In early cellular networks, deciding which base station a mobile device should connect to was simple: you connected to the closest physical tower. Because all towers were uniform macro-stations, geographic proximity correlated perfectly with signal strength.
In modern Heterogeneous Networks (HetNets), this geographic intuition breaks down. Today's network landscape is a multi-tier mix of high-power macro towers, low-power terrestrial small cells (picos and femtos), unmanned aerial vehicles (UAVs), and low Earth orbit (LEO) satellites. In this environment, a device might be physically closer to a low-power small cell, yet receive a significantly stronger signal from a massive, distant macro tower.
To navigate this complexity without incurring massive computational overhead, modern wireless systems rely on Maximum Average Received Power (MARP) association. Instead of tracking volatile, instantaneous channel fluctuations, MARP provides a long-term, statistically robust framework for device connectivity.
Why Average Power Trumps Instantaneous Tracking
Wireless signals experience two primary types of degradation: large-scale fading (caused by distance-based path loss and physical obstacles) and small-scale fading (caused by multi-path reflections as a user moves).
If a device attempted to associate with a base station based on instantaneous signal strength, it would constantly hop between towers due to rapid small-scale fading. This phenomenon, known as "ping-ponging," wastes critical network signaling overhead.
MARP solves this by averaging out small-scale fluctuations. It focuses entirely on long-term channel statistics, ensuring that user equipment (UE) connects to the node providing the most robust baseline link. As demonstrated by Matracia et al. in rural UAV-assisted networks, this strategy allows devices to seamlessly evaluate and switch between terrestrial and aerial infrastructure based on which tier offers the most effective long-term link, irrespective of the structural differences between the tiers.
Balancing Transmit Powers and Altitudes Across Tiers
The core strength of MARP lies in its ability to act as an equalizer across highly diverse network architectures. Consider a hybrid network containing both terrestrial towers and LEO satellites, as modeled by Ben Salem et al.. A satellite possesses immense transmit power but suffers from extreme path loss due to its orbital distance. By evaluating the maximum average received power, the user's device automatically balances these competing factors to select the optimal access point.
This balancing act becomes even more dynamic when networks expand vertically. In multi-layer UAV networks, Kim et al. show that assigning ground nodes to the aerial access point with the strongest average received power provides a practical, decentralized mechanism for maximizing transmission success. Similarly, Bushnaq et al. leverage this principle to map out distinct "regions of connectivity" between tethered drones and terrestrial base stations based on maximum signal-to-noise ratio (SNR) thresholds.
Overcoming Geometry with Antenna Gains and Line-of-Sight
MARP accounts for the fact that physical distance is often a poor proxy for channel quality due to two critical factors: directional antenna patterns and line-of-sight (LOS) dynamics.
As Rosas points out, a user might receive a superior connection from a distant base station if they happen to fall within its high-gain directional mainlobe, rather than a closer base station whose signal is attenuated because the user is caught in its sidelobe. This insight is echoed by Chen et al., who show that a strongest-power association policy naturally incorporates directional antenna gains to compensate for physical path loss over longer distances.
[Distant Tower] ========(High-Gain Mainlobe)========> [User Device] (Stronger Signal)
[Closer Tower] --------(Low-Gain Sidelobe)--------> [User Device] (Weaker Signal)
Furthermore, in high-frequency bands like millimeter-wave (mmWave), physical blockages drastically alter signal propagation. In mmWave vehicle-to-everything (V2X) platoon networks, Yi et al. demonstrate that MARP allows vehicles to dynamically bypass nearby obstructed base stations in favor of more distant, unobstructed line-of-sight links, shifting connections based on actual signal viability rather than proximity.
Simplifying Complex Network Optimization
Jointly optimizing user association and power control across thousands of devices is an NP-hard problem. However, under specific network conditions, MARP and its variants provide elegant mathematical simplifications that make real-time optimization tractable.
For instance, Hong et al. demonstrate that when a network requires a minimum signal-to-interference-plus-noise ratio (SINR) threshold of 0 dB, the complex joint optimization problem simplifies entirely into a linear assignment problem. In this regime, the optimal connection strategy can be determined simply by maximizing the sum of the logarithms of channel gains.
In uplink scenarios, the problem simplifies even further. Lin et al. prove that when users employ power control to meet a uniform signal-to-interference ratio target, the optimal association policy collapses back into a simple, distance-based rule: connect to the nearest base station. This distance-based approach remains optimal regardless of the transmit power or density of the base station tiers, drastically reducing computational overhead. A similar geometric simplification is identified by Shabanighazikelayeh et al., who show that connecting to the geographically closest UAV remains the rate-maximizing strategy even when accounting for probabilistic line-of-sight channel models.
Adapting to Next-Generation Surfaces and Multiplexing
As wireless networks evolve to include passive components and advanced multiplexing, the definition of "received power" has expanded, but the core MARP mechanism remains intact.
With the advent of Reconfigurable Intelligent Surfaces (RIS)—"smart mirrors" that reflect and focus wireless signals—a user's received power may come from an indirect path. Xu et al. demonstrate that MARP effectively scales to these environments by evaluating the cumulative power of both direct communication links and indirect links established via RIS reflections. In post-disaster scenarios where terrestrial infrastructure is damaged, Matracia et al. utilize a hybrid MARP rule to allow user equipment to dynamically bypass ground-level obstructions by toggling between surviving direct links and aerial RIS reflections.
Finally, MARP integrates seamlessly with advanced spectral efficiency techniques like Non-Orthogonal Multiple Access (NOMA). When multiple users share the same frequency band via power-domain multiplexing, Qin et al. show that user association can still be governed by maximum average received power, provided the policy explicitly accounts for the power-sharing coefficients inherent to NOMA. This ensures that connectivity decisions accurately reflect the true achievable SINR under shared resource conditions.
Key Foundational Work
The impact of association on the capacity of WLANs (2009) — contributed 1 concept elements to this direction.
Joint Downlink Base Station Association and Power Control for Max-Min Fairness: Computation and Complexity (2014) — contributed 1 concept elements to this direction.