Search results for: pink noise

Authors: Rahil Bahrami, Kaveh Ashenayi

Abstract:

This paper presents a systematic classification and comparative analysis of non-isolated bi-directional DC-DC converters. The increasing demand for efficient energy conversion in diverse applications has spurred the development of various converter topologies. In this study, we categorize bi-directional converters into three distinct classes: Inverting, Non-Inverting, and Interleaved. Each category is characterized by its unique operational characteristics and benefits. Furthermore, a practical comparison is conducted by evaluating the results of simulation of each bi-directional converter. BDCs can be classified into isolated and non-isolated topologies. Non-isolated converters share a common ground between input and output, making them suitable for applications with minimal voltage change. They are easy to integrate, lightweight, and cost-effective but have limitations like limited voltage gain, switching losses, and no protection against high voltages. Isolated converters use transformers to separate input and output, offering safety benefits, high voltage gain, and noise reduction. They are larger and more costly but are essential for automotive designs where safety is crucial. The paper focuses on non-isolated systems.The paper discusses the classification of non-isolated bidirectional converters based on several criteria. Common factors used for classification include topology, voltage conversion, control strategy, power capacity, voltage range, and application. These factors serve as a foundation for categorizing converters, although the specific scheme might vary depending on contextual, application, or system-specific requirements. The paper presents a three-category classification for non-isolated bi-directional DC-DC converters: inverting, non-inverting, and interleaved. In the inverting category, converters produce an output voltage with reversed polarity compared to the input voltage, achieved through specific circuit configurations and control strategies. This is valuable in applications such as motor control and grid-tied solar systems. The non-inverting category consists of converters maintaining the same voltage polarity, useful in scenarios like battery equalization. Lastly, the interleaved category employs parallel converter stages to enhance power delivery and reduce current ripple. This classification framework enhances comprehension and analysis of non-isolated bi-directional DC-DC converters. The findings contribute to a deeper understanding of the trade-offs and merits associated with different converter types. As a result, this work aids researchers, practitioners, and engineers in selecting appropriate bi-directional converter solutions for specific energy conversion requirements. The proposed classification framework and experimental assessment collectively enhance the comprehension of non-isolated bi-directional DC-DC converters, fostering advancements in efficient power management and utilization.The simulation process involves the utilization of PSIM to model and simulate non-isolated bi-directional converter from both inverted and non-inverted category. The aim is to conduct a comprehensive comparative analysis of these converters, considering key performance indicators such as rise time, efficiency, ripple factor, and maximum error. This systematic evaluation provides valuable insights into the dynamic response, energy efficiency, output stability, and overall precision of the converters. The results of this comparison facilitate informed decision-making and potential optimizations, ensuring that the chosen converter configuration aligns effectively with the designated operational criteria and performance goals.

Keywords: bi-directional, DC-DC converter, non-isolated, energy conversion

Procedia PDF Downloads 101

Authors: Ching-Sung Lee, Wei-Chou Hsu, Han-Yin Liu, Hung-Hsi Huang, Si-Fu Chen, Yun-Jung Yang, Bo-Chun Chiang, Yu-Chuang Chen, Shen-Tin Yang

Abstract:

AlGaN/GaN high electron mobility transistors (HEMTs) have been intensively studied due to their intrinsic advantages of high breakdown electric field, high electron saturation velocity, and excellent chemical stability. They are also suitable for ultra-violet (UV) photodetection due to the corresponding wavelengths of GaN bandgap. To improve the optical responsivity by decreasing the dark current due to gate leakage problems and limited Schottky barrier heights in GaN-based HEMT devices, various metal-oxide-semiconductor HEMTs (MOS-HEMTs) have been devised by using atomic layer deposition (ALD), molecular beam epitaxy (MBE), metal-organic chemical vapor deposition (MOCVD), liquid phase deposition (LPD), and RF sputtering. The gate dielectrics include MgO, HfO2, Al2O3, La2O3, and TiO2. In order to provide complementary circuit operation, enhancement-mode (E-mode) devices have been lately studied using techniques of fluorine treatment, p-type capper, piezoneutralization layer, and MOS-gate structure. This work reports an Al2O3-dielectric Al0.25Ga0.75N/GaN E-mode MOS-HEMT design by using a cost-effective ozone water oxidization technique. The present ozone oxidization method advantages of low cost processing facility, processing simplicity, compatibility to device fabrication, and room-temperature operation under atmospheric pressure. It can further reduce the gate-to-channel distance and improve the transocnductance (gm) gain for a specific oxide thickness, since the formation of the Al2O3 will consume part of the AlGaN barrier at the same time. The epitaxial structure of the studied devices was grown by using the MOCVD technique. On a Si substrate, the layer structures include a 3.9 m C-doped GaN buffer, a 300 nm GaN channel layer, and a 5 nm Al0.25Ga0.75N barrier layer. Mesa etching was performed to provide electrical isolation by using an inductively coupled-plasma reactive ion etcher (ICP-RIE). Ti/Al/Au were thermally evaporated and annealed to form the source and drain ohmic contacts. The device was immersed into the H2O2 solution pumped with ozone gas generated by using an OW-K2 ozone generator. Ni/Au were deposited as the gate electrode to complete device fabrication of MOS-HEMT. The formed Al2O3 oxide thickness 7 nm and the remained AlGaN barrier thickness is 2 nm. A reference HEMT device has also been fabricated in comparison on the same epitaxial structure. The gate dimensions are 1.2 × 100 µm 2 with a source-to-drain spacing of 5 μm for both devices. The dielectric constant (k) of Al2O3 was characterized to be 9.2 by using C-V measurement. Reduced interface state density after oxidization has been verified by the low-frequency noise spectra, Hooge coefficients, and pulse I-V measurement. Improved device characteristics at temperatures of 300 K-450 K have been achieved for the present MOS-HEMT design. Consequently, Al2O3-dielectric Al0.25Ga0.75N/GaN E-mode MOS-HEMTs by using the ozone water oxidization method are reported. In comparison with a conventional Schottky-gate HEMT, the MOS-HEMT design has demonstrated excellent enhancements of 138% (176%) in gm, max, 118% (139%) in IDS, max, 53% (62%) in BVGD, 3 (2)-order reduction in IG leakage at VGD = -60 V at 300 (450) K. This work is promising for millimeter-wave integrated circuit (MMIC) and three-terminal active UV photodetector applications.

Keywords: MOS-HEMT, enhancement mode, AlGaN/GaN, passivation, ozone water oxidation, gate leakage

Procedia PDF Downloads 263

Authors: Zdravko Trivic, Kelvin E. Y. Low, Darko Radovic, Raymond Lucas

Abstract:

Our understanding and experience of the built environment are primarily shaped by multi‐sensory, emotional and symbolic modes of exchange with spaces. Associated sensory and cognitive declines that come with ageing substantially affect the overall quality of life of the elderly citizens and the ways they perceive and use urban environment. Reduced mobility and increased risk of falls, problems with spatial orientation and communication, lower confidence and independence levels, decreased willingness to go out and social withdrawal are some of the major consequences of sensory declines that challenge almost all segments of the seniors’ everyday living. However, contemporary urban environments are often either sensory overwhelming or depleting, resulting in physical, mental and emotional stress. Moreover, the design and planning of housing neighbourhoods hardly go beyond the passive 'do-no-harm' and universal design principles, and the limited provision of often non-integrated eldercare and inter-generational facilities. This paper explores and discusses the largely neglected relationships between the 'hard' and 'soft' aspects of housing neighbourhoods and urban experience, focusing on seniors’ perception and multi-sensory experience as vehicles for design and planning of high-density housing neighbourhoods that are inclusive and empathetic yet build senior residents’ physical and mental abilities at different stages of ageing. The paper outlines methods and key findings from research conducted in two high-density housing neighbourhoods in Singapore with aims to capture and evaluate multi-sensorial qualities of two neighbourhoods from the perspective of senior residents. Research methods employed included: on-site sensory recordings of 'objective' quantitative sensory data (air temperature and humidity, sound level and luminance) using multi-function environment meter, spatial mapping of patterns of elderly users’ transient and stationary activity, socio-sensory perception surveys and sensorial journeys with local residents using eye-tracking glasses, and supplemented by walk-along or post-walk interviews. The paper develops a multi-sensory framework to synthetize, cross-reference, and visualise the activity and spatio-sensory rhythms and patterns and distill key issues pertinent to ageing-friendly and health-supportive neighbourhood design. Key findings show senior residents’ concerns with walkability, safety, and wayfinding, overall aesthetic qualities, cleanliness, smell, noise, and crowdedness in their neighbourhoods, as well as the lack of design support for all-day use in the context of Singaporean tropical climate and for inter-generational social interaction. The (ongoing) analysis of eye-tracking data reveals the spatial elements of senior residents’ look at and interact with the most frequently, with the visual range often directed towards the ground. With capacities to meaningfully combine quantitative and qualitative, measured and experienced sensory data, multi-sensory framework shows to be fruitful for distilling key design opportunities based on often ignored aspects of subjective and often taken-for-granted interactions with the familiar outdoor environment. It offers an alternative way of leveraging the potentials of housing neighbourhoods to take a more active role in enabling healthful living at all stages of ageing.

Keywords: ageing-friendly neighbourhoods, eye-tracking, high-density environment, multi-sensory approach, perception

Procedia PDF Downloads 156

Authors: Mohamed Gazzeh, Bradley Null, Fethi Tlili, Hichem Besbes

Abstract:

The burgeoning global ownership of personal vehicles has posed a significant strain on urban infrastructure, notably parking facilities, leading to traffic congestion and environmental concerns. Effective parking management systems (PMS) are indispensable for optimizing urban traffic flow and reducing emissions. The most commonly deployed systems nowadays rely on computer vision technology. This paper explores the integration of radar sensors and simulation in the context of smart parking management. We concentrate on radar sensors due to their versatility and utility in automotive applications, which extends to PMS. Additionally, radar sensors play a crucial role in driver assistance systems and autonomous vehicle development. However, the resource-intensive nature of radar data collection for algorithm development and testing necessitates innovative solutions. Simulation, particularly the monoDrive simulator, an internal development tool used by NI the Test and Measurement division of Emerson, offers a practical means to overcome this challenge. The primary objectives of this study encompass simulating radar sensors to generate a substantial dataset for algorithm development, testing, and, critically, assessing the transferability of models between simulated and real radar data. We focus on occupancy detection in parking as a practical use case, categorizing each parking space as vacant or occupied. The simulation approach using monoDrive enables algorithm validation and reliability assessment for virtual radar sensors. It meticulously designed various parking scenarios, involving manual measurements of parking spot coordinates, orientations, and the utilization of TI AWR1843 radar. To create a diverse dataset, we generated 4950 scenarios, comprising a total of 455,400 parking spots. This extensive dataset encompasses radar configuration details, ground truth occupancy information, radar detections, and associated object attributes such as range, azimuth, elevation, radar cross-section, and velocity data. The paper also addresses the intricacies and challenges of real-world radar data collection, highlighting the advantages of simulation in producing radar data for parking lot applications. We developed classification models based on Support Vector Machines (SVM) and Density-Based Spatial Clustering of Applications with Noise (DBSCAN), exclusively trained and evaluated on simulated data. Subsequently, we applied these models to real-world data, comparing their performance against the monoDrive dataset. The study demonstrates the feasibility of transferring models from a simulated environment to real-world applications, achieving an impressive accuracy score of 92% using only one radar sensor. This finding underscores the potential of radar sensors and simulation in the development of smart parking management systems, offering significant benefits for improving urban mobility and reducing environmental impact. The integration of radar sensors and simulation represents a promising avenue for enhancing smart parking management systems, addressing the challenges posed by the exponential growth in personal vehicle ownership. This research contributes valuable insights into the practicality of using simulated radar data in real-world applications and underscores the role of radar technology in advancing urban sustainability.

Keywords: autonomous vehicle simulator, FMCW radar sensors, occupancy detection, smart parking management, transferability of models

Procedia PDF Downloads 82

Authors: Subhash Nerella, Ziyuan Guan, Azra Bihorac, Parisa Rashidi

Abstract:

Intensive care units (ICUs) provide care to critically ill patients in severe and life-threatening conditions. However, patient monitoring in the ICU is limited by the time and resource constraints imposed on healthcare providers. Many critical care indices such as mobility are still manually assessed, which can be subjective, prone to human errors, and lack granularity. Other important aspects, such as environmental factors, are not monitored at all. For example, critically ill patients often experience circadian disruptions due to the absence of effective environmental “timekeepers” such as the light/dark cycle and the systemic effect of acute illness on chronobiologic markers. Although the occurrence of delirium is associated with circadian disruption risk factors, these factors are not routinely monitored in the ICU. Hence, there is a critical unmet need to develop systems for precise and real-time assessment through novel enabling technologies. We have developed the mobility and circadian disruption quantification system (Mobi-DiQ) by augmenting biomarker and clinical data with pervasive sensing data to generate mobility and circadian cues related to mobility, nightly disruptions, and light and noise exposure. We hypothesize that Mobi-DiQ can provide accurate mobility and circadian cues that correlate with bedside clinical mobility assessments and circadian biomarkers, ultimately important for delirium risk assessment and prevention. The collected multimodal dataset consists of depth images, Electromyography (EMG) data, patient extremity movement captured by accelerometers, ambient light levels, Sound Pressure Level (SPL), and indoor air quality measured by volatile organic compounds, and the equivalent CO₂ concentration. For delirium risk assessment, the system recognizes mobility cues (axial body movement features and body key points) and circadian cues, including nightly disruptions, ambient SPL, and light intensity, as well as other environmental factors such as indoor air quality. The Mobi-DiQ system consists of three major components: the pervasive sensing system, a data storage and analysis server, and a data annotation system. For data collection, six local pervasive sensing systems were deployed, including a local computer and sensors. A video recording tool with graphical user interface (GUI) developed in python was used to capture depth image frames for analyzing patient mobility. All sensor data is encrypted, then automatically uploaded to the Mobi-DiQ server through a secured VPN connection. Several data pipelines are developed to automate the data transfer, curation, and data preparation for annotation and model training. The data curation and post-processing are performed on the server. A custom secure annotation tool with GUI was developed to annotate depth activity data. The annotation tool is linked to the MongoDB database to record the data annotation and to provide summarization. Docker containers are also utilized to manage services and pipelines running on the server in an isolated manner. The processed clinical data and annotations are used to train and develop real-time pervasive sensing systems to augment clinical decision-making and promote targeted interventions. In the future, we intend to evaluate our system as a clinical implementation trial, as well as to refine and validate it by using other data sources, including neurological data obtained through continuous electroencephalography (EEG).

Keywords: deep learning, delirium, healthcare, pervasive sensing

Procedia PDF Downloads 93

Authors: Vincent Schlageter, Maroua Mestiri, Florian Denzinger, Hugo Raetzo, Michel Demierre

Abstract:

Precise landslide monitoring with differential global navigation satellite system (GNSS) is well known, but technical or economic reasons limit its application by geotechnical companies. This study demonstrates the reliability and the usefulness of Geomon (Infrasurvey Sàrl, Switzerland), a stand-alone and cost-effective rover network. The system permits deploying up to 15 rovers, plus one reference station for differential GNSS. A dedicated radio communication links all the modules to a base station, where an embedded computer automatically provides all the relative positions (L1 phase, open-source RTKLib software) and populates an Internet server. Each measure also contains information from an internal inclinometer, battery level, and position quality indices. Contrary to standard GNSS survey systems, which suffer from a limited number of beacons that must be placed in areas with good GSM signal, Geomon offers greater flexibility and permits a real overview of the whole landslide with good spatial resolution. Each module is powered with solar panels, ensuring autonomous long-term recordings. In this study, we have tested the system on several sites in the Swiss mountains, setting up to 7 rovers per site, for an 18 month-long survey. The aim was to assess the robustness and the accuracy of the system in different environmental conditions. In one case, we ran forced blind tests (vertical movements of a given amplitude) and compared various session parameters (duration from 10 to 90 minutes). Then the other cases were a survey of real landslides sites using fixed optimized parameters. Sub centimetric-accuracy with few outliers was obtained using the best parameters (session duration of 60 minutes, baseline 1 km or less), with the noise level on the horizontal component half that of the vertical one. The performance (percent of aborting solutions, outliers) was reduced with sessions shorter than 30 minutes. The environment also had a strong influence on the percent of aborting solutions (ambiguity search problem), due to multiple reflections or satellites obstructed by trees and mountains. The length of the baseline (distance reference-rover, single baseline processing) reduced the accuracy above 1 km but had no significant effect below this limit. In critical weather conditions, the system’s robustness was limited: snow, avalanche, and frost-covered some rovers, including the antenna and vertically oriented solar panels, leading to data interruption; and strong wind damaged a reference station. The possibility of changing the sessions’ parameters remotely was very useful. In conclusion, the rover network tested provided the foreseen sub-centimetric-accuracy while providing a dense spatial resolution landslide survey. The ease of implementation and the fully automatic long-term survey were timesaving. Performance strongly depends on surrounding conditions, but short pre-measures should allow moving a rover to a better final placement. The system offers a promising hazard mitigation technique. Improvements could include data post-processing for alerts and automatic modification of the duration and numbers of sessions based on battery level and rover displacement velocity.

Keywords: GNSS, GSM, landslide, long-term, network, solar, spatial resolution, sub-centimeter.

Procedia PDF Downloads 113

Authors: Qiaosi Li, Frankie Kwan Kit Wong, Tung Fung

Abstract:

Lightweight unmanned aerial vehicle (UAV) loading with novel sensors offers a low cost approach for data acquisition in complex environment. This study established a framework for applying UAV system in complex environment quick mapping and assessed the performance of UAV-based hyperspectral image and digital surface model (DSM) derived from photogrammetric point clouds for 13 species classification in wetland area Mai Po Inner Deep Bay Ramsar Site, Hong Kong. The study area was part of shallow bay with flat terrain and the major species including reedbed and four mangroves: Kandelia obovata, Aegiceras corniculatum, Acrostichum auerum and Acanthus ilicifolius. Other species involved in various graminaceous plants, tarbor, shrub and invasive species Mikania micrantha. In particular, invasive species climbed up to the mangrove canopy caused damage and morphology change which might increase species distinguishing difficulty. Hyperspectral images were acquired by Headwall Nano sensor with spectral range from 400nm to 1000nm and 0.06m spatial resolution image. A sequence of multi-view RGB images was captured with 0.02m spatial resolution and 75% overlap. Hyperspectral image was corrected for radiative and geometric distortion while high resolution RGB images were matched to generate maximum dense point clouds. Furtherly, a 5 cm grid digital surface model (DSM) was derived from dense point clouds. Multiple feature reduction methods were compared to identify the efficient method and to explore the significant spectral bands in distinguishing different species. Examined methods including stepwise discriminant analysis (DA), support vector machine (SVM) and minimum noise fraction (MNF) transformation. Subsequently, spectral subsets composed of the first 20 most importance bands extracted by SVM, DA and MNF, and multi-source subsets adding extra DSM to 20 spectrum bands were served as input in maximum likelihood classifier (MLC) and SVM classifier to compare the classification result. Classification results showed that feature reduction methods from best to worst are MNF transformation, DA and SVM. MNF transformation accuracy was even higher than all bands input result. Selected bands frequently laid along the green peak, red edge and near infrared. Additionally, DA found that chlorophyll absorption red band and yellow band were also important for species classification. In terms of 3D data, DSM enhanced the discriminant capacity among low plants, arbor and mangrove. Meanwhile, DSM largely reduced misclassification due to the shadow effect and morphological variation of inter-species. In respect to classifier, nonparametric SVM outperformed than MLC for high dimension and multi-source data in this study. SVM classifier tended to produce higher overall accuracy and reduce scattered patches although it costs more time than MLC. The best result was obtained by combining MNF components and DSM in SVM classifier. This study offered a precision species distribution survey solution for inaccessible wetland area with low cost of time and labour. In addition, findings relevant to the positive effect of DSM as well as spectral feature identification indicated that the utility of UAV-borne hyperspectral and photogrammetry deriving 3D data is promising in further research on wetland species such as bio-parameters modelling and biological invasion monitoring.

Keywords: digital surface model (DSM), feature reduction, hyperspectral, photogrammetric point cloud, species mapping, unmanned aerial vehicle (UAV)

Procedia PDF Downloads 257

Authors: Fani Sakellariadou

Abstract:

The Blue Economy includes all marine-based and marine-related activities. They correspond to established, emerging as well as unborn ocean-based industries. Seabed mining is an emerging marine-based activity; its operations depend particularly on cutting-edge science and technology. The 21st century will face a crisis in resources as a consequence of the world’s population growth and the rising standard of living. The natural capital stored in the global ocean is decisive for it to provide a wide range of sustainable ecosystem services. Seabed mineral deposits were identified as having a high potential for critical elements and base metals. They have a crucial role in the fast evolution of green technologies. The major categories of marine mineral deposits are deep-sea deposits, including cobalt-rich ferromanganese crusts, polymetallic nodules, phosphorites, and deep-sea muds, as well as shallow-water deposits including marine placers. Seabed mining operations may take place within continental shelf areas of nation-states. In international waters, the International Seabed Authority (ISA) has entered into 15-year contracts for deep-seabed exploration with 21 contractors. These contracts are for polymetallic nodules (18 contracts), polymetallic sulfides (7 contracts), and cobalt-rich ferromanganese crusts (5 contracts). Exploration areas are located in the Clarion-Clipperton Zone, the Indian Ocean, the Mid Atlantic Ridge, the South Atlantic Ocean, and the Pacific Ocean. Potential environmental impacts of deep-sea mining include habitat alteration, sediment disturbance, plume discharge, toxic compounds release, light and noise generation, and air emissions. They could cause burial and smothering of benthic species, health problems for marine species, biodiversity loss, reduced photosynthetic mechanism, behavior change and masking acoustic communication for mammals and fish, heavy metals bioaccumulation up the food web, decrease of the content of dissolved oxygen, and climate change. An important concern related to deep-sea mining is our knowledge gap regarding deep-sea bio-communities. The ecological consequences that will be caused in the remote, unique, fragile, and little-understood deep-sea ecosystems and inhabitants are still largely unknown. The blue economy conceptualizes oceans as developing spaces supplying socio-economic benefits for current and future generations but also protecting, supporting, and restoring biodiversity and ecological productivity. In that sense, people should apply holistic management and make an assessment of marine mining impacts on ecosystem services, including the categories of provisioning, regulating, supporting, and cultural services. The variety in environmental parameters, the range in sea depth, the diversity in the characteristics of marine species, and the possible proximity to other existing maritime industries cause a span of marine mining impact the ability of ecosystems to support people and nature. In conclusion, the use of the untapped potential of the global ocean demands a liable and sustainable attitude. Moreover, there is a need to change our lifestyle and move beyond the philosophy of single-use. Living in a throw-away society based on a linear approach to resource consumption, humans are putting too much pressure on the natural environment. Applying modern, sustainable and eco-friendly approaches according to the principle of circular economy, a substantial amount of natural resource savings will be achieved. Acknowledgement: This work is part of the MAREE project, financially supported by the Division VI of IUPAC. This work has been partly supported by the University of Piraeus Research Center.

Keywords: blue economy, deep-sea mining, ecosystem services, environmental impacts

Procedia PDF Downloads 85

Authors: Reza Nematirad, Anil Pahwa, Balasubramaniam Natarajan

Abstract:

Time series data consist of successive data points collected over a period of time. Accurate prediction of future values is essential for informed decision-making in several real-world applications, including electricity load demand forecasting, lifetime estimation of industrial machinery, traffic planning, weather prediction, and the stock market. Due to their critical relevance and wide application, there has been considerable interest in time series forecasting in recent years. However, the proliferation of sensors and IoT devices, real-time monitoring systems, and high-frequency trading data introduce significant intricate temporal variations, rapid changes, noise, and non-linearities, making time series forecasting more challenging. Classical methods such as Autoregressive integrated moving average (ARIMA) and Exponential Smoothing aim to extract pre-defined temporal variations, such as trends and seasonality. While these methods are effective for capturing well-defined seasonal patterns and trends, they often struggle with more complex, non-linear patterns present in real-world time series data. In recent years, deep learning has made significant contributions to time series forecasting. Recurrent Neural Networks (RNNs) and their variants, such as Long short-term memory (LSTMs) and Gated Recurrent Units (GRUs), have been widely adopted for modeling sequential data. However, they often suffer from the locality, making it difficult to capture local trends and rapid fluctuations. Convolutional Neural Networks (CNNs), particularly Temporal Convolutional Networks (TCNs), leverage convolutional layers to capture temporal dependencies by applying convolutional filters along the temporal dimension. Despite their advantages, TCNs struggle with capturing relationships between distant time points due to the locality of one-dimensional convolution kernels. Transformers have revolutionized time series forecasting with their powerful attention mechanisms, effectively capturing long-term dependencies and relationships between distant time points. However, the attention mechanism may struggle to discern dependencies directly from scattered time points due to intricate temporal patterns. Lastly, Multi-Layer Perceptrons (MLPs) have also been employed, with models like N-BEATS and LightTS demonstrating success. Despite this, MLPs often face high volatility and computational complexity challenges in long-horizon forecasting. To address intricate temporal variations in time series data, this study introduces Times2D, a novel framework that parallelly integrates 2D spectrogram and derivative heatmap techniques. The spectrogram focuses on the frequency domain, capturing periodicity, while the derivative patterns emphasize the time domain, highlighting sharp fluctuations and turning points. This 2D transformation enables the utilization of powerful computer vision techniques to capture various intricate temporal variations. To evaluate the performance of Times2D, extensive experiments were conducted on standard time series datasets and compared with various state-of-the-art algorithms, including DLinear (2023), TimesNet (2023), Non-stationary Transformer (2022), PatchTST (2023), N-HiTS (2023), Crossformer (2023), MICN (2023), LightTS (2022), FEDformer (2022), FiLM (2022), SCINet (2022a), Autoformer (2021), and Informer (2021) under the same modeling conditions. The initial results demonstrated that Times2D achieves consistent state-of-the-art performance in both short-term and long-term forecasting tasks. Furthermore, the generality of the Times2D framework allows it to be applied to various tasks such as time series imputation, clustering, classification, and anomaly detection, offering potential benefits in any domain that involves sequential data analysis.

Keywords: derivative patterns, spectrogram, time series forecasting, times2D, 2D representation

Procedia PDF Downloads 44