Low Collision Rate Research Articles

Multiagent trajectory decision‐making for complex systems based on transformer large model structure

AbstractThe article explores the application of transformer‐based large models for trajectory decision‐making in complex decision systems, focusing on multiagent nodes in a topological structure. By leveraging the capabilities of large models, the proposed approach enhances decision‐making in model‐free control environments characterized by numerous agents, diverse internal tasks, and unknown conditions. The proposed framework integrates advanced transformer models with a sophisticated decision‐making framework to enhance the efficiency, safety, and reliability of multiagent operations in dynamic environments. The results demonstrate the superior performance of the transformer model, achieving a trajectory decision‐making accuracy of 97.8%, significantly outperforming other models such as long short‐term memory (LSTM), gate recurrent unit (GRU), and traditional approaches. The visualizations highlight the close alignment between the true and predicted outcomes of complex decision‐making systems, showcasing the model's ability to capture complex dependencies and interactions within the multiagent environment. High task completion rates and low collision rates further underscore the effectiveness of the decision‐making framework in optimizing agent coordination and ensuring safe operations.

Safe Hybrid-Action Reinforcement Learning-Based Decision and Control for Discretionary Lane Change

Autonomous lane-change, a key feature of advanced driver-assistance systems, can enhance traffic efficiency and reduce the incidence of accidents. However, safe driving of autonomous vehicles remains challenging in complex environments. How to perform safe and appropriate lane change is a popular topic of research in the field of autonomous driving. Currently, few papers consider the safety of reinforcement learning in discretionary lane-change scenarios. We introduce safe hybrid-action reinforcement learning into discretionary lane change for the first time and propose the Parameterized Soft Actor–Critic with PID Lagrangian (PASAC-PIDLag) algorithm. Furthermore, we conduct a comparative analysis with Parameterized Soft Actor–Critic (PASAC), which is an unsafe version of PASAC-PIDLag. Both algorithms are employed to train the lane-change strategy to output both discrete lane-change decisions and continuous longitudinal vehicle acceleration. Our simulation results indicate that at a traffic density of 15 vehicles per kilometer (15 veh/km), the PASAC-PIDLag algorithm exhibits superior safety with a collision rate of 0%, outperforming the PASAC algorithm, which has a collision rate of 1%. The generalization assessments reveal that at low traffic density levels, both the PASAC-PIDLag and PASAC algorithms are proficient in attaining zero collision rates. However, at high traffic density levels, although both algorithms result in collisions, PASAC-PIDLag has a much lower collision rate than PASAC.

Multi-degree-of-freedom unmanned aerial vehicle control combining a hybrid brain-computer interface and visual obstacle avoidance

ObjectiveThe difficulty of unmanned aerial vehicle (UAV) control recently lies in multidirectional movement in 3-dimensional space, improving control accuracy and manipulation safety. To address these challenges, a UAV control system that incorporates a hybrid brain-computer interface (hBCI), gyroscope and visual obstacle avoidance based on monocular depth estimation is proposed. Approach. We propose an efficient steady-state visual evoked potential (SSVEP) classification network (CL-NET) featuring a one-dimensional convolutional neural network, a long short-term memory module and an attention module to identify the user's intention for UAV movement in the front, back, left and right directions. The take-off, landing and rising control of the UAV is realized by an electrooculogram (EOG) signal detection algorithm, a blink state detector. In addition, the UAV can fly in an oblique state and rotate according to the current head posture detected by a gyroscope. Furthermore, an improved monocular depth estimation network is employed to design the autonomous obstacle avoidance module of the UAV, ensuring the safety of the brain-controlled system in practice. Main results. The proposed CL-NET delivers an accuracy of 98.67% on the public dataset and an accuracy of 97.92% on the self-collected dataset, both of which surpass the performance of state-of-the-art models. Additionally, we set up a brain control group and a remote control group to conduct practical experiments in a realistic environment. In the experiments involving sixteen subjects, the proposed UAV control system reached an average information transfer rate (ITR) of 44.09 bits/min, and the brain control group had a lower collision rate than the remote control group. Significance. The hybrid control method ensures that the multi-degree-of-freedom (multi-DOF) UAV control system maintains outstanding performance while ensuring good safety.

Sensor Fish Deployments at the Xayaburi Hydropower Plant: Measurements and Simulations

Fish protection is a priority in regions with ongoing and planned development of hydropower production, like the Mekong River system. The evaluation of the effects of turbine passage on the survival of migratory fish is a primary task for informing hydropower plant operators and authorities about the environmental performance of plant operations. The present work characterizes low pressures and collision rates through the Kaplan-type runners of the Xayaburi hydropower station. Both an experimental method based on the deployment of Sensor Fish and a numerical strategy based on flow and passage simulations were implemented on the analysis of two release elevations at one operating point. Nadir pressures and pressure drops through the runner were very sensitive to release elevation, but collision rates on the runner were not. The latter showed a frequency of occurrence of 8.2–9.3%. Measured magnitudes validated the corresponding simulation outcomes in regard to the averaged magnitudes as well as to the variability. Central to this study is that simulations were conducted based on current industry practices for designing turbines. Therefore, the reported agreement helps turbine engineers gain certainty about the prediction power of flow and trajectory simulations for fish passage assessments. This can accelerate the development of environmentally enhanced technology with minimum impact on natural resources.

Conflict-free joint decision by lag and zero-lag synchronization in laser network

With the end of Moore’s Law and the increasing demand for computing, photonic accelerators are garnering considerable attention. This is due to the physical characteristics of light, such as high bandwidth and multiplicity, and the various synchronization phenomena that emerge in the realm of laser physics. These factors come into play as computer performance approaches its limits. In this study, we explore the application of a laser network, acting as a photonic accelerator, to the competitive multi-armed bandit problem. In this context, conflict avoidance is key to maximizing environmental rewards. We experimentally demonstrate cooperative decision-making using zero-lag and lag synchronization within a network of four semiconductor lasers. Lag synchronization of chaos realizes effective decision-making and zero-lag synchronization is responsible for the realization of the collision avoidance function. We experimentally verified a low collision rate and high reward in a fundamental 2-player, 2-slot scenario, and showed the scalability of this system. This system architecture opens up new possibilities for intelligent functionalities in laser dynamics.

An automatic driving trajectory planning approach in complex traffic scenarios based on integrated driver style inference and deep reinforcement learning.

As autonomous driving technology continues to advance and gradually become a reality, ensuring the safety of autonomous driving in complex traffic scenarios has become a key focus and challenge in current research. Model-free deep reinforcement learning (Deep Reinforcement Learning) methods have been widely used for addressing motion planning problems in complex traffic scenarios, as they can implicitly learn interactions between vehicles. However, current planning methods based on deep reinforcement learning exhibit limited robustness and generalization performance. They struggle to adapt to traffic conditions beyond the training scenarios and face difficulties in handling uncertainties arising from unexpected situations. Therefore, this paper addresses the challenges presented by complex traffic scenarios, such as signal-free intersections. It does so by first utilizing the historical trajectories of adjacent vehicles observed in these scenarios. Through a Variational Auto-Encoder (VAE) based on the Gated Recurrent Unit (GRU) recurrent neural network, it extracts driver style features. These driver style features are then integrated with other state parameters and used to train a motion planning strategy within an extended reinforcement learning framework. This approach ultimately yields a more robust and interpretable mid-to-mid motion planning method. Experimental results confirm that the proposed method achieves low collision rates, high efficiency, and successful task completion in complex traffic scenarios.

Hierarchical Relational Graph Learning for Autonomous Multirobot Cooperative Navigation in Dynamic Environments

As a specific kind of Cyber-Physical Systems (CPSs), autonomous robot clusters play an important role in various intelligent manufacturing fields. However, due to the increasing design complexity of robot clusters, it is becoming more and more challenging to guarantee the safety and efficiency for multi-robot cooperative navigation in dynamic and complex environments. Although Deep Reinforcement Learning (DRL) shows great potential in learning multi-robot cooperative navigation policies, existing DRL-based approaches suffer from scalability issues and rarely consider the transferability of trained policies to new tasks. To address these problems, this paper presents a novel DRL-based multi-robot cooperative navigation approach named HRMR-Navi that equips each robot with both a two-layered hierarchical graph network model and an attention-based communication model. In our approach, the hierarchical graph network model can efficiently figure out hierarchical relations among all agents that either cooperate for efficiency or avoid obstacles for safety to derive more advanced strategies, and the communication model can accurately form a global view of the environment for a specific robot, thus the multi-robot cooperation efficiency can be further strengthened. Meanwhile, we propose an improved Proximal Policy Optimization (PPO) algorithm based on the Maximum Entropy reinforcement learning, named MEPPO, to enhance the robot exploration ability. Comprehensive experimental results demonstrate that, compared with state-of-the-art approaches, HRMR-Navi can achieve more efficient cooperative navigation with less time cost, lower collision rate, higher scalability, and better knowledge transferability.

Self-Learned Autonomous Driving at Unsignalized Intersections: A Hierarchical Reinforced Learning Approach for Feasible Decision-Making

Reinforcement learning-based techniques, empowered by deep-structured neural nets, have demonstrated superiority over rule-based methods in terms of making high-level behavioral decisions due to qualities related to handling large state spaces. Nonetheless, their training time, sample efficiency and the feasibility of the learnt behaviors remain key concerns. In this paper, we propose a novel hierarchical reinforcement learning-based decision-making architecture for learning left-turn policies at unsignalized intersections with feasibility guarantees. The proposed technique is comprised of two layers; a high-level learning-based behavioral planning layer which adopts soft actor-critic (SAC) principles to learn high-level, non-conservative yet safe, driving behaviors, and a low-level motion planning layer that uses Model Predictive Control (MPC) framework to ensure feasibility of the two-dimensional left-turn maneuver. The high-level layer generates reference signals of velocity and yaw angles for the ego vehicle taking into account safety and collision avoidance with the intersection vehicles, whereas the low-level motion planning layer solves an optimization problem to track these reference commands taking into account several vehicle dynamic constraints and ride comfort. While training the behavioral SAC-based planning layer, We develop an adaptive entropy regularization technique that results in faster convergence, higher mean rewards, and lower collision rates. We validate the proposed decision-making scheme in simulated environments and compare with other model free Reinforcement Learning (RL) baselines. The results demonstrate that the proposed integrated framework possesses better training and navigation capabilities.

Nexus of Deep Reinforcement Learning and Leader–Follower Approach for AIoT Enabled Aerial Networks

The Industrial Internet of Things (IIoT) is a new industrial 4.0 paradigm that combines IoT, robotics, cyber-physical systems, and other future industrial advancements. Unmanned Aerial Vehicles (UAVs), part of the IIoT infrastructure, have a significant potential for civil and military purposes. Through the Artificial Intelligence of Things (AIoT), a well-organized group of UAVs outperforms a single large UAV in terms of device scalability, maintenance, and expense. Therefore, the UAV swarm with industry 4.0 intelligence can be used for a wide range of 24/7 security and remote monitoring applications. Though multi-UAV systems are beneficial, their application has many challenges. There is a high risk of collision in the multi-UAV system without coordination. This paper proposes an AIoT-based Navigation and Formation Control (AIoT-NFC) mechanism to scale down the collision risk by combining Deep Reinforcement Learning (DRL) with the leader-follower approach. In AIoT-NFC, a Deep Deterministic Policy Gradient (DDPG) based algorithm is proposed to navigate UAVs in remote surveillance without colliding with obstacles and other UAVs. Furthermore, the AIoT-NFC system incorporates a fault tolerance mechanism that can handle the scenario of a leader's failure due to actuator malfunction. Experimental results show that the AIoT-NFC achieves faster convergence with a lower collision rate. AIoT-NFC reduced the collision rate by 14.99% compared to existing navigation methods in successful formation without colliding with the other UAVs.

A novel distributed multi-slot TDMA-based MAC protocol for LED-based UOWC networks

Underwater optical wireless communication (UOWC) networks are promising for many civilian and industrial applications due to underwater high-speed data transmission demand. This paper focuses on the distributed time division multiple access (TDMA) protocol design for UOWC networks. Unlike existing research, we focus on a more general underwater communication scenario where nodes are scattered, and node mobility is considered. We propose a distributed TDMA-based medium access control(MAC) protocol, called cluster-based cross-layer multi-slot MAC(CCM-MAC), which can assign multiple slots to each node according to the slot-occupying information. When using CCM-MAC, the network is divided into clusters. Each node maintains knowledge of the cluster topology and slot-occupying information of surrounding nodes through routing techniques; by updating this knowledge, collisions can also be detected and eliminated dynamically. Analysis results are presented to evaluate the performance of CCM-MAC, and simulations show that high network throughput and low collision rate can be realized in comparison with existing contention-based MAC protocols.

A Survey on Safety-Critical Driving Scenario Generation—A Methodological Perspective

Autonomous driving systems have witnessed significant development during the past years thanks to the advance in machine learning-enabled sensing and decision-making algorithms. One critical challenge for their massive deployment in the real world is their safety evaluation. Most existing driving systems are still trained and evaluated on naturalistic scenarios collected from daily life or heuristically-generated adversarial ones. However, the large population of cars, in general, leads to an extremely low collision rate, indicating that safety-critical scenarios are rare in the collected real-world data. Thus, methods to artificially generate scenarios become crucial to measure the risk and reduce the cost. In this survey, we focus on the algorithms of safety-critical scenario generation in autonomous driving. We first provide a comprehensive taxonomy of existing algorithms by dividing them into three categories: data-driven generation, adversarial generation, and knowledge-based generation. Then, we discuss useful tools for scenario generation, including simulation platforms and packages. Finally, we extend our discussion to five main challenges of current works – fidelity, efficiency, diversity, transferability, controllability – and research opportunities lighted up by these challenges.

Enhanced image similarity search: Boundary of similarity approach with similarity hashing for Image Datasets

Image similarity searching is a well-explored topic with numerous techniques and approaches, each offering unique advantages. In this paper, we present an approach aimed at accelerating the search for the most similar images to a given query image within an image database. Our method leverages the principles of Locality-Sensitive Hashing (LSH) but relies on a concept of similarity boundaries to reach the search to most appropriate section and in turn facilitates a rapid retrieval of most similar content.We assume the availability of a reliable hash function for image similarity that offers high accuracy and a low collision rate and focus exclusively on the algorithm designed to maintain and manage similarity boundaries for each image within the hashtable. This strategic organization ensures that similar images are grouped together logically, significantly enhancing the efficiency and speed of the search process.

Efficient Cluster Tree Topology Operation and Routing for IEEE 802.15.4-Based Smart Grid Networks

Wireless sensor networks (WSNs) have been utilized as communication infrastructure for smart grid applications. The primary requirement of WSNs for smart grid applications is to transmit delay-critical data from smart grid assets ether at the maximum rate or by reducing collision rates. Additionally, WSNs should utilize the limited resources of the network to provide the required long-term QoS. The achievement of these objectives requires a remarkable design of WSN protocols to satisfy the requirements of smart grid applications. In this study, a multi-channel cluster tree protocol is proposed to prevent collisions and increase network performance. In the proposed scheme, the cluster head serves to broadcast a beacon frame containing information on the allocated channels and time slots. This enables the new node to determine its channel and timeslot. A performance analysis reveals that the proposed scheme can achieve a low end-to-end delay and low collision rates compared with the well-known IEEE 802.15.4 MAC protocols widely used in the literature to provide QoS to smart-grid applications.

Thermalization and Dephasing in Collisional Reservoirs.

We introduce a wide class of quantum maps that arise in collisional reservoirs and are able to thermalize a system if they operate in conjunction with an additional dephasing mechanism. These maps describe the effect of collisions and induce transitions between populations that obey detailed balance, but also create coherences that prevent the system from thermalizing. We combine these maps with a unitary evolution acting during random Poissonian times between collisions and causing dephasing. We find that, at a low collision rate, the nontrivial combination of these two effects causes thermalization in the system. This scenario is suitable for modeling collisional reservoirs at equilibrium. We justify this claim by identifying the conditions for such maps to arise within a scattering theory approach and provide a thorough characterization of the resulting thermalization process.

Vehicular intelligent collaborative intersection driving decision algorithm in Internet of Vehicles

Connected and Automated Vehicles (CAV) are regarded as the developing trend of future transportation due to the advantages in terms of increasing traffic throughput, safety, and reducing energy consumption. One of the challenging CAV research problems is intelligent collaborative driving decision in dynamically changing traffic scenarios. The sixth-generation mobile communication technology (6G) plays an important role in attaining CAV intelligent driving in Internet of Vehicles (IoV) by provide significantly higher system performance, higher spectral efficiency, higher reliability, and higher security. This paper formulates the collaborative decision problem of CAVs at unsignalized intersections as a multi-agent reinforcement learning (MARL) problem, where CAVs entering an intersection safely cross the intersection and minimize their traffic time through collaborating to learn a strategy. An efficient and scalable MARL algorithm based on Proximal Policy Optimization (PPO) is applied to the dynamic intersection scenarios, where the parameter-sharing mechanism is used to improve the PPO algorithm to be a multi-agent version. Instead of the global reward, the local reward is used to promote cooperation between agents and achieve scalability. In addition, the action masking mechanism is adopted to improve the learning efficiency by filtering out invalid actions at each step. A joint simulation platform is built to verify the performance of the Parameter-Sharing PPO algorithm (PS-PPO). The simulation results show that the PS-PPO algorithm not only guarantees a low collision rate but also greatly reduces the vehicular travel time. This algorithm promotes cooperation among CAVs and performs well in the task of collaborative traffic flow at unsignalized intersections, especially in high-traffic volume scenarios. It is helpful to improve the safety and efficiency of traffic flow at unsignalized intersections.

Three-dimensional building of anisotropic gold nanoparticles under confinement in submicron capsules.

The low collision rate and contact time of gold nanoparticles (NPs) in solution afford a low welding probability, which hinders their welding structure, orientation, and dimension. Encapsulated anisotropic NPs, gold nanotriangles (AuNTs), were successfully assembled into a three-dimensional structure inside a permeable silica nanocapsule under light illumination to generate localized surface plasmon resonance (LSPR). AuNTs were trapped in the permeable silica nanocapsules and diffused in the nanospace because of copolymer release, which increased the contact probability of AuNTs and promoted the three-dimensional building of AuNTs. Electron energy loss mapping simulations revealed that the obtained three-dimensional AuNT structure exhibited spatially separated multiple LSPR modes with different energies of incident light, which are photophysically attractive beyond the facet-selective chemical growth of NPs, and postmodification for anchoring substances with site-selective attachment to the obtained structure will be applicable to expand the sensing design and class of substances for sensing.

OSSOS. XXVI. On the Lack of Catastrophic Collisions in the Present Kuiper Belt

We investigate different conditions, including the orbital and size–frequency distribution (SFD) of the early Kuiper Belt, that can trigger catastrophic planetesimal destruction. The goal of this study is to test if there is evidence for collisional grinding in the Kuiper Belt that has occurred since its formation. This analysis has important implications for whether the present-day SFD of the cold classical trans-Neptunian objects (TNOs) is a result of collisional equilibrium or if it reflects the primordial stage of planetesimal accretion. As an input to our modeling, we use the most up-to-date debiased OSSOS++ ensemble sample of the TNO population and orbital model based on the present-day architecture of the Kuiper Belt. We calculate the specific impact energies between impactor–target pairs from different TNO groups and compare our computed energies to catastrophic disruption results from smoothed particle hydrodynamics simulations. We explore different scenarios by considering different total primordial Kuiper Belt masses and power slopes of the SFD and allowing collisions to take place over different timescales. The collisional evolution of the Kuiper Belt is a strong function of the unknown initial mass in the trans-Neptunian region, where collisional grinding of planetesimals requires a total primordial Kuiper Belt mass of M > 5 M ⊕, collision speeds as high as 3 km s−1, and collisions over at least 0.5 Gyr. We conclude that presently, most of the collisions in the trans-Neptunian region are in the cratering rather than disruption regime. Given the low collision rates among the cold classical Kuiper Belt objects, their SFD most likely represents the primordial planetesimal accretion.

Learning-Based UAV Path Planning for Data Collection With Integrated Collision Avoidance

Unmanned aerial vehicles (UAVs) are expected to be an integral part of wireless networks, and determining collision-free trajectory in multi-UAV noncooperative scenarios while collecting data from distributed Internet of Things (IoT) nodes is a challenging task. In this article, we consider a path-planning optimization problem to maximize the collected data from multiple IoT nodes under realistic constraints. The considered multi-UAV noncooperative scenarios involve a random number of other UAVs in addition to the typical UAV, and UAVs do not communicate or share information among each other. We translate the problem into a Markov decision process (MDP) with parameterized states, permissible actions, and detailed reward functions. Dueling double deep <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$Q$ </tex-math></inline-formula> -network (D3QN) is proposed to learn the decision-making policy for the typical UAV, without any prior knowledge of the environment (e.g., channel propagation model and locations of the obstacles) and other UAVs (e.g., their missions, movements, and policies). The proposed algorithm can adapt to various missions in various scenarios, e.g., different numbers and positions of IoT nodes, different amount of data to be collected, and different numbers and positions of other UAVs. Numerical results demonstrate that real-time navigation can be efficiently performed with high success rate, high data collection rate, and low collision rate.

Beta‐Dependent Constraints on Ion Temperature Anisotropy in Jupiter’s Magnetosheath

AbstractMost space plasmas are weakly collisional. Due to the low collision rate, the plasmas are routinely out of local thermal equilibrium and often exhibit non‐Maxwellian velocity distributions. The charged particles typically exhibit anisotropy in temperature measurements, that is, distinct temperatures are observed perpendicular and parallel (T⊥i and T||i) to the local magnetic field. Numerous prior studies have shown that for increasing values of parallel ion beta (β||i), the range of ion temperature anisotropy (Ri = T⊥i/T||i) values becomes narrower. Conventionally, this behavior has been attributed to the actions of kinetic microinstabilities. This study is the first to explore such β||i‐dependent limits on ion temperature anisotropy in Jupiter's magnetosheath. We use linear Vlasov theory to compute contours of constant growth rate for different instability thresholds, which closely align with the limits of the data distribution, supporting that these instabilities are acting to limit extremes of ion temperature anisotropy in the Jovian magnetosheath.

Conditions for obtaining positronium Bose–Einstein condensation in a micron-sized cavity

The quest for making a triplet positronium (Ps) Bose–Einstein condensate confined in a micron-sized cavity in a material such as porous silica faces at least three interrelated problems: (1) About 10^7 spin polarized Ps atoms must be injected into a small cavity within a porous solid material without vaporizing it. (2) It is known that Ps atoms confined in 30–100 nm diameter cavities in porous silica do not remain in the gas phase, but become stuck to the cavity walls at room temperature (Cooper et al., Phys. Rev. B 97:205302, 2018). (3) Cooling a gas of Ps atoms to cryogenic temperatures by energy exchange with the walls would be a very slow process (Saito and Hyodo, in: Surko, Gianturco (eds) New Directions in Antimatter Chemistry and Physics, Springer Dordrecht, Netherlands, 2001) because of the relatively low collision rate with the walls and the large mismatch between the masses of the Ps and the wall atoms. A possible solution of these difficulties is presented, based on cooling the implanted positrons in an isotopically pure diamond single crystal target, subsequent saturating of the wall Ps coverage so that a substantial portion of the Ps will be in the gaseous state, and thermalizing the gas-phase Ps via collisions with the low effective mass wall Ps. A design process for the target material is outlined as well, including preliminary results in porous cavity fabrication using focused ion beam milling.Graphical abstract