Advanced driver-assistance systems (ADASs) have become a salient feature for safety in modern vehicles. They are also key underlying technology emerging autonomous State-of-the-art ADASs primarily vision based, but light detection and ranging (lidar), radio (radar), other advanced-sensing technologies becoming popular. In this article, we present survey of different hardware software ADAS their capabilities limitations. We discuss approaches used vision-based recognition sensor fusion...

Mobile and IoT devices have proliferated our daily lives. However, these miniaturized computing systems should be highly energy-efficient due to their ultrasmall form factor. Hence, energy management is of utmost importance for both mobile devices. This article presents a comprehensive survey on this topic.

Deep learning has led to unprecedented successes in solving some very difficult problems domains such as computer vision, natural language processing, and general pattern recognition. These achievements are the culmination of decades-long research into better training techniques deeper neural network models, well improvements hardware platforms that used train execute deep models. Many application-specific integrated circuit (ASIC) accelerators for have garnered interest recent years due...

Autonomous vehicles are on the horizon and will be transforming transportation safety comfort. These connected to various external systems utilize advanced embedded perceive their environment make intelligent decisions. However, this increased connectivity makes these vulnerable cyber-attacks that can have catastrophic effects. Attacks automotive already rise in today's expected become more commonplace future autonomous vehicles. Thus, there is a need strengthen cybersecurity In article, we...

Object detection is a computer vision task that has become an integral part of many consumer applications today such as surveillance and security systems, mobile text recognition, diagnosing diseases from MRI/CT scans. also one the critical components to support autonomous driving. Autonomous vehicles rely on perception their surroundings ensure safe robust driving performance. This system uses object algorithms accurately determine objects pedestrians, vehicles, traffic signs, barriers in...

Silicon photonics (SiPh) technology has facilitated the deployment of integrated across different application domains, from ultra-fast communication in Datacom applications to energy-efficient optical computation emerging hardware accelerators for machine learning. More recently, integration SiPh and phase change materials created a unique opportunity realize adaptable, reconfigurable, programmable photonic platforms. In particular, nonvolatile programmability made them promising candidate...

Three-dimensional integrated circuits (3D ICs) offer a promising solution to overcome the on-chip communication bottleneck and improve performance over traditional two-dimensional (2D) ICs. Long interconnects can be replaced by much shorter vertical through silicon via (TSV) in 3D This enables faster more power efficient inter-core across multiple layers. However, IC technology also faces challenges due higher densities routing congestion TSV pads distributed on each layer. In this paper,...

Indoor localization is emerging as an important application domain for enhanced navigation (or tracking) of people and assets in indoor locales such buildings, malls, underground mines. Most solutions proposed prior work do not deliver good accuracy without expensive infrastructure (and even then, the results may lack consistency). Ambient wireless received signal strength indication (RSSI) based fingerprinting using smart mobile devices a low-cost approach to problem. However, creating...

Domain-specific neural network accelerators have seen growing interest in recent years due to their improved energy efficiency and performance compared CPUs GPUs. In this paper, we propose a novel cross-layer optimized accelerator called CrossLight that leverages silicon photonics. includes device-level engineering for resilience process variations thermal crosstalk, circuit-level tuning enhancements inference latency reduction, architecture-level optimizations enable better resolution,...

System-on-Chip (SoC) designs are increasingly becoming more complex. Efficient on-chip communication architectures critical for achieving desired performance in these systems. System designers typically use Bus Cycle Accurate (BCA) models written high level languages such as C/C++ to explore the design space. These capture all of bus signals and strictly maintain cycle accuracy, which is useful reliable exploration but results slow simulation speeds complex designs, even when they modeled...

Three-dimensional integrated circuits (3D-ICs) offer a significant opportunity to enhance the performance of emerging chip multiprocessors (CMPs) using high density stacked device integration and shorter through silicon via (TSV) interconnects that can alleviate some problems associated with interconnect scaling in s ub-65nm CMOS technologies. However, network-on-chip (NoC) fabrics will connect cores together 3D-ICs increasingly be susceptible permanent intermittent faults, which cause...

Within the past decade, mobile computing has morphed into a principal form of human communication, business, and social interaction. Unfortunately, energy demands newer ambient intelligence collaborative technologies on devices have greatly overwhelmed modern storage abilities. This paper proposes several novel techniques that exploit spatiotemporal device context to predict wireless data location interface configurations can optimize consumption in devices. These techniques, which include...

Many of today's data centers experience physical limitations on the power needed to run center. The first problem that we study is maximizing performance (quantified by reward collected for completing tasks their individual deadlines) a center subject total consumption (of compute nodes and CRAC units) thermal constraints. second how minimize in while guaranteeing overall does not drop below specified threshold. For both problems, develop novel optimization techniques assigning states cores...

The need for indoor localization systems that can provide reliable access to location information in areas are not serviced sufficiently by a global positioning system (GPS) has continued grow. There wide variety of use cases this data and increasing interest from industry, academia, government agencies fueled research area. Smartphones uniquely positioned be critical part solution based on the proliferation these devices diverse array sensors radios they contain. In article, capabilities...

With increasing application complexity and improvements in process technology, Chip MultiProcessors (CMPs) with tens to hundreds of cores on a chip are becoming reality. Networks-on-Chip (NoCs) have emerged as scalable communication fabrics that can support high bandwidths for these massively parallel multicore systems. However, traditional electrical NoC implementations still need overcome the challenges data transfer latencies large power consumption. On-chip photonic interconnects...

Silicon photonics is an emerging paradigm for designing energy -efficient and high-performance communication backbone manycore chips. This article presents a survey about photonic on-chip infrastructure.

Today's vehicles are complex distributed embedded systems that increasingly being connected to various external systems. Unfortunately, this increased connectivity makes the vulnerable security attacks can be catastrophic. In article, we present a novel intrusion detection system (IDS) called INDRA utilizes gated recurrent unit (GRU)-based autoencoder detect anomalies in controller area network (CAN) bus-based automotive We evaluate our proposed framework under different attack scenarios and...

GPS technology has revolutionized the way we localize and navigate outdoors. However, poor reception of signals in buildings makes it unsuitable for indoor localization. WiFi fingerprinting-based localization is one most promising ways to meet this demand. Unfortunately, work domain fails resolve challenges associated with deployability on resource-limited embedded devices. In work, propose a compression-aware high-accuracy deep learning framework called CHISEL that outperforms best-known...

Cloud service providers are distributing data centers geographically to minimize energy costs through intelligent workload distribution. With increasing volumes in emerging cloud workloads, it is critical factor the network for transferring workloads across centers. For geo-distributed centers, many researchers have been exploring strategies cost minimization and inter-data-center distribution separately. However, prior work does not comprehensively simultaneously consider center costs,...

Modern vehicles have multiple electronic control units (ECUs) that are connected together as part of a complex distributed cyber-physical system (CPS). The ever-increasing communication between ECUs and external systems has made these particularly susceptible to variety cyber-attacks. In this work, we present novel anomaly detection framework called TENET detect anomalies induced by cyber-attacks on vehicles. uses temporal convolutional neural networks with an integrated attention mechanism...

Coherent integrated photonic neural networks (IPNNs) are increasingly being explored for rapidly growing artificial intelligence applications. However, the principal roadblocks in scalability of IPNNs their large area footprint and high tuning power consumption during both training inferencing. In deep (DNNs), software techniques to prune redundant weights often utilized reduce resource (e.g., memory, computation, power) overheads. due complex nature which mapped building blocks IPNNs, prior...

Transformer neural networks are rapidly being integrated into state-of-the-art solutions for natural language processing (NLP) and computer vision. However, the complex structure of these models creates challenges accelerating their execution on conventional electronic platforms. We propose first silicon photonic hardware network accelerator called TRON transformer-based such as BERT, Vision Transformers. Our analysis demonstrates that exhibits at least 14× better throughput 8× energy...