Page 164 - ITU Journal Future and evolving technologies – Volume 2 (2021), Issue 2

ITU Journal on Future and Evolving Technologies, Volume 2 (2021), Issue 2

Mr-UAVs
6000
Helicopters
Fixed wing UAVs
5000 Small fixed wing planes
Max. velocity (m/s) 4000 Cruise missiles
Large fixed wing planes
Fighter jets
Birds
Ballistic missiles
3000
Rockets & artillery
HGVs
2000
1000
0
0 10 20 30 40 50 60 70
Length of central section (m)
Fig. 9 – Visualization of the length of the central section and maximum velocity of training targets.

Reflected laser beams
Transmitted laser beams

(a)

(b)
Fig. 10 – (a) Re lected laser beams to provide spatial diversity, (b) laser beam transmitted and re lected towards the source.

5.2 Localization and tracking of the target

ℳ = (NB) (M (Tr) , C (Tr) ), (15) The ( , , ) coordinates of the intersection of the laser
1
beams are unique as shown in Fig. 5(a. If one or more

laser intersections are blocked by a target, the coordi‑

ℳ = (LDA) (M (Tr) , C (Tr) ), (16) nates of the blocked intersection positions provide the lo‑
2

calization of the target at a given time instance. As the

ℳ = (KNN) (M (Tr) , C (Tr) ), (17) target moves, the localization information is also updated
3
in time. The center of the target is the center position of

ℳ = (RF) (M (Tr) , C (Tr) ). (18) the blocked laser intersections and mesh. For example in
4
Fig. 5(b, the center of the target is at the third blocked in‑
tersection of = 2 laser mesh. The tracking and mapping
of the target’s trajectory are obtained as the blocked laser
The interpolated target’s data and the classi ier models

areusedtoestimatetheparticular class ofthetargetusing intersection positions are updated as the target moves.

the prediction function given as The overall SDCLT process of a target is given in Algo‑
rithm 2. In Algorithm 2 at lines 4 and 5, a detection test

(est,ℳ q ) =predict(ℳ , M (eval) ), (19) is carried out / G < to determine the presence of

a target for each steering position. Once a target is de‑

tected, the shape features of the target are extracted (that
where = 1, 2, 3, 4 stands for the model used, and are provided in Table 1 . Next, the target is classi ied and
(est,ℳ q ) is the estimated class from the th model. The categorized based on the extracted features. The localiza‑
predictions are made using the relevant predict functions tion and tracking of the target are also carried out at dif‑
speci ic to the models deployed. ferent steering positions (discussed in Section 5.2 . The

)
extraction of features, classi ication, and localization, and
tracking of the target are provided at lines 6 to 9 of
Algorithm 2.
150 © International Telecommunication Union, 2021