Source-linked AI summary
Integrating Sensing, Computing, and Communication in 6G Wireless Networks: Design and Optimization
Qiao Qi, Xiaoming Chen, Ata Khalili, Caijun Zhong, Zhaoyang Zhang, Derrick Wing Kwan Ng
TL;DR
Emerging wireless services require 6G networks to combine sensing, computing, and communication over shared resources. The paper develops a unified ISCC framework with two MOOP-based joint beamforming algorithms, and simulations validate their effectiveness while revealing parameter impacts.
Problem
Emerging wireless services require 6G networks to provide target sensing, intelligent computing, and information communication over the same radio spectrum.
Method
The paper develops a unified ISCC framework and two MOOP-based joint beamforming algorithms for weighted overall performance maximization and total transmit power minimization.
Results
The proposed algorithms show excellent performance in extensive simulations, with system priorities, sensor counts, and base-station antenna counts affecting overall performance.
Takeaways & Limitations
ISCC performance can be adjusted to application preferences through system priorities and improved by selecting suitable numbers of sensors and base-station antennas.
Abstract
from arXiv · showhide
The roll-out of various emerging wireless services has triggered the need for the sixth-generation (6G) wireless networks to provide functions of target sensing, intelligent computing and information communication over the same radio spectrum. In this paper, we provide a unified framework integrating sensing, computing, and communication to optimize limited system resource for 6G wireless networks. In particular, two typical joint beamforming design algorithms are derived based on multi-objective optimization problems (MOOP) with the goals of the weighted overall performance maximization and the total transmit power minimization, respectively. Extensive simulation results validate the effectiveness of the proposed algorithms. Moreover, the impacts of key system parameters are revealed to provide useful insights for the design of integrated sensing, computing, and communication (ISCC).
I. INTRODUCTION
Emerging wireless services require 6G networks to integrate sensing, computing, and communication over shared resources. The paper proposes an ISCC framework, two joint beamforming designs, and parameter-based design guidance.
- 6G wireless networks need integrated target sensing, intelligent computing, and information communication to support emerging services.
- Prior work mainly integrates sensing and communication, computing and communication, or sensing and computing rather than all three functions simultaneously.
- The paper introduces a general ISCC framework enabling multiple-target sensing, multidimension computing, and multistream communication over shared wireless resources.
- Two joint beamforming algorithms respectively maximize weighted overall performance under transmit-power constraints and minimize total transmit power under performance requirements.
- The study uses simulations to evaluate the algorithms and examines system-parameter impacts to guide practical ISCC design.
II. SYSTEM MODEL
The system model uses intelligent multi-antenna sensors and a multi-antenna base station to jointly perform target sensing, model computing, and information communication. Sensors spatially beamform superposed signals toward the base station, which applies corresponding receive processing.
- The modeled 6G network contains a base station with N antennas and K multifunction sensors with M antennas each, where KM ≤ N.
- The region of interest is partitioned into equal-size pixels, whose size determines sensing accuracy, while clutters act as sensing interference.
- Each sensor superposes sensing, computing, and communication signals and uses spatial directional beamforming adapted to channel conditions.
- The base station estimates target reflection coefficients from received reflections, aggregates models through AirFL, and receives information communication signals.
- Sensing, computing, and communication signals are assumed mutually independent and Gaussian distributed for analysis.
A. Sensing for Target Observation
Target sensing uses collaborative sensor transmission and base-station receive beamforming to estimate target reflection coefficients. The resulting MSE measures sensing quality and directly affects target imaging accuracy.
- The base station uses a linear unbiased estimator and receive beamforming to enhance target signals, suppress interference, and estimate reflection coefficients.
- Sensing MSE measures the error between estimated and actual reflection coefficients, while interference from other sensing, computing, and communication signals affects the estimate.
- Higher sensing accuracy, represented by smaller reflection-coefficient MSE, produces more accurate target imaging.
B. Computing for Model Aggregation
The computing task uses AirFL to aggregate local model parameters from intelligent sensors into a global model at the base station. Receive beamforming reduces aggregation distortion, which is evaluated by computing MSE and linked to image-recognition accuracy.
- Multiple intelligent sensors cooperatively train and transmit local model parameters for AirFL aggregation into a global model at the base station.
- Each local model parameter is weighted by ξ_k = |D_k|/|D| before transmission, reflecting the sensor dataset size.
- Base-station receive beamforming reduces model-aggregation distortion caused by interference and channel fading.
- Computing accuracy is measured by the MSE between the aggregated computing signal and the desired global model, and higher accuracy improves image recognition.
C. Communication for Information Transmission
Communication quality is improved through BS receive beamforming that mitigates interference, with SINR serving as the communication performance metric.
- BS receive beamforming mitigates interference to improve the quality of the communication signal.
- SINR determines communication-signal quality and is therefore used as the communication performance metric.
- Sensing, computing, and communication performance are jointly determined by sensor transmit beams and BS receive beams, creating resource competition.
- The integrated system differs from multi-user communication because its sensing, computing, and communication signals have different types and purposes.
III. JOINT DESIGN OF SENSING, COMPUTING AND COMMUNICATION
The paper jointly designs transmit and receive beamforming for ISCC by formulating two multi-objective optimization categories that represent different system priorities.
- Two MOOP categories are formulated: weighted overall performance maximization under a transmit-power budget and total transmit-power minimization under sensing, computing, and communication QoS constraints.
A. Weighted Overall Performance Maximization Design
The WOPM design combines sensing, computing, and communication objectives through weighted optimization and solves the resulting non-convex problem with alternating optimization.
- A. Weighted Overall Performance Maximization Design: The individual SOOP building blocks separately optimize sensing distortion, computing distortion, or communication weighted sum-rate under transmit-power limits.
- A. Weighted Overall Performance Maximization Design: Communication weighted sum-rate maximization is transformed into weighted sum-modified-MSE minimization using an MMSE receiver and an auxiliary weight variable.
- A. Weighted Overall Performance Maximization Design: The WOPM formulation combines normalized sensing, computing, and communication objectives using priorities whose weights sum to one.
- A. Weighted Overall Performance Maximization Design: The WOPM problem is non-convex because transmit beams, receive beams, and weight variables are inter-coupled.
- A. Weighted Overall Performance Maximization Design: Alternating optimization divides WOPM into receive-beam, weight-variable, and transmit-beam subproblems and iterates until the objective converges.
- A. Weighted Overall Performance Maximization Design: The transmit-beam subproblem becomes a convex QCQP solved with an interior-point method, while Algorithm 1 summarizes the WOPM joint design.
B. Total Transmit Power Minimization Design
The TTPM design minimizes total sensor transmit power while enforcing sensing, computing, and communication QoS, using alternating optimization and semidefinite relaxation.
- B. Total Transmit Power Minimization Design: TTPM minimizes total transmit power while constraining sensing distortion, computation error, and communication SINR to their required QoS levels.
- B. Total Transmit Power Minimization Design: The TTPM problem is non-convex because transmit and receive beamforming variables are coupled in the QoS constraints.
- B. Total Transmit Power Minimization Design: The transmit-beam subproblem introduces Ck,j = ck,jcHk,j and converts the formulation into a semidefinite program after relaxing the rank-one constraint.
- B. Total Transmit Power Minimization Design: After relaxation, the transmit-beam problem is jointly convex and can be solved directly with CVX.
- B. Total Transmit Power Minimization Design: The optimal relaxed matrices have rank one, allowing eigenvalue decomposition to recover the corresponding transmit beamforming vectors.
- B. Total Transmit Power Minimization Design: Algorithm 2 summarizes the joint ISCC design based on total transmit-power minimization.
C. Convergence and Complexity Analysis of Proposed Algorithms
Both proposed algorithms converge through monotonic, lower-bounded objective sequences, while their per-iteration computational complexity is characterized using interior-point methods for the underlying conic and matrix constraints.
- Convergence Analysis: Algorithm 1 converges because each convex subproblem yields feasible solutions for the next iteration, producing a monotonically non-increasing, lower-bounded objective.The monotone bounded convergence theorem establishes convergence for a suitable number of iterations.
- Convergence Analysis: Algorithm 2 converges because total transmit power is monotonically non-increasing and lower bounded by its QoS constraints.MMSE receivers ensure the QoS requirements, while convex transmit-beam subproblems can be solved optimally via CVX.
- Complexity Analysis: The dominant per-iteration operations solve the optimization subproblems in Algorithms 1 and 2, whose SOC and LMI constraints can be handled by standard interior-point methods.For Algorithm 1, the relevant subproblem contains K SOC constraints and decision-variable dimension n1 = O(K2M 2).
- Complexity Analysis: The per-iteration complexities of Algorithms 1 and 2 scale with ln(1/ǫ)ς1,2 for solution precision ǫ > 0.The complexity expressions are summarized in Table II.
IV. SIMULATION RESULTS
Simulations evaluate convergence, parameter sensitivity, and algorithm comparisons for the proposed ISCC beamforming methods. The results show stable iterative behavior, performance gains from suitable priorities, SNR, sensors, and antennas, and strong performance of the proposed algorithms.
- Algorithm 1 converges to a stable equilibrium within few iterations, with sensing and computation errors decreasing while communication rate increases.
- Increasing communication priority improves communication performance while reducing sensing and computation performance, so priorities should balance the three objectives.
- Higher sensor SNR, more BS antennas, and more sensors improve Algorithm 1's overall performance across sensing, computing, and communication metrics.The sensor-count result is attributed to more accurate reflection estimation, model aggregation, and communication weighted sum-rate, while extra BS antennas provide spatial multiplexing gain.
- Algorithm 2 converges within no more than 10 iterations on average under different minimum communication SINR requirements.
- Relaxing computation and sensing requirements reduces total transmit power, while additional BS antennas improve performance but eventually provide diminishing gains and increase RF-chain cost.The simulations motivate selecting antenna count and performance requirements to balance system performance, power consumption, and implementation cost.
- Algorithm 2 achieves the best total-transmit-power performance across the entire required communication SINR range, especially at high communication requirements.
V. CONCLUSION
The paper proposes an integrated sensing, computing, and communication architecture for 6G and studies how system design choices affect overall performance. It also acknowledges that the integration remains general and that deeper ISCC challenges remain unresolved.
- The proposed ISCC architecture integrates sensing, computing, and communication for 6G wireless networks.
- Extensive simulations are reported to confirm the performance of the proposed algorithms.
- Varying system priorities can provide desired sensing, computing, and communication performance according to application preferences.
- Overall performance can improve by increasing suitable numbers of sensors and BS antennas.
- The study makes only a general integration of the three functions, while deeper ISCC challenges remain insufficiently addressed.
APPENDIX A PROOF OF THEOREM 1
The appendix derives an optimization solution using rank inequalities, Lagrangian and KKT conditions, and related matrix arguments. The proof concludes that the adopted semidefinite relaxation is tight.
- The appendix connects communication MMSE to SINR through an inverse relationship to transform the optimization objective.
- The proof uses Sylvester’s rank inequality and a rank-subadditivity lemma for matrix sums.
- Lagrangian multipliers and KKT conditions are introduced to characterize the optimal solution under Slater’s condition.
- The matrix-rank derivation concludes that the adopted semidefinite relaxation is tight.