English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

The integration of different types of navigation systems is frequently used in the automatic motion control systems due to the fact that particular errors existing in anyone of them are usually of different physical natures. The autonomous navigation systems are always preferred from many of reasons and the inertial navigation systems (INS) are traditionally considered as the main representative of this class. The integration of such systems based on the inertial sensors (rate gyros and linear accelerometers) and other navigation systems is very popular nowadays, especially as a combination with global positioning systems [Farrel & Barth, 1999], [Grewal et al., 2001]. The vision based navigation systems (VNS) are also of autonomous type and there is a reasonable intention to make the fusion of these two systems in some type of integrated INS/VNS system. This paper is oriented toward the possibility of fusion of data originated from a strap-down INS on one side, and from a dynamic vision based navigation system (DVNS), on the other. Such an approach offers the wide area of potential applications including the mobile robots and a number of automatically controlled ground, submarine, and aerial vehicles. The most usual approach in navigation systems integration is of “optimal filter” type (typical INS/VNS examples are given in [Kaminer et al., 1999] and [Roumeliotis et al., 2002]) In such an approach one of the systems is considered as the main one and the other supplies less frequently made measurements (corrupted by the noise, but still considered as the more precise) used in order to estimate in an optimal fashion the navigation states as well as the error parameters of the main system’s sensors. The approach adopted in this paper considers both systems in an equal way. It is based on the weighted averaging of their outputs, allowing some degrees of freedom in this procedure regarding to the actually estimated likelihood of their data. These estimates are based on reasoning related to the physical nature of system errors. The errors characterizing one typical strap-down INS are of slowly varying oscillatory nature and induced by the inaccuracies of inertial sensors. On the other hand, the errors in any VNS are mainly due to a finite resolution of a TV camera, but there is a significant influence of the actual scene structure and visibility conditions, also. In other words, it could be said that the accuracy of an INS is gradually decreasing in time while it is not affected by the fact where the moving object actually is. The accuracy of a DVNS is generally better in all situations where the

One Approach to the Fusion of Inertial Navigation and Dynamic Vision