Citation Context

...oblem of tracking a target when the sensors are asynchronous and their offsets are unknown. B. Brief Literature Survey The issue of handling the asynchronous activity of sensors has been addressed in =-=[3]-=-, in the context of a multisensormultitarget bias estimation problem, but assuming that all timing offsets are known. Similarly, in [4] sensor registration is performed using Kalman filtering for asyn...

Citation Context

...ror performance in the absence of a closed-form solution [7]. This bound has been used increasingly in the context of sensor management in radar scheduling, submarine tracking, and terrain navigation =-=[8]-=-, [9]. We employ the Principle Component Analysis (PCA) to map the camera images into the eigenspace. The PCA has shortcomings in handling structured noise such as occlusions due to its global modelli...

Citation Context

...e been multiple attempts within either software or hardware to mitigate the errors from these components by modeling the error with Markov models and estimating the bias using a Kalman filter [22][23]=-=[24]-=-. One limitation of modeling the bias within an algorithm is that it does not account for calibrations of the bias that might occur over the period an object is being localized. Calibrations are used ...

Citation Context

...ystem by partitioning the state. More recently in Refs. [32] and [54], bias estimation in a radar tracking context with multiple targets was addressed. Bias estimation was further considered in Refs. =-=[37]-=- and [45], where multiple sensors were used to track multiple targets in a centralized formulation. Bias estimation has also been considered in a least squares estimation context by Dogancay in Ref. [...

Citation Context

...nabling fast manoeuvre detectionsand possibly airport map information.sMuch effort has been devoted in the last years to thesdefinition of bias estimation procedures for MMTT systemss(among others [2]=-=[3]-=-[4]). The basic idea is estimating allsbias terms in the measurements potentially causingsconsistency mismatch, and removing them from the rawsmeasures, providing the tracking filters with bias correc...

Citation Context

... distributed fusion and the data processing order does not matter at all. Remark: It is assumed above that all Ns sensors work synchronously. But in practice, 107most sensors may work asynchronously =-=[1, 71, 82, 88, 122]-=- due to different sampling rates, different initial times, etc. Following a similar sequential processing idea, the optimal asynchronous distributed fusion algorithm which uses the transformed data in...