Robotic Monitoring of Habitats: The Natural Intelligence Approach

In this paper, we first discuss the challenges related to habitat monitoring and review possible robotic solutions. Then, we propose a framework to perform terrestrial habitat monitoring exploiting the mobility of legged robotic systems. The idea is to provide the robot with the Natural Intelligence introduced as the combination of the environment in which it moves, the intelligence embedded in the design of its body, and the algorithms composing its mind. This approach aims to solve the challenges of deploying robots in real natural environments, such as irregular and rough terrains, long-lasting operations, and unexpected collisions, with the final objective of assisting humans in assessing the habitat conservation status. Finally, we present examples of robotic monitoring of habitats in four different environments: forests, grasslands, dunes, and screes.

View this article on IEEE Xplore

Increasing Light Load Efficiency in Phase-Shifted, Variable Frequency Multiport Series Resonant Converters

Multiport power conversion topologies provide the capability of multiple independent converters with a single transformer having multiple windings (i.e., ports) potentially increasing power densities and enabling flexible (and bidirectional) power routing. In automotive onboard charger (OBC), the multiport approach combined with symmetrical series resonant circuits, the so-called multiport series resonant converter (MSRC), allows for a galvanic isolated connection between all ports: the grid-side converter (i.e., usually an AC/DC power factor correction (PFC) stage), vehicle’s main and the auxiliary low-voltage (LV) battery. The variation of the battery voltage significantly affects the MSRC operation, particularly for light loads at a low state-of-charge, and high losses can be experienced since zero-voltage-switching (ZVS) conditions are lost. In addition to the conventional control approach of the MSRC, where the power flow is set with a phase-shift between the individual full bridges or by changing the switching frequency, this paper proposes a novel and coordinated approach, including the manipulation of both and the additional modulation of the duty cycle as a function of the DC-link voltages, aiming to introduce a zero-voltage interval on the full bridge output voltages. A full mathematical description of the adopted converter topology is provided, including accurate simulation models that allow a comparison between the proposed duty cycle mode and the conventional control strategy. A detailed description of achieving ZVS within the connected full bridges is also included. Experimental results validate the proposal and demonstrate significant efficiency improvements compared to standard control approaches.

View this article on IEEE Xplore

Interference-Aware Intelligent Scheduling for Virtualized Private 5G Networks

Private Fifth Generation (5G) Networks can quickly scale coverage and capacity for diverse industry verticals by using the standardized 3rd Generation Partnership Project (3GPP) and Open Radio Access Network (O-RAN) interfaces that enable disaggregation, network function virtualization, and hardware accelerators. These private network architectures often rely on multi-cell deployments to meet the stringent reliability and latency requirements of industrial applications. One of the main challenges in these dense multi-cell deployments is the interference to/from adjacent cells, which causes packet errors due to the rapid variations from air-interface transmissions. One approach towards this problem would be to use conservative modulation and coding schemes (MCS) for enhanced reliability, but it would reduce spectral efficiency and network capacity. To unlock the utilization of higher efficiency schemes, in this paper, we present our proposed machine-learning (ML) based interference prediction technique that exploits channel state information (CSI) reported by 5G User Equipments (UEs). This method is integrated into an in-house developed Next Generation RAN (NG-RAN) research platform, enabling it to schedule transmissions over the dynamic air-interface in an intelligent way. By achieving higher spectral efficiency and reducing latency with fewer retransmissions, this allows the network to serve more devices efficiently for demanding use cases such as mission critical Internet-of-Things (IoT) and extended reality applications. In this work, we also demonstrate our over-the-air (OTA) testbed with 8 cells and 16 5G UEs in an Industrial IoT (IIoT) Factory Automation layout, where 5G UEs are connected to various industrial components like automatic guided vehicles (AGVs), supply units, robotics arms, cameras, etc. Our experimental results show that our proposed Interference-aware Intelligent Scheduling (IAIS) method can achieve up to 39% and 70% throughput gains in low and high interference scenarios, respectively, compared to a widely adopted link-adaptation scheduling approach.

View this article on IEEE Xplore

On-Road Trajectory Planning of Connected and Automated Vehicles in Complex Traffic Settings: A Hierarchical Framework of Trajectory Refinement

This paper presents a hierarchical framework for on-road trajectory planning in complex traffic environments. Firstly, the processing of sparse coarse trajectories involves the utilization of DP (Dynamic Programming) generation and interpolation techniques. Then, for the waypoints with collision risk in the smoothed trajectory, the spiral search method is used to find some safe alternate waypoints. The alternate waypoints and the previous ones without collision risk form the amended trajectory. Concurrently, safety tunnels are constructed along the amended trajectory for the ego vehicle. Furthermore, with the constraint conditions of vehicle kinematics model and safety tunnels, nonlinear program (NLP) optimization is carried out for the amended trajectory of ego vehicle to obtain smooth and safe trajectories. For typical cases, simulation experiments demonstrate that the ego vehicle can ensure collision safety in dynamic traffic scenarios, while maintaining smooth vehicle velocity and small jitter of the front wheel angle. The proposed trajectory planning framework provides a novel decision-making method for connected and automated vehicles (CAVs).

View this article on IEEE Xplore

Randomized Rank-Revealing QLP for Low-Rank Matrix Decomposition

The pivoted QLP decomposition is computed through two consecutive pivoted QR decompositions. It is an approximation to the computationally prohibitive singular value decomposition (SVD). This work is concerned with a partial QLP decomposition of matrices through the exploitation of random sampling. The method presented is tailored for low-rank matrices and called Randomized Unpivoted QLP (RU-QLP). Like pivoted QLP, RU-QLP is rank-revealing and yet it utilizes randomized column sampling and the unpivoted QR decomposition. The latter modifications allow RU-QLP to be highly scalable and parallelizable on advanced computational platforms. We provide an analysis for RU-QLP, thereby bringing insights into its characteristics and performance behavior. In particular, we derive bounds in terms of both spectral and Frobenius norms on: i) the rank-revealing property; ii) principal angles between approximate subspaces and exact singular subspaces and vectors; and iii) the errors of low-rank approximations. Effectiveness of the bounds is illustrated through numerical tests. We further use a modern, multicore machine equipped with a GPU to demonstrate the efficiency of RU-QLP. Our results show that compared to the randomized SVD, RU-QLP achieves a speedup of up to 7.1 and 8.5 times using the CPU and up to 2.3 and 5.8 times using the GPU for the decomposition of dense and sparse matrices, respectively.

View this article on IEEE Xplore

Scalable Empirical Dynamic Modeling With Parallel Computing and Approximate k-NN Search

Empirical Dynamic Modeling (EDM) is a mathematical framework for modeling and predicting non-linear time series data. Although EDM is increasingly adopted in various research fields, its application to large-scale data has been limited due to its high computational cost. This article presents kEDM, a high-performance implementation of EDM for analyzing large-scale time series datasets. kEDM adopts the Kokkos performance-portable programming model to efficiently run on both CPU and GPU while sharing a single code base. We also conduct hardware-specific optimization of performance-critical kernels. kEDM achieved up to 6.58× speedup in pairwise causal inference of real-world biology datasets compared to an existing EDM implementation. Furthermore, we integrate multiple approximate k-NN search algorithms into EDM to enable the analysis of extremely large datasets that were intractable with conventional EDM based on exhaustive k-NN search. EDM-based time series forecast enhanced with approximate k-NN search demonstrated up to 790× speedup compared to conventional Simplex projection with less than 1% increase in MAPE.

View this article on IEEE Xplore

Toward Precise Ambiguity-Aware Cross-Modality Global Self-Localization

There are significant advances in GNSS-free cross-modality self-localization of self-driving vehicles. Recent methods focus on learnable features for both cross-modal global localization via place recognition (PR) and local pose tracking, however they lack means of combining them in a complete localization pipeline. That is, a pose retrieved from PR has to be validated if it actually represents the true pose. Performing this validation without GNSS measurements makes the localization problem significantly more challenging. In this contribution, we propose a method to precisely localize the ego-vehicle in a high resolution map without GNSS prior. Furthermore, sensor and map data may be of different dimensions (2D / 3D) and modality, i.e. radar, lidar or aerial imagery. We initialize our system with multiple hypotheses retrieved from a PR method and infer the correct hypothesis over time. This multi-hypothesis approach is realized using a Gaussian sum filter which enables an efficient tracking of a low number of hypotheses and further facilitates the inference of our deep sensor-to-map matching network at arbitrarily distant regions simultaneously. We further propose a method to estimate the probability that none of the currently tracked hypotheses is correct. We achieve successful global localization in extensive experiments on the MulRan dataset, outperforming comparative methods even if none of the initial poses from PR was close to the true pose. Due to the flexibility of the approach, we can show state-of-the-art accuracy in lidar-to-aerial-imagery localization on a custom dataset using our pipeline with only minor modifications of the matching model.

View this article on IEEE Xplore

A Survey of Scheduling Algorithms for the Time-Aware Shaper in Time-Sensitive Networking (TSN)

Time-Sensitive Networking (TSN) is an enhancement of Ethernet which provides various mechanisms for real-time communication. Time-triggered (TT) traffic represents periodic data streams with strict real-time requirements. Amongst others, TSN supports scheduled transmission of TT streams, i.e., the transmission of their frames by end stations is coordinated in such a way that none or very little queuing delay occurs in intermediate nodes. TSN supports multiple priority queues per egress port. The TAS uses so-called gates to explicitly allow and block these queues for transmission on a short periodic timescale. The TAS is utilized to protect scheduled traffic from other traffic to minimize its queuing delay. In this work, we consider scheduling in TSN which comprises the computation of periodic transmission instants at end stations and the periodic opening and closing of queue gates. In this paper, we first give a brief overview of TSN features and standards. We state the TSN scheduling problem and explain common extensions which also include optimization problems. We review scheduling and optimization methods that have been used in this context. Then, the contribution of currently available research work is surveyed. We extract and compile optimization objectives, solved problem instances, and evaluation results. Research domains are identified, and specific contributions are analyzed. Finally, we discuss potential research directions and open problems.

View this article on IEEE Xplore

Object Detection Using Deep Learning, CNNs and Vision Transformers: A Review

Detecting objects remains one of computer vision and image understanding applications’ most fundamental and challenging aspects. Significant advances in object detection have been achieved through improved object representation and the use of deep neural network models. This paper examines more closely how object detection has evolved in the era of deep learning over the past years. We present a literature review on various state-of-the-art object detection algorithms and the underlying concepts behind these methods. We classify these methods into three main groups: anchor-based, anchor-free, and transformer-based detectors. Those approaches are distinct in the way they identify objects in the image. We discuss the insights behind these algorithms and experimental analyses to compare quality metrics, speed/accuracy tradeoffs, and training methodologies. The survey compares the major convolutional neural networks for object detection. It also covers the strengths and limitations of each object detector model and draws significant conclusions. We provide simple graphical illustrations summarising the development of object detection methods under deep learning. Finally, we identify where future research will be conducted.

View this article on IEEE Xplore

AMS Circuit Design Optimization Technique Based on ANN Regression Model With VAE Structure

The advanced design of an analog mixed-signal circuit is not simple enough to meet the requirements of the performance matrix as well as robust operations under process-voltage-temperature (PVT) changes. Even commercial products demand stringent specifications while maintaining the system’s performance. The main objectives of this study are to increase the efficiency of the design optimization process by configuring the design process in multiple regression modeling stages, to characterize our target circuit into a regression model including PVT variations, and to enable a search for co- optimum design points while simultaneously checking performance sensitivity. We used an artificial neural network (ANN) to develop a regression model and divided the ANN modeling process into coarse and fine simulation steps. In addition, we applied a variational autoencoder (VAE) structure to the ANN model to reduce the training error due to an insufficient input sample. According to the proposed algorithm, the AMS circuit designer can quickly search for the co- optimum point, which results in the best performance, while the least sensitive operation as the design process uses a regression model instead of launching heavy SPICE simulations. In this study, a voltage-controlled oscillator (VCO) is selected to prove the proposed algorithm. Under various design conditions (CMOS 180 nm, 65 nm, and 45 nm processes), we proceed with the proposed design flow to obtain the best performance score that can be evaluated by a figure-of-merit (FoM). As a result, the proposed regression model-based design flow achieves twice accurate results in comparison to that of the conventional single-step design flow.

View this article on IEEE Xplore