Pr(AC1 alert)  1.70 · 10.1  2.50 · 10.1  2.50 · 10.1  4.77 · 10.1 

Pr(AC1 strengthening)  5.60 · 10.3  5.90 · 10.3  7.90 · 10.3  6.78 · 10.3 

Pr(AC1 reversal)  1.94 · 10.5  5.04 · 10.4  2.51 · 10.5  3.91 · 10.3 

E[AC1 RA changes]  1.80 · 10.1  2.57 · 10.1  2.56 · 10.1  6.48 · 10.1 

E[AC1 RA duration]  2.30 · 100  4.59 · 100  4.60 · 100  1.08 · 101 

The Joint-Indep strategy outperforms TCAS on all metrics: Joint-Indep is almost twice as safe and alerts almost three times less frequently. Because there is no sensor noise, Uncoord-Indep is just as safe as Joint-Indep but at the expense of more alerts. The Uncoord-Compat strategy is over twice as safe as Uncoord-Indep with the same alert rate. This is because Uncoord-Compat, unlike Uncoord-Indep, ensures compatibility. Uncoord-Compat is over four times safer than TCAS while alerting almost half the time.
4000 3000 2000 1000 0 .1000 .2000

Altitude (ft)
North (ft)
1
0.5
0
.0.5
.1
.1.5 East (ft) ×104
(b) Horizontal pro.le.
The
following
four
coordination
strategies
were
evaluated
with
sensor
noise
(Section
7.1)
on
the same encounters selected from the correlated encounter model:
.
joint policy / independent action selection (Joint-Indep),

.
uncoordinated policy / independent action selection (Uncoord-Indep),

.
joint policy / centralized action selection (Joint-Cent), and

.
uncoordinated policy / compatibility action selection (Uncoord-Compat).

The
results
are
summarized
in
Table
12.

TABLE 12
Performance evaluation of coordination strategies with the TCAS sensor

Joint-Indep  Uncoord-Indep  Joint-Cent  Uncoord-Compat  TCAS 

Pr(AC1 alert)  2.97 · 10.1  4.07 · 10.1  3.05 · 10.1  4.07 · 10.1  5.12 · 10.1 

Pr(AC1 strengthening)  1.74 · 10.2  1.92 · 10.2  1.84 · 10.2  1.87 · 10.2  6.54 · 10.3 

Pr(AC1 reversal)  8.99 · 10.4  5.00 · 10.3  1.53 · 10.4  1.40 · 10.3  3.60 · 10.3 

E[AC1 RA changes]  3.27 · 10.1  4.31 · 10.1  3.36 · 10.1  4.26 · 10.1  6.94 · 10.1 

E[AC1 RA duration]  4.73 · 100  8.44 · 100  4.84 · 100  8.48 · 100  1.13 · 101 

The Joint-Indep strategy performs markedly worse than TCAS. Because the aircraft can disagree over the belief state, they can issue incompatible advisories that if followed lead to collision. The Uncoor-Indep strategy, though it also has the potential of issuing incompatible advisories, is safer than Joint-Indep because it alerts earlier, perhaps allowing additional time to reverse the advisory if necessary. Joint-Cent, which solves this problem in part by using only the belief state of the aircraft with the lower address, has an NMAC probability ten times lower than Joint-Indep with comparable alert probability. The Joint-Cent strategy still has a somewhat greater probability of NMAC compared to TCAS but alerts signi.cantly less. The Uncoord-Compat strategy is over twice as safe as TCAS and alerts less often. One potential reason why Uncoord-Compat is safer than Joint-Cent is because Uncoord-Compat, using the uncoordinated logic which assumes that intruders do not maneuver to avoid collision, alerts more aggressively.
　
中國航空網 m.k6050.com
航空翻譯 www.aviation.cn
本文鏈接地址：Robust Airborne Collision Avoidance through Dynamic Programm(43)