Search

Situational awareness (SA) of unmanned aerial vehicles (UAVs) has been a research hotspot over the decades. Existing research mainly focuses on the SA of a single UAV in two-dimensional planes, whereas obstacles are assumed a priori. To eliminate these constraints, this study considers the cooperative situational awareness (CSA) problem of multi-UAV systems in the scenario of crossing a three-dimensional (3D) obstacle belt while no prior information of obstacles is required. First, the distribution models of the multi-UAV system and the obstacles are built based on two reference frames. Second, various types of uncertainties are characterized, which reflect an urgent need for CSA. Thus, a centralized CSA scheme is proposed and conducted on multiple UAVs and at different times, and the Dempster-Shafer (D-S) evidence theory is introduced to address information uncertainties and achieve high-accuracy information fusion. Next, to deal with the high-conflict evidence situations that are common in practice, a modified D-S fusion method is further developed. A modified Pearson coefficient is utilized to measure the correlation between different pieces of evidence. Both information credibility and uncertainty are taken into account to evaluate the evidence from different perspectives, and a novel evidence weight assignment method is presented to treat high-conflict situations. Numerical simulations validate the effectiveness of the proposed CSA method. Compared to existing studies, the proposed method is applicable to different trust paradoxes and achieves the best performance among various fusion methods. [ABSTRACT FROM AUTHOR]

Situational awareness (SA) of unmanned aerial vehicles (UAVs) has been a research hotspot over the decades. Existing research mainly focuses on the SA of a single UAV in two-dimensional planes, whereas obstacles are assumed a priori. To eliminate these constraints, this study considers the cooperative situational awareness (CSA) problem of multi-UAV systems in the scenario of crossing a three-dimensional (3D) obstacle belt while no prior information of obstacles is required. First, the distribution models of the multi-UAV system and the obstacles are built based on two reference frames. Second, various types of uncertainties are characterized, which reflect an urgent need for CSA. Thus, a centralized CSA scheme is proposed and conducted on multiple UAVs and at different times, and the Dempster-Shafer (D-S) evidence theory is introduced to address information uncertainties and achieve high-accuracy information fusion. Next, to deal with the high-conflict evidence situations that are common in practice, a modified D-S fusion method is further developed. A modified Pearson coefficient is utilized to measure the correlation between different pieces of evidence. Both information credibility and uncertainty are taken into account to evaluate the evidence from different perspectives, and a novel evidence weight assignment method is presented to treat high-conflict situations. Numerical simulations validate the effectiveness of the proposed CSA method. Compared to existing studies, the proposed method is applicable to different trust paradoxes and achieves the best performance among various fusion methods. [ABSTRACT FROM AUTHOR]