Low-Complexity and Low-Energy RNN-Enhanced Kalman Filter for Channel Estimation: A Case Study

In radiocommunications, multipath propagation often induces Inter-Symbol Interference (ISI), which requires accurate channel estimation and equalization. Neural networks have recently shown promising performance for this task but remain limited by their high computational cost and lack of interpretability, hindering their adoption in mission-critical systems. To overcome these limitations, this study proposes a hybrid estimation architecture that combines a linear Kalman filter with two lightweight neural networks that dynamically and deterministically adjust the noise covariance matrices of the filter, enabling efficient and explainable channel tracking. Experimental results show performance gains of up to 5 dB at low SNR and 3 dB at high SNR compared with the Least Squares (LS) algorithm. This low-dimensional design enables real-time deployment on energy-constrained programmable platforms. Implementations on an ARM Cortex-A72 CPU and Artix-7/Virtex-7 FPGAs demonstrate throughputs of 53.7 Mbps and 8-20 Mbps, respectively, with FPGA energy consumption over $20\times $ lower than on the CPU. These results confirm the effectiveness of the proposed NN-enhanced Kalman estimator for interpretable and energy-efficient channel estimation in embedded and defense applications.

• Publication year

  2026

• Venue

  IEEE Open Journal of the Communications Society

• Publication date

  Unknown publication date

• Fields of study

  Not labeled

• Identifiers
  DOI 10.1109/OJCOMS.2026.3665311
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• A Review on Large Language Models: Architectures, Applications, Taxonomies, Open Issues and Challenges

  2024cited by this paper
• RTiL: Real-Time Inference of Large Language Models on Memory-Constrained GPU Devices

  2024cited by this paper
• Large Language Model Inference Acceleration: A Comprehensive Hardware Perspective

  2024cited by this paper
• Contemporary Advances in Neural Network Quantization: A Survey

  2024cited by this paper
• A Comprehensive Survey of Convolutions in Deep Learning: Applications, Challenges, and Future Trends

  2024cited by this paper
• LCRM: Layer-Wise Complexity Reduction Method for CNN Model Optimization on End Devices

  2023cited by this paper
• The Smart Kalman Filter: A Deep Learning-Based Approach for Time-Varying Channel Estimation

  2023cited by this paper
• Gradient distribution-aware INT8 training for neural networks

  2023cited by this paper
• A Survey on Convolutional Neural Network Accelerators: GPU, FPGA and ASIC

  2022cited by this paper
• FlexCNN: An End-to-end Framework for Composing CNN Accelerators on FPGA

  2022cited by this paper
• Towards deep learning-aided wireless channel estimation and channel state information feedback for 6G

  2022cited by this paper
• A novel framework for deployment of CNN models using post-training quantization on microcontroller

  2022cited by this paper
• FP8 Quantization: The Power of the Exponent

  2022cited by this paper
• Deep Neural Network Based Channel Estimation for Massive MIMO-OFDM Systems With Imperfect Channel State Information

  2022cited by this paper
• 3D Trajectory and Transmit Power Optimization for UAV-Enabled Multi-Link Relaying Systems

  2021cited by this paper
• CNN Inference Costs Estimation on Microcontrollers: the EST Primitive-based Model

  2021cited by this paper
• A Survey on the Optimization of Neural Network Accelerators for Micro-AI On-Device Inference

  2021cited by this paper
• Dual CNN-Based Channel Estimation for MIMO-OFDM Systems

  2021cited by this paper
• KalmanNet: Neural Network Aided Kalman Filtering for Partially Known Dynamics

  2021cited by this paper
• Nonlinear Kalman Filter-Based Robust Channel Estimation for High Mobility OFDM Systems

  2021cited by this paper
• A Survey of Convolutional Neural Networks: Analysis, Applications, and Prospects

  2020cited by this paper
• High Dimensional Channel Estimation Using Deep Generative Networks

  2020cited by this paper
• Deep Learning Based Channel Estimation Schemes for IEEE 802.11p Standard

  2020cited by this paper
• A Memory-Optimized and Energy-Efficient CNN Acceleration Architecture Based on FPGA

  2019cited by this paper
• DynaNet: Neural Kalman Dynamical Model for Motion Estimation and Prediction

  2019cited by this paper
• SIMD stealing: Architectural support for efficient data parallel execution on multicores

  2019cited by this paper
• Neural Network-Based Uncertainty Quantification: A Survey of Methodologies and Applications

  2018cited by this paper
• Efficient SIMD implementation for accelerating convolutional neural network

  2018cited by this paper
• FPGA-Based Convolutional Neural Network Architecture with Reduced Parameter Requirements

  2018cited by this paper
• Long Short-Term Memory Kalman Filters: Recurrent Neural Estimators for Pose Regularization

  2017cited by this paper
• LS and MMSE estimation with different fading channels for OFDM system

  2017cited by this paper
• Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inference

  2017cited by this paper
• A neural network based channel estimation scheme for OFDM system

  2016cited by this paper
• SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <1MB model size

  2016cited by this paper
• A scalable GPU-enabled framework for training deep neural networks

  2016cited by this paper
• Improving neural network performance on SIMD architectures

  2015cited by this paper
• Kalman filter based target tracking for track while scan data processing

  2015cited by this paper
• On the Properties of Neural Machine Translation: Encoder–Decoder Approaches

  2014cited by this paper
• Data-Aided Kalman Tracking for Channel Estimation in Doppler-Affected OFDM Systems

  2007cited by this paper
• IMM Based Kalman Filter for Channel Estimation in UWB OFDM Systems

  2007cited by this paper
• Automatic digital modulation recognition using artificial neural network and genetic algorithm

  2004cited by this paper
• Adaptive Kalman Filtering for Low-cost INS/GPS

  2002cited by this paper
• A New Approach to Linear Filtering and Prediction Problems

  2002cited by this paper
• A formal model for SIMD computation

  1988cited by this paper
• The interacting multiple model algorithm for systems with Markovian switching coefficients

  1988cited by this paper
• Adaptive equalization

  1985cited by this paper

• No citing papers are available for this paper.