Network slicing, a cornerstone technology for future networks, enables the creation of customized virtual networks on a shared physical infrastructure. This fosters innovation and agility by providing dedicated resources tailored to specific applications. However, current orchestration and management approaches face limitations in handling the complexity of new service demands within multi-administrative domain environments. This paper proposes a future vision for network slicing powered by Large Language Models (LLMs) and multi-agent systems, offering a framework that can be integrated with existing Management and Orchestration (MANO) frameworks. This framework leverages LLMs to translate user intent into technical requirements, map network functions to infrastructure, and manage the entire slice lifecycle, while multi-agent systems facilitate collaboration across different administrative domains. We also discuss the challenges associated with implementing this framework and potential solutions to mitigate them.

In today’s fast-paced business environment, the abundance of data is both a boon and a challenge for companies. The sheer volume of information generated internally and acquired from vendors is immense, and technologies like Tableau have revolutionized data analytics and visualizations. However, the key to success lies in delivering the right information to the sales team promptly, enabling them to make informed decisions in real time. This article explores how harnessing Tableau subscriptions, user filters, and machine learning-driven recommendations can efficiently provide real-time, personalized insights directly to sales teams via email and phone. The goal is to transform the conventional approach, where end-users actively seek out information, into a proactive model where insights are delivered seamlessly, optimizing the efficiency of the sales team. The traditional paradigm often sees sales professionals spending valuable time searching for relevant insights instead of focusing on selling. By proactively delivering hidden trends and actionable insights, businesses can enhance the productivity of their sales teams, allowing them to allocate more time to revenue-generating activities. One of the key features of this approach is its emphasis on user-friendly tools like Tableau, demonstrating that Next Best Action can be seamlessly integrated within the Tableau ecosystem without the need for additional tools such as CRM systems. This not only streamlines the workflow for sales teams but also highlights the versatility and power of Tableau as an all-encompassing analytics platform. Furthermore, the solution presented in this article leverages the convenience of delivering insights directly to sales representatives through email and text messages. This eliminates the need for complex system logins, making the process more accessible and user-friendly. By providing insights directly to the fingertips of sales teams, this approach ensures that critical information is readily available, facilitating quick decision-making and boosting overall engagement.

The traditional approach to applying safety limits in electro-mechanical systems across various industries, including automated vehicles, robotics, and aerospace, involves hard-coding control and safety limits into production firmware, which remains fixed throughout the product life cycle. However, with the evolving needs of automated systems like automated vehicles and robots, this approach falls short in addressing all use cases and scenarios to ensure safe operation. Particularly for data-driven machine learning applications that continuously evolve, there’s a need for a more flexible and adaptable safety limits application strategy based on different Operational Design Domains (ODDs) and scenarios. Our ITSC conference paper [(1)] introduced the Dynamic Control Limits Application (DCLA) strategy, supporting the flexible application of diverse limits profiles based on dynamic scenario parameters across different layers of the Autonomy software stack. This paper extends the DCLA strategy by outlining a methodology for safety limits application based on ODD elements, scenario identification, and classification using Decision Making Engines. It also utilizes a layered architecture and cloud infrastructure based on Vehicle-to-Infrastructure (V2I) technology to store scenarios and limits mapping as a ground truth or backup mechanism for the Decision Making Engine. Additionally, the paper focuses on providing a comprehensive list of scenarios and an experimental dataset covering maximum ODD elements, along with multiple tables of safety limits to create various application profiles. These profiles, based on perceived scenario parameters, can be systematically applied or trained on Decision Making algorithms, offering a scalable solution for future automated vehicles and systems up to Level 5 Autonomy within the industry.

Integrating radar and communication systems for economical use of hardware and spectrum resources is projected to be a crucial aspect of sixth-generation (6G) systems, leading to extensive research in the integrated sensing and communication (ISAC) area. In this article, we propose a radar detector structure for a multiple-input multiple-output (MIMO) ISAC system using orthogonal frequency-division multiplexing (OFDM) modulated waveforms. These waveforms are utilized to communicate with downlink (DL) users while receiving echoes from targets and clutter, in addition to separate OFDM-modulated uplink (UL) communication waveforms. The transmitter (Tx) and receiver (Rx) employ hybrid beamformers, with analog beamformers precisely designed to cover flat angular sectors. Tx and Rx beams are directed towards DL and UL users, respectively, with the possibility of overlapping or separate angular sectors. We introduce a Doppler-aware Code Bank (DACB) as the initial processing stage, thoroughly investigating the effects of Doppler mismatch. Following DACB, a sub-optimal sequential angle-range processing (SARP) method is proposed to maximize the output signal-to-interference-plus-noise ratio (SINR) while maintaining feasible processing. A two-variant detector scheme is proposed to address this processing's suboptimality. Four different detectors with varying complexities, including a fully adaptive detector, are introduced. The potential beam-squint effect due to increased bandwidth is also considered, and a subband approach is proposed to mitigate these effects. Simulation results for all four detectors, as well as a conventional 3-dimensional periodogram detector commonly used as a benchmark in the literature, are provided. The results demonstrate that the proposed detectors can significantly enhance SINR and the probability of detection, particularly when accounting for the coupling between the clutter Doppler and Tx OFDM symbols.

The sustainability of Machine Learning-Enabled Systems (MLS), particularly with regard to energy efficiency, is an important challenge in their development and deployment. Self-adaptation techniques, recognized for their potential in energy savings within software systems, have yet to be extensively explored in Machine Learning-Enabled Systems (MLS), where runtime uncertainties can significantly impact model performance and energy consumption. This variability, alongside the fluctuating energy demands of ML models during operation, necessitates a dynamic approach. Addressing these challenges, we introduce EcoMLS approach, which leverages the Machine Learning Model Balancer concept to enhance the sustainability of MLS through runtime ML model switching. By adapting to monitored runtime conditions, EcoMLS optimally balances energy consumption with model confidence, demonstrating a significant advancement towards sustainable, energy-efficient machine learning solutions. Through an object detection exemplar, we illustrate the application of EcoMLS, showcasing its ability to reduce energy consumption while maintaining high model accuracy throughout its use. This research underscores the feasibility of enhancing MLS sustainability through intelligent runtime adaptations, contributing a valuable perspective to the ongoing discourse on energy-efficient machine learning.

Changes in gait are associated with an increased risk of falling and may indicate the presence of movement disorders in connection with a neurological disease or age-related weakness. Continuous monitoring, based on inertial measurement unit (IMU) sensor data can be used to estimate gait parameters that indicate changes in gait. The monitoring from a waist-level IMU sensor is applicable for the assessment of such data, as it can be easily worn either as specific sensor-integrated belts or being suitable as a smartphone application. However, there is also the challenge that data from a waist-level mounted sensor provides a weaker representation of gait compared to signals from sensors placed closer to the legs and feet as sources of motion. Our work examines how well gait parameters can be estimated from the data of a waist worn IMU sensor. The approach employs the processing of the IMUs triaxial accelerometer and gyroscope data to detect gait events and estimate spatio-temporal gait parameters. In addition to the gait events heel strike (HS) and toe off (TO), it estimates the gait parameters step length, cadence, speed, stance time, swing time, and stride time. The results are compared to data measured by a GAITRite ® system as a reference. The present study investigates and compares the accuracy of two different approaches for gait event detection: a rule-based method and a machine learning (ML) approach based on a sequence-to-sequence convolutional neural network (CNN) architecture. The efficacy of these approaches in accurately determining gait parameters is evaluated using a dataset consisting of 21,193 recorded steps from 70 subjects, who performed 4,590 walks, each with a distance of approximately 4 meters. The ML-based algorithm performs better than the rule-based achieving an accuracy of over 99% in detecting gait events and estimating step lengths with a mean error of-0.12±4.49 cm.

The conventional use of the Retrieval-Augmented Generation (RAG) architecture has proven effective for retrieving information from diverse documents. However, challenges arise in handling complex table queries, especially within PDF documents containing intricate tabular structures. This research introduces an innovative approach to enhance the accuracy of complex table queries in RAG-based systems. Our methodology involves storing PDFs in the retrieval database and extracting tabular content separately. The extracted tables undergo a process of context enrichment, concatenating headers with corresponding values. To ensure a comprehensive understanding of the enriched data, we employ a fine-tuned version of the Llama-2-chat language model for summarisation within the RAG architecture. Furthermore, we augment the tabular data with contextual sense using the ChatGPT 3.5 API through a one-shot prompt. This enriched data is then fed into the retrieval database alongside other PDFs. Our approach aims to significantly improve the precision of complex table queries, offering a promising solution to a longstanding challenge in information retrieval.

An optimal reference voltage saturation (ORVS) for nonlinear current control of synchronous machine drives is presented. It is integrated into a nonlinear current control system based on input/output (I/O) linearization and proportional-integral (PI) controllers with anti-windup. The ORVS is derived from an optimization problem that can be solved analytically. The deviation between the desired current dynamics obtained by the proportional-integral (PI) current control system if the voltage constraints were not present and the feasible current dynamics due to the saturated voltage is minimized. Finally, the proposed ORVS is compared to (i) classical (vector-valued) reference voltage saturation and (ii) the absence of any saturation and/or anti-windup. The simulative comparison is done for a highly nonlinear, anisotropic reluctance synchronous machine (RSM) whose parameters and look-up tables were obtained from a real machine. The simulation results show that the proposed ORVS outperforms the other two approaches in terms of current tracking accuracy and faster current dynamics.

A document by surendra kumar. Click on the document to view its contents.

A simple two-dimensional model of an infinite array of circular conductors is presented. The model is developed in two variants, a thin model and a thick model. It allows the evaluation of the shielding effectiveness of the array when the wires distances their and radius changes. The shielding effect is expressed through the transmission coefficient of a plane wave normally impinging on the screen. The model is limited to a plane wave impinging normally on the wire's plane and linearly polarized along the wires.

Trajectory planning for automated vehicles in traffic has been a challenging task and a hot topic in recent research. The need for flexibility, transparency, interpretability and predictability poses challenges in deploying data-driven approaches in this safety-critical application. This paper proposes DeepGame-TP, a game-theoretical trajectory planner that uses deep learning to model each agent's cost function and adjust it based on observed behavior. In particular, a LSTM network predicts each agent's desired speed, forming a penalizing term that reflects aggressiveness in the cost function. Experiments demonstrated significant advantages of this innovative framework, highlighting the adaptability of DeepGame-TP in intersection, overtaking, and merging scenarios. It effectively avoids dangerous situations that could arise from incorrect cost function estimates. The approach is suitable for real-time applications, solving the Generalized Nash Equilibrium Problem (GNEP) in scenarios with up to three vehicles in under 100 milliseconds on average.

This paper proposes an "Ice Cube Algorithm" for adaptive time stretching in video editing after AI translation. The algorithm tackles the challenge of maintaining audiovisual synchronization despite variations in translated text length. It dynamically adjusts audio speed to ensure consistent video duration, offering an efficient solution for multilingual content creators.

We consider a cellular network equipped with a retransmission mechanism where the signal transmissions in a single cell depends on interference with simultaneous traffic in surrounding cells. In this framework where failed signals are either retransmitted or lost we study the channel capacity performance of a single-tier model with downlink or uplink traffic. For this purpose, a tractable model that allows for a precise theoretical analysis of coverage probability and coverage rate is developed further. Specifically, we obtain the Shannon capacity in the model and introduce relevant performance measures to guide in the identification of those systems which in a precise sense are able to process all incoming work. To emphasize the generic patterns that arise we extend and simplify the results under a scaling regime of balanced densification, which highlights that performance essentially falls into three categories, for pure-loss, buffered, and no-loss systems.

Friction phenomena may strongly deteriorate performance and reliability of force-controlled systems. Although different solutions exist, ensuring accurate and robustly stable friction compensation is still a challenging task due to model inaccuracies and parameter identification issues. This paper proposes a novel friction observer architecture designed to achieve accurate friction compensation in force control applications, without relying on a specific friction model. Since passivity is a fundamental requirement for interaction control, a passivity analysis is reported showing that the proposed solution preserves passivity at the environment, control and friction ports. This implies that the proposed observer can provide a passive friction compensation action for any friction dynamics and can be combined with any passive controller while guaranteeing a stable interaction with any passive environment. The proposed approach is validated through simulation and experimental trials, considering different interacting environments and diverse friction phenomena affecting the actuator dynamics. A main impact of the proposed approach is expected on collaborative and assistive robotics, including active industrial exoskeletons where interaction reliability and transparency are of paramount importance.

In recent years, the concept of the digital divide has gained significant attention across various sectors, including agriculture. The digital divide refers to the gap between those who have access to digital technologies and those who do not, and this divide is particularly pronounced in rural communities. This paper explores the digital divide in agriculture, identifying its key factors and discussing strategies to bridge this gap, thereby ensuring equitable access to technology in rural areas.

Spinal cord injuries (SCI) substantially affect sensory, motor, and autonomous functions below the level of injury, reducing the independence and quality of life for affected individuals. Specifically, people with SCI between C5 and C7 cervical levels encounter limitations in voluntary finger and wrist flexion, reducing grasp capability. Compensatory strategies like tenodesis grasp, whereby wrist extension passively closes the fingers, remain; this is effective for lighter objects but insufficient for heavier ones. Typically, wearable assistive exoskeletons are designed to actuate a person's fingers, however, such devices are sensitive to anatomical variability, such as hand size and joint contractures. Addressing this challenge, here we present a new version of the Dorsal Grasper, a wearable device designed for those with voluntary wrist extension, providing human-robot collaborative grasping capabilities with underactuated supernumerary fingers on the back of the hand. We show that the Dorsal Grasper expands the graspable workspace and reduces trunk motion, especially in situations where the use of a wheelchair restricts the individual's posture. Our experiments with SCI participants demonstrate the Dorsal Grasper 's potential as a versatile assistive solution for enhancing grasping capability in individuals with distinct SCI profiles.

Deploying convolution-based algorithms into SRAM computing-in-memory (CIM) systems faces various challenges, such as operator incompatibility and intrinsic non-ideal error. This paper proposes a compilation framework to address this issue. Efficient weight mapping strategies are introduced to improve the utilization of SRAM-CIM macro. The intrinsic non-ideal errors of SRAM-CIM macro are also taken into consideration, and two efficient error correction schemes are proposed, which include calibration of computation voltage linear error (CCVLE) and the mitigation of analog-to-digital quantization error (MAQE). In addition, bit-width flexibility and signed-unsigned reconfigurability are also supported to facilitate the deployment of various convolution-based algorithms. ResNet18, finite impulse response (FIR) filtering, and Gaussian image filtering are deployed into a multi-macro SRAM-CIM system. These algorithms serve as deployment representatives of convolutional neural network (CNN), digital signal processing (DSP), and digital image processing (DIP), respectively. The results show that the introduced weight mapping strategies improve the macro utilization by 63.29% and 21.10% for two types of frequently used convolution layers compared to the traditional strategy. Moreover, the proposed error correction schemes achieve similar algorithm accuracy to the floating-point results, and the deployment result of ResNet18 achieves 66.3%~70.1% top-1 classification accuracy evaluated on the ImageNet dataset with different throughput tradeoffs.

Recent findings reveal that a simple one-layer linear model can significantly outperform transformer-based models. To further enhance time series forecasting performance, we introduce a feedforward neural network (FNN) -based deep learning model, which surpassed most existing models. The novel FNN-based model presents three innovations: input and output structure, FNN division and combination, and an FNN sharing mechanism. The model utilizes three inputs: historical values, historical timestamps, and future timestamps, in order to predict the future demand as its output. FNN division means different inputs corresponding to different FNNs, and the outputs of these FNNs are combined to form the final prediction. The FNN sharing mechanism involves mapping each column (or row) of the matrix inputs to the column (or row) using a common FNN, termed FNN kernel. This FNN neural network framework reduces computational complexity and the number of weights significantly while still providing interpretability and strong performance. Our comparative experiments demonstrate that all three models outperform existing models on two real-world datasets.