China
Africa
Explore chapters and articles related to this topic
Search, Tracking, and Surveillance
Published in Yasmina Bestaoui Sebbane, Intelligent Autonomy of Uavs, 2018
The problem is to track a ground vehicle reactively while it is in view and plan a recovery strategy that relocates the target every time it is lost. The approach presented in [9] aims to handle big geographic areas, complex target motion models and long-term operations. The UAV is equipped with imaging systems allowing the observation of the target, prone to error and interference from terrain. The probability of observing the target on each observation cycle depends on how recently the target was last observed, the distance between the actual position of the target, the terrain and the mode of the imaging system. The camera has two modes:Wide-angle mode used to increase the area being scanned when the target is not currently observed, at the cost of a lower probability of successfully observing the target in any specific part of the image.Narrow-angle mode in which the viewing area is reduced, but the probability of detecting the target is higher.The effect of terrain is to reduce the probability of spotting the target in urban, suburban, forested and mountainous areas, while in rough or open rural areas the probability is higher. A faster moving target in the viewing zone is considered easier to spot. The probabilistic approach relies on the use of Recursive Bayesian Estimation (RBE) techniques that recursively update and predict the probability density function of the targets state with respect to time, under the assumption that the prior distribution and the probabilistic motion model of the target are known. Another strategy can be used as the orienteering problem with time windows (OPTW). The set of search patterns corresponds to the set of vertices of the OPTW, whereas the time slots in which the search patterns are active correspond to the time windows. If the UAV observer loses the target beyond the short period for which it tracks the predicted location of the target, it must follow a search strategy such as spiral or Boustrophedon to attempt to rediscover the target.
Stream travel time prediction using particle filtering approach
Published in Transportation Letters, 2018
B. Dhivyabharathi, E. S. Hima, L. Vanajakshi
Particle filtering is a technique to implement recursive Bayesian estimation using Sequential Monte Carlo methodology. The basic idea is that a posterior probability density function (PDF) of state can be represented by a set of particles with associated weights, and the estimate can be computed as the expected value of the discrete PDF (Arulampalam et al. 2002). It is based on point mass (particle) representations of probability densities which can be applied to any state space model including highly non-linear models with non-Gaussian noise densities. Initially, random samples (particles of filter) are generated using Monte Carlo methodology and are propagated and updated according to the system dynamics and measurement models.
Tsallis Divergence as Strategy for Radioactive Sources Search and Location
Published in Nuclear Science and Engineering, 2022
Chen Fu, Peng Xu, Yonggang Huo, Sufen Li, Xingfu Cai
According to Bayes’ theorem, the recursive Bayesian estimation method can be used to update the posterior probability density function (PDF) of the system state using new data obtained from detection. The posterior PDF is then used to find the optimal estimation of the state parameters: