TY - JOUR TI - THERMAL VISION BASED INTELLIGENT SYSTEM FOR HUMAN DETECTION AND TRACKING IN MOBILE ROBOT CONTROL SYSTEM AU - Ćirić Ivan T AU - Ćojbašić Žarko M AU - Ristić-Durrant Danijela D AU - Nikolić Vlastimir D AU - Ćirić Milica V AU - Simonović Miloš B AU - Pavlović Ivan R JN - Thermal Science PY - 2016 VL - 20 IS - 15 SP - 1553 EP - 1559 PT - Article AB - This paper presents the results of the authors in thermal vision based mobile robot control. The most important segment of the high level control loop of mobile robot platform is an intelligent real-time algorithm for human detection and tracking. Temperature variations across same objects, air flow with different temperature gradients, reflections, person overlap while crossing each other, and many other non-linearities, uncertainty and noise, put challenges in thermal image processing and therefore the need of computationally intelligent algorithms for obtaining the efficient performance from human motion tracking system. The main goal was to enable mobile robot platform or any technical system to recognize the person in indoor environment, localize it and track it with accuracy high enough to allow adequate human-machine interaction. The developed computationally intelligent algorithms enables robust and reliable human detection and tracking based on neural network classifier and autoregressive neural network for time series prediction. Intelligent algorithm used for thermal image segmentation gives accurate inputs for classification.