.e. resolve the data association problem. For the box particle filter (PF) we\nderive a theoretical expression of the generalised likelihood function in the\npresence of clutter. An adaptive convolution particle filter (CPF) is also\ndeveloped and the performance of the two filters is compared with the standard\nsequential importance resampling (SIR) PF. The pros and cons of the two filters\nare illustrated over a realistic scenario (representing a crowd motion in a\nstadium) for a large crowd of pedestrians. Accurate estimation results are\nachieved.\n