A Hierarchical Optimization Framework Using Deep Reinforcement Learning for Task-Driven Bandwidth Allocation in 5G Teleoperation

The evolution of 5G wireless technology has revolutionized connectivity, enabling a diverse range of applications. Among these are critical use cases such as real time teleoperation, which demands ultra reliable low latency communications (URLLC) to ensure precise and uninterrupted control, and enhanced mobile broadband (eMBB) services, which cater to data-intensive applications requiring high throughput and bandwidth. In our scenario, there are two queues, one for eMBB users and one for URLLC users. In teleoperation tasks, control commands are received in the URLLC queue, where communication delays occur. The dynamic index (DI) controls the service rate, affecting the telerobotic (URLLC) queue. A separate queue models eMBB data traffic. Both queues are managed through network slicing and application delay constraints, leading to a unified Lagrangian-based Lyapunov optimization for efficient resource allocation. We propose a DRL based hierarchical optimization framework that consists of two levels. At the first level, network optimization dynamically allocates resources for eMBB and URLLC users using a Lagrangian functional and an actor critic network to balance competing objectives. At the second level, control optimization finetunes the best gains for robots, ensuring stability and responsiveness in network conditions. This hierarchical approach enhances both communication and control processes, ensuring efficient resource utilization and optimized performance across the network.

Personalized Speech Emotion Recognition in Human-Robot Interaction using Vision Transformers

Emotions are an essential element in verbal communication, so understanding individuals' affect during a human-robot interaction (HRI) becomes imperative. This paper investigates the application of vision transformer models, namely ViT (Vision Transformers) and BEiT (BERT Pre-Training of Image Transformers) pipelines, for Speech Emotion Recognition (SER) in HRI. The focus is to generalize the SER models for individual speech characteristics by fine-tuning these models on benchmark datasets and exploiting ensemble methods. For this purpose, we collected audio data from different human subjects having pseudo-naturalistic conversations with the NAO robot. We then fine-tuned our ViT and BEiT-based models and tested these models on unseen speech samples from the participants. In the results, we show that fine-tuning vision transformers on benchmark datasets and and then using either these already fine-tuned models or ensembling ViT/BEiT models gets us the highest classification accuracies per individual when it comes to identifying four primary emotions from their speech: neutral, happy, sad, and angry, as compared to fine-tuning vanilla-ViTs or BEiTs.