IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS Flying through cluttered and dynamic environments with LiDAR Huajie Wu, Wenyi Liu, Y unfan Ren, Zheng Liu, Hairuo Wei, Fangcheng Zhu, Haotian Li, and Fu Zhang Abstract --Navigating unmanned aerial vehicles (UAVs) through cluttered and dynamic environments remains a significant challenge, particularly when dealing with fast-moving or sudden-appearing obstacles. This paper introduces a complete LiDAR-based system designed to enable UAVs to avoid various moving obstacles in complex environments. Benefiting the high computational efficiency of perception and planning, the system can operate in real time using onboard computing resources with low latency. For dynamic environment perception, we have integrated our previous work, M-detector, into the system. M-detector ensures that moving objects of different sizes, colors, and types are reliably detected. For dynamic environment planning, we incorporate dynamic object predictions into the integrated planning and control (IPC) framework, namely DynIPC. This integration allows the UAV to utilize predictions about dynamic obstacles to effectively evade them. We validate our proposed system through both simulations and real-world experiments. In simulation tests, our system outperforms state-of-the-art baselines across several metrics, including success rate, time consumption, average flight time, and maximum velocity. Index Terms --LiDAR-based UAV, dynamic obstacle avoidance, cluttered and dynamic environment I. I NTRODUCTION I N recent years, the development of lightweight and high-precision sensors, such as Light Detection and Ranging sensors (LiDAR), event cameras, and depth cameras, has significantly advanced the autonomous flight capabilities of unmanned aerial vehicles (UA Vs) or drones. This technological progress has facilitated the widespread application of drones across various industries, including agricultural spraying [1], product delivery [2], inspection [3], and search and rescue [4]. These applications have notably enhanced production efficiency, reduced costs, and driven economic growth within these sectors.

In addition to revealing its initial plans for AI and annual updates to iOS, macOS and more at WWDC 2024, Apple also discussed new capabilities coming to the second-gen AirPods Pro. Siri Interactions will allow you to respond to the assistant by nodding your head yes or shaking your head no. Apple also plans to introduce improved Voice Isolation that further reduces background noise when you're on a call. Both of these items are exclusive to the most recent AirPods Pro, because they rely on the company's H2 chip like existing Adaptive Audio, Personalized Volume and Conversation Awareness features. Like those advanced audio tools that are already available on AirPods Pro, Siri Interactions and Voice Isolation use the processing abilities of the H2 chip in tandem with the power of a source device -- an iPhone or MacBook Pro, for example.

Neural implicit representations have emerged as a promising solution for providing dense geometry in Simultaneous Localization and Mapping (SLAM). However, existing methods in this direction fall short in terms of global consistency and low latency. This paper presents NGEL-SLAM to tackle the above challenges. To ensure global consistency, our system leverages a traditional feature-based tracking module that incorporates loop closure. Additionally, we maintain a global consistent map by representing the scene using multiple neural implicit fields, enabling quick adjustment to the loop closure. Moreover, our system allows for fast convergence through the use of octree-based implicit representations. The combination of rapid response to loop closure and fast convergence makes our system a truly low-latency system that achieves global consistency. Our system enables rendering high-fidelity RGB-D images, along with extracting dense and complete surfaces. Experiments on both synthetic and real-world datasets suggest that our system achieves state-of-the-art tracking and mapping accuracy while maintaining low latency.

Simultaneous machine translation (SiMT) models are trained to strike a balance between latency and translation quality. However, training these models to achieve high quality while maintaining low latency often leads to a tendency for aggressive anticipation. We argue that such issue stems from the autoregressive architecture upon which most existing SiMT models are built. To address those issues, we propose non-autoregressive streaming Transformer (NAST) which comprises a unidirectional encoder and a non-autoregressive decoder with intra-chunk parallelism. We enable NAST to generate the blank token or repetitive tokens to adjust its READ/WRITE strategy flexibly, and train it to maximize the non-monotonic latent alignment with an alignment-based latency loss. Experiments on various SiMT benchmarks demonstrate that NAST outperforms previous strong autoregressive SiMT baselines.

Spiking Neural Network (SNN) is a promising energy-efficient AI model when implemented on neuromorphic hardware. However, it is a challenge to efficiently train SNNs due to their non-differentiability. Most existing methods either suffer from high latency (i.e., long simulation time steps), or cannot achieve as high performance as Artificial Neural Networks (ANNs). In this paper, we propose the Differentiation on Spike Representation (DSR) method, which could achieve high performance that is competitive to ANNs yet with low latency. First, we encode the spike trains into spike representation using (weighted) firing rate coding. Based on the spike representation, we systematically derive that the spiking dynamics with common neural models can be represented as some sub-differentiable mapping. With this viewpoint, our proposed DSR method trains SNNs through gradients of the mapping and avoids the common non-differentiability problem in SNN training. Then we analyze the error when representing the specific mapping with the forward computation of the SNN. To reduce such error, we propose to train the spike threshold in each layer, and to introduce a new hyperparameter for the neural models. With these components, the DSR method can achieve state-of-the-art SNN performance with low latency on both static and neuromorphic datasets, including CIFAR-10, CIFAR-100, ImageNet, and DVS-CIFAR10.

The fifth generation of mobile networks, or 5G, claims to be faster, more dependable, and more effective than the generations before it. It's a technology that makes use of novel modulation techniques, a wider spectrum, and numerous antennas, among other innovations, to enable quicker and more dependable data delivery. A variety of novel use cases that were not feasible with prior network generations, such as augmented and virtual reality, autonomous vehicles, and smart cities, will be made viable by the enhanced capacity and speed of 5G networks. Worldwide deployment of 5G technology has already begun in many nations, and in the upcoming years, it is expected to overtake all other mobile network technologies. The Internet of Things (IoT) refers to a network of internet-connected devices that can communicate and share data with one another. This network includes everything from wearables and smart home appliances to heavy machinery and transportation vehicles.

As new technology is constantly being developed, trends are merged and combined to enhance functionality and improve old systems. The future of fifth-generation cellular technology and artificial intelligence is a perfect example of how today's innovators can apply two separate concepts together to develop new use cases and refine the inventions of the past to better serve the needs of the future. Read on to learn more about how 5G and AI technology will work together to produce exciting new developments. With their powers combined, AI and 5G technologies are a superforce. This dynamic duo has the potential to transform many different industries -- from healthcare and transportation to entertainment and beyond.

AMD announced that Japanese automotive systems supplier Aisin is using its Xilinx Automotive (XA) Zynq UltraScale MPSoC platform for its Automated Parking-Assist (APA) system. The highly adaptable platform enables the next-generation Aisin APA system to detect pedestrians, vehicles and free space efficiently and at extremely low latency. The Aisin APA system will begin shipping in model year 2024. The Aisin APA system uses four cameras and 12 ultrasonic sensors mounted on the vehicle to recognize the surrounding environment and calculate the driving route. The system then controls the vehicle according to the calculated route to park itself.

Simultaneous speech translation (SimulST) is a challenging task aiming to translate streaming speech before the complete input is observed. A SimulST system generally includes two components: the pre-decision that aggregates the speech information and the policy that decides to read or write. While recent works had proposed various strategies to improve the pre-decision, they mainly adopt the fixed wait-k policy, leaving the adaptive policies rarely explored. This paper proposes to model the adaptive policy by adapting the Continuous Integrate-and-Fire (CIF). Compared with monotonic multihead attention (MMA), our method has the advantage of simpler computation, superior quality at low latency, and better generalization to long utterances. We conduct experiments on the MuST-C V2 dataset and show the effectiveness of our approach.

Computing specialist Blaize has agreed a strategic partnership with Accton, a premier provider of networking and communications solutions, to bring edge AI computing to the AI inspection market. The Accton Smart Automated Optical Inspection (AOI) solution will utilize the Blaize Pathfinder P1600 Embedded System on Module (SoM) to add visual AI production line inspection for assembly, manufacturing, packaging, and appearance activities. Colby Chou, IoT business unit head of Accton, says: "We are pleased to partner with Blaize to provide our customers with a cost-effective AI inspection service. Our solution helps our customer reduces up to 85 percent of the operators' workload and significantly improves product quality. The Accton product Pallas, uses Blaize's P1600 SoM, leveraging the programmability and efficiency benefits of the Blaize Graph Streaming Processor (GSP) architecture. The SoM is ideal for rugged and challenging environments and offers the processing power, low latency, and energy efficiency crucial for AI inferencing workloads at the edge and the inherent stringent inspection requirements. Accton will be able to implement computer vision applications and new AI inferencing solutions across a range of edge smart vision use cases using the Blaize architecture. Dinakar Munagala, co-founder and CEO of Blaize, says: "Blaize looks forward to providing Accton with a solution that enables stricter quality standards, higher yield, and more efficient manufacturing and inspection processes.