An electric vehicle needs a forecast of its battery state of health (SOH). A useful forecast uses past data only, learns from few vehicles and needs little memory. Quantum circuits can build small models, but their benefit over classical layers of the same size is not established. This paper proposes MQAttn-QLSTM-VQR, a small model that combines multi-qubit quantum attention (MQAttn), a quantum long short-term memory (QLSTM) and a variational quantum residual (VQR). Quantum circuits compute only the query, key and value vectors; the similarity and softmax stay classical. The tests use 20 electric vehicles and two public battery datasets. Equal-size classical models and a classical twin, which replaces every quantum circuit with a classical layer, serve as references. The model forecasts better than the rule that repeats the last SOH on every held-out vehicle and cell. It shows no significant difference to a much larger classical network or to its classical twin. The benefit comes from the compact design and not from the quantum circuits. Simulated noise and limited measurements show how far the model tolerates hardware errors.
An electric vehicle needs a forecast of its battery state of health (SOH). A useful forecast uses past data only, learns from few vehicles and needs little memory. Quantum circuits can build small models, but their benefit over classical layers of the same size is not established. This paper proposes MQAttn-QLSTM-VQR, a small model that combines multi-qubit quantum attention (MQAttn), a quantum long short-term memory (QLSTM) and a variational quantum residual (VQR). Quantum circuits compute only the query, key and value vectors; the similarity and softmax stay classical. The tests use 20 electric vehicles and two public battery datasets. Equal-size classical models and a classical twin, which replaces every quantum circuit with a classical layer, serve as references. The model forecasts better than the rule that repeats the last SOH on every held-out vehicle and cell. It shows no significant difference to a much larger classical network or to its classical twin. The benefit comes from the compact design and not from the quantum circuits. Simulated noise and limited measurements show how far the model tolerates hardware errors.
Accurate prediction of the State of Health (SoH) in Lithium-ion (Li-ion) batteries is critical for increasing Electric Vehicles (EV) performance and lifespan. However, capturing complex patterns remains challenging as traditional Machine Learning (ML) models struggle with non-linear and high-dimensional data. We propose a Variational Quantum Neural Network (VQNN) model that uses the principles of quantum mechanics, such as superposition, entanglement, and parallelism, to predict SoH. QML applications in EVs are scarce, and VQNN addresses this gap by applying quantum computing in battery SoH prediction. Our VQNN model achieves a Root Mean Squared Error (RMSE) of 0.0346 on the NASA B0005 and B0006 battery datasets, outperforming the LSTM model by 0.0017. However, the VQNN requires a longer training time of 8021.81 s compared to the Neural Network (NN) at 4.45 s and the Long Short Term Memory (LSTM) at 9.46 s. We demonstrate that Quantum Machine Learning (QML) can be used to capture complex degradation patterns, paving the way for quantum-driven Battery Management Systems (BMS) in connected EVs. Our findings suggest that despite the computational overheads, advancements in quantum hardware could enable real-time SoH prediction, enhancing EV reliability in the transport ecosystem.
Accurate state-of-health (SOH) estimation for lithium-ion batteries remains a challenging problem due to complex electrochemical degradation mechanisms and long-range temporal dependencies. In this work, we propose a quantum-enhanced recurrent framework, termed QLSTM, in which variational quantum circuits are directly embedded into the gating mechanisms of long short-term memory networks. By replacing classical affine transformations with parameterized unitary operations, the proposed model introduces structured nonlinear transformations into the recurrent state-transition process. Extensive experiments on multiple benchmark battery datasets demonstrate that QLSTM consistently outperforms classical sequence models in both predictive accuracy and robustness, achieving significant reductions in mean absolute error (MAE), with improvements on the order of 20% compared with classical LSTM baselines. Ablation studies further confirm that these improvements arise primarily from quantum-enhanced gating rather than input-level transformations. Additional analyses on qubit scaling and noise robustness reveal that model performance is governed by a balance between expressive capacity and trainability. These results provide empirical evidence that embedding quantum computational primitives within recurrent architectures offers a structurally grounded approach to improving sequence modeling capability. The proposed framework establishes a new design paradigm for integrating quantum operators into temporal learning models, with potential applications in complex dynamical system prediction tasks.
Accurate battery state-of-health (SOH) estimation is essential for electric vehicles and energy-storage systems. We propose Coupled Quantum Reservoir Features (CQRF), a fixed ten-qubit two-register quantum reservoir. It processes battery cycles independently, producing 57 observables without recurrent feedback. With a Compact Bottleneck Readout (CBR) containing only 125 trainable parameters, CQRF achieves a macro mean absolute percentage error (MAPE) of 0.0172 across four datasets spanning nickel cobalt manganese (NCM), lithium iron phosphate (LFP), and nickel cobalt aluminum (NCA)/NCM-blend batteries, across 398 evaluated cells and ten seeds. Under matched internal settings, CQRF+CBR outperforms the recurrent Quantum Recurrent Reservoir (QRR)+CBR baseline at equal parameter count on all four dataset-level means and 9 of 11 evaluation contexts, while providing at least a 184-fold reduction in trainable parameters relative to the quantum physics-informed neural network (QPINN). Cross-dataset evaluation yields approximate MAPE-derived scores of 86% under leave-one-dataset-out and 85% under single-source transfer. Because normalization uses unlabeled target-dataset statistics, this represents cross-dataset generalization with mild unsupervised target-domain adaptation rather than strict zero-shot transfer. Architecture ablations, capacity-input removal, and cycle-summary controls support the coupled non-recurrent design as an alternative to recurrent quantum memory. Finally, 150 samples (50 each from HUST, MIT, and TJU) are executed on IBM’s 156-qubit ibm_marrakesh processor in an unmitigated quantum processing unit (QPU) audit. Simulator–QPU feature correlations of 0.18–0.41 and a TJU score reduction from 99.02% to 63.53% characterize current hardware execution as a feasibility and noise-sensitivity audit rather than performance validation.
Accurate prediction of the remaining useful life (RUL) of lithium-ion batteries is critical for enhancing the reliability and safety of battery-powered systems, especially in electric vehicles. While recent machine learning approaches such as Transformers and recurrent neural networks have shown promising results, their deployment in resource-constrained environments is limited due to high computational costs and large model sizes. In this paper, we propose a novel hybrid quantum-classical denoising Transformer (HQDeTransformer) architecture that integrates variational quantum circuits (VQCs) and quantum-enhanced attention mechanisms into a Transformer-based framework. The proposed HQDeTransformer employs a hybrid quantum denoising autoencoder to mitigate noise in capacity degradation sequences and utilizes hybrid quantum-classical neural networks to construct lightweight multi-head attention and feedforward blocks. Experimental evaluations on the NASA and CALCE benchmark datasets demonstrate that HQDeTransformer achieves superior performance compared to conventional machine learning models and quantum machine learning models, while reducing the number of trainable parameters. Furthermore, the use of angle-amplitude quantum embeddings and deeper quantum layers contributes to enhanced generalization and expressivity. These results highlight the potential of hybrid quantum-classical models for lightweight, accurate, and noise-resilient RUL prediction in practical battery management systems.
Energy storage devices such as lithium-ion batteries and supercapacitors are essential for portable electronics, electric vehicles, and renewable energy systems, yet their long-term reliability is limited by capacity degradation. Accurate prediction of state-of-health (SOH), state-of-charge (SOC), and capacity retention are therefore critical for extending device lifetimes and improving safety. Classical machine learning (ML) methods including LSTMs, CNNs, and ensemble approaches have achieved success in forecasting degradation trends, but they face limitations in scalability, data requirements, and capturing complex electrochemical behaviors. Quantum computing, and in particular quantum machine learning (QML), offers new opportunities by exploiting superposition and entanglement to process information more efficiently and compactly. This review surveys recent advances in applying QML to energy storage, with a focus on CQ approaches (classical data–quantum processing), distinguishing Hybrid-CQ architectures. Comparative analyses highlight trade-offs: CQ and Hybrid-CQ models achieve higher accuracy and parameter efficiency but are constrained by noise, limited qubits, and slower runtimes on real devices. Looking forward, integrating error-mitigation strategies, benchmarking on actual quantum hardware, and embedding physics-informed modeling are critical to closing the gap between theoretical promise and practical deployment. Beyond prediction tasks, quantum approaches hold potential for dynamic charging optimization, materials discovery, and quantum battery concepts. Collectively, these developments underscore how QML can complement and eventually surpass classical ML, paving the way toward more accurate, sustainable, and efficient energy storage systems.
Abstract Hybrid quantum recurrent models are commonly evaluated against compact classical networks, making it diffcult to determine whether observed gains arise from the variational quantum circuit (VQC) or from the surrounding recurrent architecture. This study presents an architecture and parameter-controlled evaluation of quantum gated recurrent unit (QGRU) and quantum long short-term memory (QLSTM) models for one-step-ahead meteorological forecasting. Each quantum model is compared with both a compact, approximately parameter-matched recurrent baseline and an exact classical gate-core surrogate that preserves the hidden width, recurrent equations, gate-specific pre-/post-projections, optimization protocol, and total trainable-parameter count. All models use 17 input features, 24-step histories, a chronological 80/10/10 train–validation–test split, trainingonly standardization, and ten matched random seeds. QGRU achieves an RMSE of 1.1045 ± 0.0693, reducing test RMSE by 19.01% relative to the compact GRU and by 5.47% relative to its exact 773-parameter surrogate. It also attains a lower mean best validation MSE, reaches 95% of its achievable validation improvement after 5.5 ± 0.5 epochs, and exhibits a smaller final train–validation gap. QLSTM reduces test RMSE by 24.80% relative to the compact LSTM, while its exact 1,025-parameter surrogate achieves comparable predictive performance with slightly greater validation stability. Analysis of 1,949 epoch-level records across convergence, checkpoint quality, early stopping, generalization, and runtime demonstrates that VQC utility depends on the recurrent architecture in which it is embedded. All experiments are conducted using ideal noiseless statevector simulation.