SAGUIN: Integrated Space-Air-Ground-Underground Network

• SAGUIN is an integrated network architecture combining satellite, aerial, ground, and underground layers to support sustainable mMTC in challenging environments.
• It leverages LoRaWAN’s long-range low-power communication and RF wireless energy transfer to power and maintain underground sensor networks.
• Simulation results identify SF9 with a ~2/3 reporting period for WET as an optimal balance between energy harvesting, collision mitigation, and network longevity.

Searching arXiv for the specified paper to ground the article in current research. Space–Air–Ground–Underground Integrated Networks (SAGUIN) denote an integrated networking paradigm that extends massive machine-type communications (mMTC) into the subsurface by coupling satellite systems, aerial platforms, terrestrial infrastructure, and wireless underground sensor networks (WUSNs). In the formulation developed in "Toward Sustainable Subterranean mMTC: Space-Air-Ground-Underground Networks Powered by LoRaWAN and Wireless Energy Transfer" (Lin et al., 20 Aug 2025), SAGUIN is specifically positioned as a means to support sustainable subterranean mMTC in remote, disaster-stricken, and hard-to-reach areas by combining LoRaWAN for long-range low-power connectivity with Radio-Frequency Wireless Energy Transfer (WET) for self-sustaining operation of underground devices.

1. Architectural model and system scope

SAGUIN comprises four tightly coupled layers—Space, Aerial, Ground, and Underground—each assigned a distinct functional role in end-to-end sensing, transport, and energy support (Lin et al., 20 Aug 2025). The space segment is based on LEO satellites that provide truly global outreach, backhaul connectivity, and low-latency uplinks for sensor data aggregation. These satellites can directly act as LoRaWAN gateways in a direct-to-satellite configuration or receive relayed traffic from High-Altitude Platforms (HAPs) and Unmanned Aerial Vehicles (UAVs). The same segment also supports backbone transport via RF or free-space–optical feeder links to ground stations.

The aerial segment includes both HAPs and UAVs. HAPs float in the stratosphere at approximately $20\,\text{km}$, offering quasi-stationary wide beams, high-gain antennas, and solar-charged endurance. UAVs operate at a few hundred meters and provide on-demand relay, site inspection, and mobile WET sources. In this architecture, the aerial layer extends coverage where terrestrial infill is impossible, relays traffic toward satellites, and delivers energy beaming through WET.

The ground segment contains fixed LoRaWAN gateways, cellular base stations, and off-grid or grid-tied power beacons. Its role is low-cost backhaul in urban and rural zones together with primary WET anchoring, including solar-battery or diesel-generator hybrid configurations. The underground segment is composed of Underground Devices (UDs), namely buried sensors equipped with LoRaWAN radios, energy-harvesting modules, and batteries. These devices report variables such as soil moisture, leak indicators, or seismic motion, and operate through single-hop uplinks to aboveground gateways while also receiving WET.

The interconnection structure is explicitly denoted by U2H (UD→Hybrid station), U2V (UD→UAV), U2P (UD→HAP), H2V (Hybrid station→UAV), V2P (UAV→HAP), and P2S (HAP→Satellite). This layered decomposition supports three design considerations emphasized in the study: coverage extension through Non-Terrestrial Networks (NTN), mobility and agility through UAV repositioning and HAP beam adaptation, and reliability and resilience through multi-layer redundancy and satellite off-loading when ground links fail. A plausible implication is that SAGUIN is not merely a connectivity overlay but a coordinated multi-domain system in which communication reachability and energy availability are co-designed.

2. LoRaWAN-based communication model

LoRaWAN is adopted in SAGUIN because of its long range, ultra-low power, and freedom from operator fees (Lin et al., 20 Aug 2025). For underground-to-air or underground-to-ground links, the devices operate in the $868\,\text{MHz}$ ISM band with carrier frequency $f_c = 868\,\text{MHz}$, spreading factors $SF \in \{7,\ldots,12\}$, bandwidth $BW = 125\,\text{kHz}$, code rate $CR = 4/5$, transmit power $P_{tx} = 14\,\text{dBm}$, and antenna gains $G_{tx} = 2.15\,\text{dBi}$ and $G_{rx} = 25\,\text{dBi}$.

Packet reception is governed by the link-budget equation

$P_{rx}(d) = P_{tx} + G_{tx} - PL(d) + G_{rx},$

where $868\,\text{MHz}$0 includes both underground attenuation and any air-path loss. The underground attenuation depends on soil permittivity, volumetric water content, and burial depth. In the simulations, the underground path-loss model follows a modified Friis law with empirically derived soil parameters.

The spread factor is treated as a principal design variable because it induces a nontrivial reliability–airtime–energy trade-off. Higher $868\,\text{MHz}$1 yields larger processing gain, a lower SNR threshold, longer range, and better $868\,\text{MHz}$2, but also longer Time-on-Air (ToA), increased collision probability, and higher energy per packet. Lower $868\,\text{MHz}$3 provides a higher data rate and shorter ToA, but with reduced sensitivity and range. A common simplification is that increasing $868\,\text{MHz}$4 always improves operating efficiency; the reported results explicitly contradict that simplification by separating success probability from energy-per-packet minimization.

Within this formulation, LoRaWAN serves not only as the PHY/MAC choice for sparse and remote connectivity but also as the mechanism whose airtime properties directly couple to the energy budget, collision process, and lifetime analysis. This suggests that, in SAGUIN, communication design cannot be decoupled from harvesting and scheduling decisions.

3. Wireless energy transfer and subterranean energy sustainability

WET is introduced as the second enabling technology of the architecture, complementing LoRaWAN by supporting sustainable operation of buried devices (Lin et al., 20 Aug 2025). During a dedicated harvesting phase, a UD receives RF power from a power beacon or overflying UAV. Under a free-space approximation, the received RF power at distance $868\,\text{MHz}$5 is modeled by the Friis equation

$868\,\text{MHz}$6

With conversion efficiency $868\,\text{MHz}$7 and harvesting duration $868\,\text{MHz}$8, the harvested energy is

$868\,\text{MHz}$9

The study further states that underground propagation introduces extra losses due to soil absorption and refraction at the surface, so that $f_c = 868\,\text{MHz}$0 is reduced by a factor $f_c = 868\,\text{MHz}$1, where $f_c = 868\,\text{MHz}$2 depends on volumetric water content and soil composition. This is important because the same subterranean medium that degrades the communication channel also attenuates the energy-transfer channel. Consequently, underground sustainability depends not only on beacon power and harvesting time but also on geophysical properties.

The protocol-level integration is divided into two phases. In the WET phase, UDs switch to energy-harvesting mode and accumulate $f_c = 868\,\text{MHz}$3 from stationary beacons and overflying UAVs. In the data phase, they consume harvested energy for sensing, processing, and LoRaWAN uplink. The architecture therefore treats energy as a periodically replenished resource rather than a fixed battery reserve. A plausible implication is that SAGUIN transforms lifetime engineering from a battery-sizing problem into a joint propagation, harvesting, and access-control problem.

4. Time allocation, lifetime formulation, and optimization

To maximize device lifetime under periodic traffic, each reporting period $f_c = 868\,\text{MHz}$4 is partitioned into a WET phase of duration $f_c = 868\,\text{MHz}$5 and a data-transmission phase of duration $f_c = 868\,\text{MHz}$6 (Lin et al., 20 Aug 2025). Let $f_c = 868\,\text{MHz}$7 denote the initial battery energy, for example $f_c = 868\,\text{MHz}$8 for a $f_c = 868\,\text{MHz}$9 battery; let $SF \in \{7,\ldots,12\}$0 be the harvested energy per period; and let $SF \in \{7,\ldots,12\}$1 denote the average energy consumed for a packet, including retransmissions due to collisions, under $SF \in \{7,\ldots,12\}$2 and success probability $SF \in \{7,\ldots,12\}$3.

The per-period net energy change is

$SF \in \{7,\ldots,12\}$4

and the device lifetime in number of periods is stated roughly as

$SF \in \{7,\ldots,12\}$5

The joint optimization is

$SF \in \{7,\ldots,12\}$6

subject to

$SF \in \{7,\ldots,12\}$7

By brute-forcing over $SF \in \{7,\ldots,12\}$8 and solving $SF \in \{7,\ldots,12\}$9, the study obtains the $BW = 125\,\text{kHz}$0–$BW = 125\,\text{kHz}$1 pair that achieves the longest lifetime.

The paper also presents a simple search procedure that iterates over $BW = 125\,\text{kHz}$2, computes ToA, sweeps $BW = 125\,\text{kHz}$3, calculates $BW = 125\,\text{kHz}$4, $BW = 125\,\text{kHz}$5, and $BW = 125\,\text{kHz}$6, and returns the best $BW = 125\,\text{kHz}$7, $BW = 125\,\text{kHz}$8, and lifetime. The significance of this formulation lies in the explicit coupling of MAC-level contention, PHY-level spread factor selection, and energy harvesting within a single lifetime objective. This suggests that, in subterranean mMTC, nominally separate layers are operationally inseparable once long-term maintenance-free operation becomes the target metric.

5. Simulation methodology and pipeline-monitoring case study

The feasibility study is based on a remote underground pipeline monitoring scenario (Lin et al., 20 Aug 2025). The underground environment is parameterized by burial depth $BW = 125\,\text{kHz}$9, volumetric water content $CR = 4/5$0, and clay content $CR = 4/5$1, with underground path-loss exponent approximately $CR = 4/5$2, described as line-of-sight through soil. The aerial geometry places a HAP at height $CR = 4/5$3, with elevation angle $CR = 4/5$4 and link distance $CR = 4/5$5.

The radio configuration uses $CR = 4/5$6, $CR = 4/5$7, $CR = 4/5$8, $CR = 4/5$9, $P_{tx} = 14\,\text{dBm}$0 channels, and noise floor $P_{tx} = 14\,\text{dBm}$1. Traffic is generated by $P_{tx} = 14\,\text{dBm}$2 UDs, each sending a 10-byte packet every $P_{tx} = 14\,\text{dBm}$3 or $P_{tx} = 14\,\text{dBm}$4 using ALOHA uplink. The energy configuration assumes initial battery $P_{tx} = 14\,\text{dBm}$5 at $P_{tx} = 14\,\text{dBm}$6, harvested power $P_{tx} = 14\,\text{dBm}$7, and conversion efficiency $P_{tx} = 14\,\text{dBm}$8.

The performance metrics are precisely defined. The success probability is

$P_{tx} = 14\,\text{dBm}$9

Additional metrics are energy per packet $G_{tx} = 2.15\,\text{dBi}$0, system lifetime measured as average years before battery depletion, and throughput measured as useful bits per second averaged over the network.

The case study is noteworthy because it grounds the architectural proposal in a parameterized scenario rather than treating SAGUIN purely as a conceptual framework. At the same time, the authors identify the scenario as a simulation-based evaluation, not a field deployment. This distinction matters when interpreting quantitative outcomes.

6. Results, operating regimes, and open research problems

The reported results establish several regime-dependent behaviors (Lin et al., 20 Aug 2025). For link reliability versus spreading factor and soil conditions, shallow depth $G_{tx} = 2.15\,\text{dBi}$1 favors $G_{tx} = 2.15\,\text{dBi}$2, which yields the highest $G_{tx} = 2.15\,\text{dBi}$3 and lowest $G_{tx} = 2.15\,\text{dBi}$4 because of its extremely short ToA of $G_{tx} = 2.15\,\text{dBi}$5. Under harsher conditions $G_{tx} = 2.15\,\text{dBi}$6, $G_{tx} = 2.15\,\text{dBi}$7 maximizes $G_{tx} = 2.15\,\text{dBi}$8, but $G_{tx} = 2.15\,\text{dBi}$9 minimizes $G_{rx} = 25\,\text{dBi}$0 by balancing reliability and airtime. This directly counters the misconception that the spread factor that maximizes reliability must also minimize energy cost.

Lifetime as a function of WET duration is non-monotonic. It first rises with $G_{rx} = 25\,\text{dBi}$1 because more energy is harvested, then declines as $G_{rx} = 25\,\text{dBi}$2 shrinks, which increases collisions and $G_{rx} = 25\,\text{dBi}$3. An explicit example is that $G_{rx} = 25\,\text{dBi}$4 attains peak lifetime of approximately $G_{rx} = 25\,\text{dBi}$5 at $G_{rx} = 25\,\text{dBi}$6 out of $G_{rx} = 25\,\text{dBi}$7. The study further reports that $G_{rx} = 25\,\text{dBi}$8 plus $G_{rx} = 25\,\text{dBi}$9 emerges as a robust choice in the examined scenario.

The sensitivity to node density and reporting interval is also quantified. With $P_{rx}(d) = P_{tx} + G_{tx} - PL(d) + G_{rx},$0 and $P_{rx}(d) = P_{tx} + G_{tx} - PL(d) + G_{rx},$1 at $P_{rx}(d) = P_{tx} + G_{tx} - PL(d) + G_{rx},$2, lifetime falls from approximately $P_{rx}(d) = P_{tx} + G_{tx} - PL(d) + G_{rx},$3 for $P_{rx}(d) = P_{tx} + G_{tx} - PL(d) + G_{rx},$4 to approximately $P_{rx}(d) = P_{tx} + G_{tx} - PL(d) + G_{rx},$5 for $P_{rx}(d) = P_{tx} + G_{tx} - PL(d) + G_{rx},$6. Eliminating WET by setting $P_{rx}(d) = P_{tx} + G_{tx} - PL(d) + G_{rx},$7 reduces lifetime by a factor of approximately $P_{rx}(d) = P_{tx} + G_{tx} - PL(d) + G_{rx},$8, indicating the central role of harvested energy in long-term sustainability. Shortening the reporting period to $P_{rx}(d) = P_{tx} + G_{tx} - PL(d) + G_{rx},$9 nearly halves lifetime because uplink attempts and collision risk both increase.

The paper closes by identifying open challenges. Channel modeling remains under-studied, especially multi-layer path loss combining soil absorption and surface refraction under varying water content. CSI acquisition is difficult because harvesting and communication beamforming require accurate channel state information in harsh underground and air links, making learning-based CSI estimation with low pilot overhead a promising direction. Network coordination must address heterogeneous routes such as direct UD→HAP, UD→UAV→HAP, and UD→ground through adaptive routing, cross-layer scheduling, and platform-aware frequency-switching protocols. Scalability is constrained by LoRa’s pure ALOHA MAC, with LR-FHSS and rate-splitting multiple access identified as mechanisms to mitigate collisions in dense deployments. Efficient WET operation motivates UAV trajectory optimization, Reconfigurable Intelligent Surfaces for beam focusing, and CSI-free WET schemes. Sustainability of the beacons themselves introduces another research axis, including solar- or wind-driven power sources and joint energy-communication scheduling among beacons, UAVs, and UDs.

Taken together, these results characterize SAGUIN as a multi-layer architecture in which subterranean communications, aerial relaying, satellite backhaul, and RF energy replenishment are jointly optimized. The study shows that long-lived underground monitoring is feasible in simulation when LoRaWAN configuration and WET time allocation are selected in concert, while also making clear that channel realism, coordination overhead, and MAC scalability remain unresolved research issues.