Abstract
In this paper, an energy-based spectrum sensing algorithm is proposed to address the problem of Uplink (UL)/Downlink (DL) resource block identification in a single cell. This framework aims to provide a general UL/DL identification tool for interference management algorithms in autonomous cells. The proposed algorithm is based on the observation that user terminals can be active from any part of the cell during UL, whereas a base station is fixed at a single position during DL. Because of these spatial properties, UL/DL identification is impossible with a single sensing equipment—or Secondary User Equipment (SUE). A multiple soft-sensing framework is introduced, using path-loss as a spatial information provider. Theoretical analysis and simulations show that the proposed algorithm achieves high reliability, with an error probability decreasing exponentially to zero as the number of SUEs increases.
Introduction: The Challenge of UL/DL Identification
As 5G and heterogeneous networks evolve, small "micro-cells" are increasingly embedded within larger "macro-cells." To manage interference effectively without high-overhead communication between layers, micro-cells must autonomously identify the macro-cell's current state: is it transmitting from the base station to a user (Downlink) or is a user transmitting to the base station (Uplink)?
Standard spectrum sensing typically only detects frequency occupancy. Distinguishing between UL and DL is significantly more difficult because of path loss. From the perspective of a single sensor, a weak signal from a nearby mobile user (UL) can appear identical to a strong signal from a distant base station (DL).
Proposed Methodology: Multiple Soft-Sensing
To overcome the limitations of single-sensor detection, the authors propose a multiple soft-sensing framework. Instead of a single SUE making a binary guess, multiple sensors collect raw energy values and transmit them to a central processor. This transforms the problem into a geometric classification task within a multi-dimensional "Energy Space."
The Two-Step Algorithm
- Direction Matching: The system compares the vector of received energy levels against a set of "test vectors" representing potential user positions. This step identifies the most likely direction and distance of the transmitter if it were a mobile user.
- Hyperplane Decision: Once a direction is matched, the algorithm utilizes a decision hyperplane. It determines if the signal's magnitude and spatial distribution are more characteristic of the fixed central Primary Base Station (PBS) or a mobile Primary User Equipment (PUE) at that specific location.
By utilizing raw energy values (soft-decisions) rather than binary flags, the system preserves critical spatial information provided by path-loss signatures.
Key Insights and Results
The core of the paper's contribution lies in the "Power of Crowdsourcing." While a single sensor is mathematically incapable of distinguishing UL from DL in many scenarios, the "spatial signature" of a fixed base station becomes increasingly distinct from that of any individual user as the number of sensors ($N_s$) grows.
Exponential Reliability
Theoretical derivations and numerical simulations confirm that the probability of an identification error decreases exponentially toward zero as the number of sensing devices increases. In tested scenarios, the error probability dropped sharply once the number of SUEs exceeded 5 to 10.
Computational Efficiency and Robustness
The algorithm maintains high performance even in noisy environments by averaging thermal noise across multiple sensors. Furthermore, the authors found that increasing the number of "test vectors" beyond a certain point does not significantly improve accuracy, suggesting that a relatively coarse spatial model is sufficient for high-reliability identification.
Conclusion and Future Work
This framework provides a robust tool for autonomous interference management in single-cell environments. By leveraging the geometric properties of cellular networks, it achieves high reliability with simple energy measurements. Future research directions include extending this framework to multi-cell environments where interference from neighboring macro-cells may complicate energy signatures, and investigating the impact of complex channel effects like shadowing and fast fading.
Summary updated