|
| Reference | Year | Category | Approach | Application | Design objective | Potentials for transmit power reduction |
|
| [82] | 2019 | Temporal data prediction | Auto regressive integrated moving average- (ARIMA-) based modeling for data prediction | Reduction in transmission energy for WSNs exploiting temporal data redundancy and data trend similarity between neighboring nodes | Energy efficiency | Yes |
| [83] | 2019 | Beamforming | Unsupervised deep neural network-based method for beamforming | Fast beamforming in MIMO systems | MIMO beamforming optimization | Yes |
| [84] | 2019 | Device identification | Robust principal component analysis- (RPCA-) based dimensional reduction and random forest-based classification for identification of rf characteristics unique to the transceiver. | Node identification | Security in node-to-node interaction | Depends on the rate of unauthorized access |
| [85] | 2019 | Compressive data reconstruction | Supervised learning approach for compressive data reconstruction | Structural health monitoring | Accuracy in signal reconstruction | Yes |
| [86] | 2018 | Transmission scheduling | Reinforcement-based method (Q-learning) for transmission through multiple channels | System throughput maximization in IOTs | Energy-efficient transmission | Yes |
| [87] | 2018 | Traffic classification | Unsupervised learning approach for fault detection | Machine health monitoring using WSNs | Predictive fault detection without data training | No |
| [88] | 2018 | Traffic classification | Survey machine learning-based approaches in traffic classification in software-defined WSNS. | For Qos implementation and unwanted traffic identification | _ | Depends on the application |
| [89] | 2018 | Topology control | Unsupervised learning approach for topology control | Ultradense WSNs | Energy management | Yes |
| [90] | 2018 | Spectrum monitoring | Convolutional neural network-based modulation and interference detection | Spectrum monitoring applications for interference and modulation management. | Detection performance | Yes |
| [91] | 2018 | Routing, traffic control | Deep convolutional neural network- (deep CNN-) based real-time learning for intelligent network traffic control | Traffic control optimization | Intelligent routing | Yes |
| [92] | 2018 | rf sensing | Deep learning framework for rf sensing using WiFi channel state information (CSI) | Fingerprinting, activity recognition, and vital sign monitoring | Prediction accuracy | Yes |
| [93] | 2018 | Location fingerprinting | Autoencoder-based machine learning for indoor localization | Indoor localization of mobile nodes | Localization accuracy | Yes |
| [94] | 2018 | Fault detection | Support vector machine classifier for WSN faulty sensors | Fault detection in WSNs | Fault detection with limited resources as in WSNs | Depends on the application |
| [79] | 2018 | Device identification | Deep neural network-based framework for device identification, exploiting die-to-die manufacturing variations | Node identification | Security in node-to-node interaction | Depends on the rate of unauthorized access |
| [95] | 2018 | Compressive data recovery | Secure data prediction using a time series trust model (TSTM) and a trust-based autoregressive (TAR) process | Secure data prediction in WSNs | Compressive sensing and resilience to node attack | Yes |
| [96] | 2017 | Mobile routing | Reinforcement-based method (Q-learning) for intelligent routing using route shortness and link stability | Real-time routing for mobile ad hoc networks (MANETS) | Routing optimization in MANETS | Yes |
| [97] | 2017 | Clusterhead selection | Naïve Bayes classifier for optimal cluster head selection | Optimal cluster head determination node for WSNs | Energy efficiency | Yes |
| [98] | 2016 | Topology control | Sensor data learning for spatial inference | Topology control in WSNS | Prediction accuracy while staying within the energy constraints | Yes |
| [99] | 2016 | Temporal data prediction | Temporal correlation-based dynamic forecasting model for data prediction | Temporal data prediction in WSNs | Prediction accuracy | Yes |
| [100] | 2016 | rf sensing | Deep learning-based device-free wireless localization and activity recognition (DFLAR) | Wireless sensing: localization and activity recognition | Sensing efficiency | Yes |
| [101] | 2016 | Prediction-based sensing and transmission reduction | Accurate data prediction while maintaining coverage requirements | Reduction in transmission energy for WSNs | Prediction accuracy and security for WSNs in cyber-physical systems | Yes |
| [102] | 2016 | Mobility prediction | Bayesian-based framework using observations in link duration to predict node velocity | Mobile WSNs | Mobility prediction accuracy vis-à-vis the resource constraints | Yes |
| [103] | 2016 | Location fingerprinting | Unsupervised learning algorithm for indoor localization based on received signal strength index (RSSI) | Indoor localization of mobile nodes | Localization accuracy | Yes |
| [104] | 2016 | Location fingerprinting | Unsupervised learning method for indoor localization | Indoor localization of mobile nodes | Unsupervised location fingerprinting | Yes |
| [105] | 2016 | Intrusion detection | Combines spectral clustering and deep neural network algorithms for intrusion detection | Malicious network traffic detection | Detection accuracy | Depends on the rate of unauthorized access |
| [106] | 2016 | Aggregation optimization | Hierarchical least-mean-square (HLMS) dual prediction algorithm for aggregation optimization | Reduction in transmissions energy for WSNs | Aggregation optimization | Yes |
| [107] | 2015 | Location estimation | Deep learning-based indoor fingerprinting system using channel state information (CSI) | Indoor location prediction | Indoor positioning accuracy | Yes |
| [108] | 2015 | Fault detection | Distributed Bayesian algorithm for distributed fault detection | Fault detection in WSNs | Data fault detection accuracy | Depends on the application |
| [109] | 2015 | Distributed data mining | Distributed data mining method based on deep neural network (DNN) | Decentralized applications | Data mining efficiency | Yes |
| [110] | 2015 | Aggregation optimization | Grey model (GM) and optimally pruned extreme learning machine- (OP-ELM-) based dual prediction scheme | Energy-constrained nodes | Transmission energy reduction | Yes |
| [111] | 2014 | Self-organization | Reinforcement learning approach-based energy cycle learning for on-demand throughput optimization | Dynamic power management in WSNs | On-demand throughput provisioning vis-à-vis resource constraints | Depends on the application |
|
