Comparative Evaluation of Deep-Learning based DoA Estimators for Real-Time ISAC Beamforming
Authors: Anjali K. S., Manoj Kumar Panda, Chandramani Singh
Organizations: Department of Electronics and Communication Engineering, Amrita School of Engineering, Bengaluru, Amrita Vishwa Vidyapeetham, India · Department of Electronic Systems Engineering, Indian Institute of Science, Bangalore, India
Integrated Sensing and Communication (ISAC) systems rely on accurate Direction-of-Arrival (DoA) estimation for effective beamforming and spectral-efficiency enhancement. While numerous classical and deep-learning-based DoA estimation methods have been reported in the literature, their suitability for real-time ISAC operation remains insufficiently understood because practical deployment requires consideration of both estimation accuracy and computational latency. This paper presents a unified evaluation framework for systematically comparing classical and learning-based DoA estimators within an ISAC beamforming pipeline. Representative classical and learning-based DoA estimators, together with an ensemble architecture, are integrated with Singular Value Decomposition (SVD)-based beamforming and evaluated under identical operating conditions. The impact of DoA estimation accuracy, inference latency, beamforming-weight generation time, and resulting effective spectral efficiency is investigated. Results show that different estimators occupy distinct operating points in the accuracy-latency-performance design space. Although learning-based approaches substantially reduce estimation time compared with MUSIC, the overall processing latency increasingly becomes dominated by the computational cost of SVD-based beamforming. Among the evaluated methods, the Regression Convolutional Neural Network (RCNN) exhibits a favorable balance between estimation accuracy, computational efficiency, and beamforming performance, making it a promising candidate for real-time ISAC deployment. The findings further suggest that future low-latency ISAC systems may benefit from beamforming strategies that reduce or avoid computationally intensive SVD-based beamforming-weight generation.
Figures & tables
Methods
References
Limitations
Traditional (e.g., MUSIC, ESPRIT)
[ 2 , 3 , 4 , 5 , 6 ]
• High execution time. • Less accurate in noisy conditions.
Deep Learning (Classification)
[ 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 ]
• Limited to predefined angle grids • Inaccurate for off-grid DoAs. • High execution time and training time. • Requires large labeled datasets.
Deep Learning (Regression)
[ 16 , 17 , 18 , 19 , 20 , 21 , 22 ]
• Sensitive to training data distribution and outliers. • Vulnerable to unstable predictions in noisy conditions.
TABLE I: DoA Estimation Methods and Their Limitations.
Fig. 1: Hybrid beamforming ISAC Scenario
Model Name
Covariance matrix input
Input Size
Layer type
CLASS [ 8 ]
Full matrix (real, imaginary, phase)
16×16×3
4 Convolutional + FC
RDNN [ 20 ]
UT elements (Real and imaginary)
1×240
FC only
LDNN [ 21 ]
Average of UT elements along super-diagonals (Real and imaginary)
1×30
FC only
RCNN [ 22 ]
Full Covariance matrix
16×16
1 Convolutional +FC
TABLE II: Model Architectures and Configurations.
Fig. 2: RMSE vs. SNR performance with sinusoidal signals.
Fig. 3: RMSE vs. SNR performance with LFM signals.
Fig. 4: Transmitter beam pattern for BS transmitting sine waves
Fig. 5: Transmitter beam pattern for BS transmitting LFM signals.
Fig. 6: Prediction time needed for each method of DoA estimation
Fig. 7: Beamforming weight generation time for each method
Fig. 8: Beamforming weight generation time Vs. SVD time analysis
Fig. 9: Spectral efficiency for DL communication for sine wave
Fig. 10: Spectral efficiency for DL communication for LFM signal
Methods
DoA RMSE
DoA Pred.Time(s)
ESE (bps/Hz)
CLASS
0.1051
2.55×10−04
9.9738
RDNN
0.2809
4.6×10−5
10.2694
LDNN
17.7031
1.22×10−4
2.2426
MUSIC
0.8011
0.001
8.8898
RCNN
0.1887
2.2805×10−5
10.4252
Ensemble
0.1744
5.4×10−5
10.2738
TABLE III: Performance comparison at 10 dB using LFM signals.
Fig. 11: Performance evaluation based on distance from ideal operating point (LFM signals)
Future integrated sensing and communication (ISAC) architectures separate the sensing entity (SE) that acquires measurements from the sensing function (SF) that performs inference, creating a need for compact, task-oriented feedback on the SE-SF interface. Forwarding the raw channel frequency response or full per-link delay-Doppler-azimuth-elevation (DDAE) tensor is prohibitively expensive, while peak-only reporting discards target-discriminative structure under clutter. We propose a learning-based coarse-to-fine sensing pipeline with candidate-latent feedback for single-target estimation. At the SE, a lightweight convolutional scorer produces a dense delay-Doppler proposal map from pilot-based OFDM channel estimates, and a learned encoder constructs K compact C-dimensional candidate tokens by fusing per-candidate azimuth-elevation patches, normalized position, and confidence cues. The latents are uniformly quantized post-training to b bits and transmitted under a finite budget B_fb = bKC + 18K + 16 bits to the SF, which performs cross-candidate refinement, reranking, and joint four-parameter estimation. On a ray-traced urban scene with static and dynamic clutter, three operating points in the (K, C, b) design space achieve 96.33-98.88% detection at 107-806 bytes per coherent processing interval, compression ratios of 1.2-9.2 x 10^4 over the 8-bit DDAE magnitude tensor, reducing the SE-SF interface from multi-Gbit/s to sub-Mbit/s rates. Cross-scene evaluation on an independent campus-scale environment achieves 98.79-99.50% detection and at-or-better angular accuracy without retraining, indicating that the learned representation captures target-relevant structure that transports across scenes of comparable or lower clutter density.
In this paper, we study an unmanned aerial vehicle (UAV)-assisted integrated sensing and communication (ISAC) system, where a UAV enhances the sensing capability of a base station (BS) towards a target while ensuring reliable communication towards a downlink user. This architecture is practically attractive for future wireless networks due to the UAV's controllable mobility and adaptive sensing coverage in wireless environments. The sensing performance is characterized by the average Cramér-Rao bound (CRB), which quantifies the minimum variance of the unbiased angle-of-arrival estimation. To enhance the sensing performance, the UAV trajectory and beamforming parameters are jointly optimized under power and mobility constraints, while satisfying communication requirements to the downlink user. To address the resulting non-convex problem, we employ null-space projection for beamforming design and adopt deep reinforcement learning for the trajectory optimization over a discrete-time scale. In each time slot, beamforming is optimized based on the channel state information to improve CRB performance while mitigating interference between the BS and the communication user. Simulation results demonstrate that the proposed method significantly reduces the time-averaged CRB by over 10%, compared with the ISAC system without UAV assistance, and also achieves a higher sensing accuracy than both the fixed-UAV-trajectory and the maximum-ratio-transmission-based beamforming benchmarks.
Yi Yang, Qianqian Zhang, Huaxia Wang
Department of Electrical and Computer Engineering, Rowan University, NJ, USA
Direction-of-arrival (DOA) estimation is an important task in microphone array processing and many downstream applications. The steered response power with phase transform (SRP-PHAT) method has been widely adopted for DOA estimation in recent years. However, accurate SRP-PHAT estimation in 3D scenarios requires evaluating steering responses over thousands of candidate directions, severely limiting real-time performance on resource-constrained platforms. This challenge becomes even more critical for planar arrays, which are widely used in robotics due to their structural simplicity. Motivated by the fact that azimuth estimation is usually more reliable than elevation estimation for most arrays, we propose ASAP, an azimuth-priority strip-based search approach to planar microphone array DOA estimation in 3D. In the first stage, ASAP performs coarse-to-fine region contraction within azimuthal strips to lock azimuth angles while retaining multiple maxima through spherical caps. In the second stage, it refines elevation along the great-circle arc between two close candidates. Extensive simulations and real-world experiments validate the efficiency and merits of the proposed method over existing approaches.
Ming Huang, Shuting Xu, Leying Yang +6
Southern University of Science and Technology, Shenzhen, China · Institute of Science Tokyo, Japan