| 1. |
- Banugondi Rajashekara, Manoj, et al.
(author)
-
Adversarial Attacks on Deep Learning Based Power Allocation in a Massive MIMO Network
- 2021
-
In: IEEE INTERNATIONAL CONFERENCE ON COMMUNICATIONS (ICC 2021). - : IEEE. - 9781728171227
-
Conference paper (peer-reviewed)abstract
- Deep learning (DL) is becoming popular as a new tool for many applications in wireless communication systems. However, for many classification tasks (e.g., modulation classification) it has been shown that DL-based wireless systems are susceptible to adversarial examples; adversarial examples are well-crafted malicious inputs to the neural network (NN) with the objective to cause erroneous outputs. In this paper, we extend this to regression problems and show that adversarial attacks can break DL-based power allocation in the downlink of a massive multiple-input-multiple-output (maMIMO) network. Specifically, we extend the fast gradient sign method (FGSM), momentum iterative FGSM, and projected gradient descent adversarial attacks in the context of power allocation in a maMIMO system. We benchmark the performance of these attacks and show that with a small perturbation in the input of the NN, the white-box attacks can result in infeasible solutions up to 86%. Furthermore, we investigate the performance of black-box attacks. All the evaluations conducted in this work are based on an open dataset and NN models, which are publicly available.
|
|
| 2. |
- Manoj, Banugondi Rajashekara, et al.
(author)
-
Sensing and Classification Using Massive MIMO : A Tensor Decomposition-Based Approach
- 2021
-
In: IEEE Wireless Communications Letters. - : IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC. - 2162-2337 .- 2162-2345. ; 10:12, s. 2649-2653
-
Journal article (peer-reviewed)abstract
- Wireless-based activity sensing has gained significant attention due to its wide range of applications. We investigate radio-based multi-class classification of human activities using massive multiple-input multiple-output (MIMO) channel measurements in line-of-sight and non line-of-sight scenarios. We propose a tensor decomposition-based algorithm to extract features by exploiting the complex correlation characteristics across time, frequency, and space from channel tensors formed from the measurements, followed by a neural network that learns the relationship between the input features and output target labels. Through evaluations of real measurement data, it is demonstrated that the classification accuracy using a massive MIMO array achieves significantly better results compared to the state-of-the-art even for a smaller experimental data set.
|
|
