Reinforced-LSTM Trajectory Prediction-Driven Dynamic Service Migration: A Case Study

Zhao, Zhongliang; Emami, Negar; Santos, Hugo; Pacheco, Lucas; Karimzadeh, Mostafa; Braun, Torsten; Braud, Arnaud; Radier, Benoît; Tamagnan, Philippe

doi:10.1109/tnse.2022.3169786

Cited by 4 publications

(3 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In recent years, data-driven social mobility predictors are gaining popularity compared to the previously proposed Social-force models, which use simple repulsive and attraction forces [7]. The vast majority of modern human-trajectory predictors are based on deep learning models, such as RNNs, Long Short-Term Memorys (LSTMs), Convolutional Neural Networks (CNNs), and attention-based neural networks, such as Transformers, which require less computation and achieve higher prediction accuracy compared to social-force models due to their better modeling of sequential patterns [1], [8], [9]. Instead of modeling kinetic forces and energy potentials as in social-force models, social-pooling [2], [3], attention [10], [5],…”

Section: Related Workmentioning

confidence: 99%

“…Hence, we consider only the information about which base station the user is connected to. For the large-scale scenario, we use a private cellular network management dataset provided by Orange telecommunication S.A., France [1]. This dataset contains the timestamps and the connected base station IDs for each of the 1.3 million users that move near a district of Paris between July and September 2019.…”

Section: A Experimental Setupmentioning

confidence: 99%

“…Accurate user mobility prediction can improve Quality of Service (QoS) of network applications through adaptive and anticipatory network management, such as proactive caching, resource allocation, service migration, and handover management [1]. On the other hand, trajectory prediction is of great importance in urban safety, autonomous driving, and robot-to-human interacting systems.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

INTRAFORCE: Intra-Cluster Reinforced Social Transformer for Trajectory Prediction

Emami

Maio

Braun

2022

2022 18th International Conference on Wireless and Mobile Computing, Networking and Communications (WiMob)

Self Cite

View full text Add to dashboard Cite

Predicting mobile users' trajectories accurately is essential for improving the performance of wireless networks and autonomous systems. In this paper, we tackle the problem of trajectory prediction in a multi-agent scenario where the social interaction among users is taken into consideration. We propose Intra-Cluster Reinforced Social Transformer (INTRAFORCE), a novel system to design and train Social-Transformer neural networks that learn the spatio-temporal interactions among neighboring mobile users and predict their joint future trajectories. Unlike state-of-the-art social-aware trajectory predictors that either miss the large-distance interactions or are computationally expensive due to the pooling of all users' interactions, INTRAFORCE clusters users with similar trajectories and learns their interactions. INTRAFORCE performs Neural Architecture Search to optimize each transformer's architecture to fit each cluster's user mobility features using Reinforcement Learning. Through experimental validation, we show that INTRAFORCE outperforms several state-of-the-art trajectory predictors on five widely used smallscale pedestrian mobility datasets and one large-scale privacyoriented cellular mobility dataset by achieving lower prediction error, training time, and computational complexity.

show abstract