Abstract:This paper proposes a hybrid quantum neural network (HQNN) for indoor user localization using received signal strength indicator (RSSI) values. We use publicly available RSSI datasets for indoor localization using WiFi, Bluetooth, and Zigbee to test the performance of the proposed HQNN. We also compare the performance of the HQNN with the recently proposed quantum fingerprinting-based user localization method. Our results show that the proposed HQNN performs better than the quantum fingerprinting algorithm since the HQNN has trainable parameters in the quantum circuits, whereas the quantum fingerprinting algorithm uses a fixed quantum circuit to calculate the similarity between the test data point and the fingerprint dataset. Unlike prior works, we also test the performance of the HQNN and quantum fingerprint algorithm on a real IBM quantum computer using cloud quantum computing services. Therefore, this paper examines the performance of the HQNN on noisy intermediate scale (NISQ) quantum devices using real-world RSSI localization datasets. The novelty of our approach lies in the use of simple feature maps and ansatz with fewer neurons, alongside testing on actual quantum hardware using real-world data, demonstrating practical applicability in real-world scenarios.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is **the accuracy and efficiency of indoor user positioning**, especially its performance in highly interfering environments. Specifically:
1. **Limitations of traditional methods**:
- Traditional Global Navigation Satellite Systems (GNSS), such as GPS, face challenges such as signal attenuation and multipath effects in indoor environments.
- Geometric methods (such as triangulation and multilateration) are limited by multipath distortion and non - line - of - sight (NLoS) problems in practical applications, and require high communication overhead and precise synchronization.
- Although the fingerprint identification method is effective, it depends on large - scale data sets and parameter tuning, and has a long computational training time, which limits its accuracy and practicality in practical applications.
2. **Advantages and challenges of quantum computing**:
- Quantum computing has the potential to handle complex optimization tasks, but the existing quantum fingerprint algorithms lack trainable parameters, are difficult to deal with highly interfering and sparse data environments, and as the size of the data set increases, the number of required qubits also increases, and there are scalability problems.
3. **The proposed new method**:
- The paper proposes an indoor user positioning method based on the Hybrid Quantum Neural Network (HQNN), which uses Received Signal Strength Indication (RSSI) values for positioning.
- HQNN combines the advantages of classical machine learning and quantum computing, realizes efficient data processing through simple feature mapping and a small number of neurons, and has been tested on actual quantum hardware, showing its application potential in the real world.
4. **Main contributions**:
- A new HQNN model for RSSI - based user positioning is proposed.
- The performance of HQNN under different wireless technologies (WiFi, Bluetooth, Zigbee) is verified using public RSSI data sets.
- The performance of HQNN and the quantum fingerprint algorithm is tested on IBM quantum computers, providing practical insights into the feasibility and advantages of NISQ computing in indoor positioning.
### Formula display
- **Mean Squared Error (MSE) loss function**:
\[
L(\hat{z}, z)=\frac{1}{n}\sum_{i = 1}^{n}\left[(\hat{x}_i - x_i)^2+(\hat{y}_i - y_i)^2\right]
\]
where \(\hat{z}\) is the predicted coordinate, \(z\) is the real coordinate, and \(n\) is the number of training data points.
- **Adam optimizer update rule**:
\[
m_t=\beta_1m_{t - 1}+(1-\beta_1)\nabla_\theta L
\]
\[
v_t=\beta_2v_{t - 1}+(1-\beta_2)(\nabla_\theta L)^2
\]
\[
\hat{m}_t=\frac{m_t}{1-\beta_1^t}
\]
\[
\hat{v}_t=\frac{v_t}{1-\beta_2^t}
\]
\[
\theta_{t + 1}=\theta_t-\eta\frac{\hat{m}_t}{\sqrt{\hat{v}_t}+\epsilon}
\]
- **Parameter Shift Rule**:
\[
\Delta\Phi=\frac{1}{2}\left(L(\Phi+\frac{\pi}{2})-L(\Phi-\frac{\pi}{2})\right)
\]
\[
\Phi_{t + 1}=\Phi_t-\eta\cdot\Delta\Phi
\]
Through these methods, the paper shows the superior performance of HQNN in processing highly interfering data and proves its feasibility on actual quantum hardware.