AI/ML based SS burst set and CSI-RS TRS configuration optimization and improving NR network power and spectral efficiency
A system for optimizing one of a 5G New Radio (NR) network or a 4G Long Term Evolution (LTE) network operation, includes: an artificial intelligence (AI) engine configured to: a) implement at least one of the following: i) collection of raw training data set; ii) training of an AI agent; iii) inference generation triggered by UE; iv) using at least one of network key performance indicator (KPI) report and network operator inputs; and v) network KPI collection after application of inferences; and b) at least one of infer and apply at least one of i) network-wide optimal per-gNB or per transmission/reception point (TRP) Synchronization Signal (SS) burst set, and ii) associated channel state information reference signal (CSI-RS) configurations for downlink (DL) reference signal transmissions, thereby enhancing at least one of network transmission power efficiency and spectral efficiency.
1. A method of optimizing one of a 5G New Radio (NR) network or a 4G Long Term Evolution (LTE) network operation, comprising:
defining a set of observation time windows T W , w≥1, at a gNB as follows:
min
η
1
,
η
2
,
…
η
S
SBeamIdx
{
S
D
∀
SS
Beams
with
#
of
RACH
detection
≥
N
th
}
Over
all
{
T
W
}
and
min
{
v
η
1
,
v
η
2
,
…
v
η
S
SBeamIdx
,
N
U
E
Connected
_
Tw
}
{
N
SS
≤
N
SS
3
GPP
max
}
over
all
{
T
W
}
such that
max
S
D
,
N
S
S
T
IA
≤
T
IA
_
max
T
IA
wherein:
T IA_max is a maximum initial access (IA) latency stipulated by a network operator, defined as
T
IA
=
(
⌈
s
D
N
S
S
⌉
-
1
)
T
S
S
+
T
last
+
T
PRACH
;
N SS is a number of synchronization signal (SS) blocks in an SS Burst Set: T SS is a SS burst set time: T last is a time to transmit the SS blocks in a last SS Burst Set: T PRACH is a time taken after a user equipment (UE) detects a beam to transmit Physical Random Access Channel (PRACH) in uplink (U): S D represents a maximum number of SS blocks needed for one
transmission/reception point (TRP) deployed in a target scenario: η 1 , . . . , η j , . . . η SSBeamldk are SS beam index at the gNB: v n j is inferred UE mobility indicator in a direction of beam index η j ;
N UE Connected_Tw is a number of connected UEs at the observation time window T W : N th is the minimum number of PRACH instance detected in a received beam direction at the gNB: N SS sGPPmax is a upper bound for a number of SS blocks in the SS Burst Set:
establishing, by an artificial intelligence (AI) engine through a learning process, a functional relationship between angular spans {Δ θ , Δ ø }, wherein Δ θ and Δ ø are respectively azimuth and elevation angles of the beams of both the gNB and the UE, and wherein S D is a function of Δ θ and Δ ø ;
selecting, by the AI engine through the learning process, a set of receive beam directions where detected PRACH transmissions exceed N th :
indirectly measuring the UE mobility in a cell by measuring a rate of change and number of PRACH detections over the set of observation time window T W , thus providing a UE mobility information:
finding, by the AI engine, a functional relationship among the UE mobility information, number of connected UEs in the gNB with N SS , and number of SS blocks in the SS Burst Set, upper bounded N SS sGPPmax ;
learning, by the AI engine, from observations available at the gNB, wherein a list of observations available at the gNB comprises at least one of:
a) receive beam index for PRACH and the statistics for a detected PRACH transmissions set:
b) cell-specific and beam-specific Reference Signal Received Power (RSRP) of the received PRACH signals at the gNB and statistics of SNR threshold, measurement of PRACH correction power;
c) number of received PRACH signals for a receive beam index at sample time instant; and
d) UE mobility or geolocation information:
deriving, by the AI engine, a set of values {S D , N SS } to maximize the value of T IA with restriction T IA ≤T IA_max ; and
wherein a training phase is further carried out to train the AI engine for the gNB, using at least one of the following inputs: the receive beam index η j over which PRACH is received: the UE mobility information; and the number of connected UEs.
2. The method according to claim 1 , wherein the list of observations available at the gNB further comprises at least one of:
e) number of detected PRACH transmission in every UL receive beams at the gNB; and
f) RSRP reported by the UJE for any other DL reference signal transmissions.
3. The method according to claim 1 , wherein the finding of the functional relationship by the AI engine is performed using deep reinforcement learning (DRL).
4. The method according to claim 1 , further comprising:
successfully detecting the U)E by the gNB when a received signal-to-noise ratio (SNR) of the received PRACH signal achieves a specified SNR threshold and the gNB detects the PRACH signal.