MLRS-RL: An Energy-Efficient Multilevel Routing Strategy Based on Reinforcement Learning in Multimodal UWSNs

Abstract

The limited energy and computing resources of unmanned aerial vehicles (UAVs) hinder the application of aerial artificial intelligence. The utilization of split inference in UAVs garners significant attention due to its effectiveness in mitigating computing and energy requirements. However, achieving energyefficient split inference in UAVs remains complex considering of various crucial parameters such as energy level and delay constraints, especially involving multiple tasks. In this paper, we present a two-timescale approach for energy minimization in split inference, where discrete and continuous variables are segregated into two timescales to reduce the size of action space and computational complexity. This segregation enables the utilization of tiny reinforcement learning (TRL) for selecting discrete transmission modes for sequential tasks. Moreover, optimization programming (OP) is embedded between TRL's output and reward function to optimize the continuous transmit power. Specifically, we replace the optimization of transmit power with that of transmission time to decrease the computational complexity of OP since we reveal that energy consumption monotonically decreases with increasing transmission time. The replacement significantly reduces the feasible region and enables a fast solution according to the closed-form expression for optimal transmit power. Simulation results show that the proposed algorithm can achieve a higher probability of successful task completion with lower energy consumption.

SALLoc: An Accurate Target Localization in WiFi-Enabled Indoor Environments via SAE-ALSTM

“…Quite a few DL-based indoor localization algorithms have been investigated for various applications. Some are device-free [26,[40][41][42][43][44] whereas others device-based [12,18,19,28,[45][46][47][48] systems. The author in [26] explored the use of DL techniques for device-free target tracking and localization in indoor environments.…”

“…The approach assumes the trajectory of users to always be in a straight line. Study in [46] presents an indoor localization based on GrowNet and LSTM network in multi-building environment to improve accuracy and robustness of WiFi fingerprinting technique. Ensemble model was used to extract mapping correlation between the uneven RSSI samples and building/floor classification while LSTM was used for location estimation.…”

“…where 𝑥 𝑖 ,𝑦 𝑖 and 𝑥 𝑖 ̂,𝑦 𝑖 ̂ are the real and estimated longitude and latitude coordinates of the 𝑖-th RSSI fingerprint respectively [46]. 𝑁 𝑅 is the number of RSSI samples in the dataset for which 1 < 𝑖 < 𝑁 𝑅 .…”

“…RSSI data preprocessing have been shown in related studies to have a substantial influence on performance of indoor localization models [14,15,19,28,46,47]. To ensure optimal performance, our proposed DL framework entail detailed preparation, involving a series of steps that were executed in multiple folds.…”

“…The dropout layer randomly set a fraction of the activation of a layer to zero (0) during training. This forces the model to learn multiple representations of the data, reducing its reliance on any single feature or group of features [46,54,55]. With the optimized SAE and ALSTM models, we evaluate our proposed SALLoc model with different dropout rates as presented in Fig.…”