This paper presents a tightly-coupled LiDAR-Visual-Inertial Odometry (LIVO) system that integrates both LIO and VIO subsystems. The system jointly estimates the state by fusing LiDAR or visual data with Inertial Measurement Units (IMUs). It employs point-to-mesh tracking to optimize LiDAR poses and leverages intensity information from LiDAR point clouds to refine camera pose estimation. The optimized camera pose, derived from VIO, plays a crucial role in texture mapping and 3D geometry synthesis (3DGS) rendering. Our experiments demonstrate a significant improvement in average Peak Signal-to-Noise Ratio (PSNR) compared to existing methods, including R3LIVE, SR-LIVO, and FAST-LIVO. Furthermore, the system features a real-time mapping module implemented on the GPU, utilizing Truncated Signed Distance Function (TSDF) fields for global map maintenance and the Marching Cubes algorithm for mesh extraction. This approach ensures rapid and precise tracking and reconstruction capabilities. Additionally, our system supports real-time remeshing of the global map upon detecting loop closures, thereby enhancing the robustness and accuracy of the overall SLAM process.

Nowadays, backscatter, radio-frequency (RF) energy harvesting (EH), and cognitive radio (CR) technologies have been widely applied in Internet of Things (IoT) to address the issues of energy supply and spectrum scarcity. This article focuses on the throughput maximization problem of backscatter-aided wireless-powered heterogeneous CR networks (WPHetCRNs), where two types of secondary transmitters (STs), i.e., the STs with backscatter units (STBs) and the STs with RF-EH units (STEs), coexist. The STBs operate in the ambient backscatter (AB) mode, and the STEs operate in the harvest-then-transmit (HTT) mode. Inspired by the potential benefits of cooperations between different users, we propose a backscatter-aided cooperative relay transmission (BaCRT) strategy to improve the sum-throughput of the secondary users (SUs). The main idea is that when the licensed spectrum of the primary users (PUs) is busy, the STBs first help to relay the primary data via the passive relay mode, and then transmit the secondary data via the AB mode, while the STEs harvest energy in the HTT mode. With the help of relaying, the target throughput of the PUs could be met in shorter duration and the licensed spectrum could become idle more quickly. When the licensed spectrum becomes idle, the STEs transmit data in the HTT mode. The goal of this article is to identify the optimal time allocation among the passive relay mode, AB mode, and data transmission of HTT mode that maximizes the sum-throughput of the SUs. To reach this goal, we first investigate the single-ST case for each type and derive the closed-form solution of the optimal time allocation. We then extend to the multiple-ST case for each type, where three scenarios are classified with respect to the fairness issue of the STBs. We prove that the sum-throughput maximization problem is convex in each scenario and employ the block coordinate descent and gradient descent iterative algorithms to solve the problem. Numerical results show that the proposed BaCRT strategy significantly improves the sum-throughput of the SUs compared with other strategies.

The estimation and reconstruction of complex fluids like wind turbine wake are challenging problems, which require approximating high-dimensional, nonlinear dynamic systems with a limited number of physical measurements. Current works mainly focus on the reduced-order approximation for fluid dynamics like dynamic mode decomposition, or linear stochastic estimation for static mapping. In this article, the problem of yaw-controlled wind turbine wake dynamics approximation and reconstruction are addressed. A Koopman-linear flow estimator is designed, which forms a control-oriented linear dynamic state-space model. The full wake flow is first approximated on its dynamic performances and then reconstructed from low-order physical states. The novelty mainly yields on the added yaw excitation that a control-oriented dynamic model is obtained. The Kalman filter is also involved that measured data in future time could be embedded while the measurement noise is filtered. At last, static and dynamic estimation tests verify the effectiveness of the proposed estimator.

An efficient multi-threaded memory allocator has great impact on performance of applications with frequent memory allocation and deallocation operations. Currently, most popular memory allocators ignore the difference between thread local and shared memory, and manage them in a unified manner, which cannot make good use of the spatiotemporal locality of memory access. To solve this problem, this paper proposes a hybrid and lock-free multi-threaded memory allocator, named HMalloc. The allocator separates local memory from shared memory. There is no false sharing and lock contentions in local memory allocation and deallocation process. Moreover, coalescence-free is used to optimize this process. Further, a flag-based shared memory allocation and deallocation method is proposed to achieve lock-free shared memory management. Experimental results show that HMalloc can achieve significant performance improvement when compared with existing well-known memory allocators.

The intra-cellular calcium signaling pathways of a neuron depends on both biochemical reactions and diffusions. Some quasi-isolated compartments (e.g., spines) are so small and calcium concentrations are so low that one extra molecule diffusing in by chance can make a nontrivial difference in concentration (percentage-wise). These rare events can affect dynamics discretely in such a way that they cannot be evaluated by a deterministic and continuous simulation. Stochastic models of such a system provide a more detailed understanding of these systems than existing deterministic models because they capture their behavior at a molecular level. Our research focuses on the development of a high performance parallel discrete event simulation environment, Neuron Time Warp (NTW), which is intended for use in the parallel simulation of stochastic reaction-diffusion systems such as intra-calcium signaling. NTW is integrated with NEURON, a simulator which is widely used within the neuroscience community. We simulate two models, a calcium buffer and a calcium wave model. The calcium buffer model is employed in order to verify the correctness and performance of NTW by comparing it to a sequential deterministic simulation in NEURON. We also derived a discrete event calcium wave model from a deterministic model using the stochastic $\text{IP}_{3}\text{R}$IP3R structure.

This paper describes a parallel discrete event simulator, Neuron Time Warp-Multi Thread (NTW-MT), developed for the simulation of reaction diffusion models of neurons. The simulator was developed as part of the NEURON project and is intended to be included in NEURON. It relies upon a stochastic discrete event model developed for chemical reactions. NTW-MT is optimistic and thread-based, in which communication latency among threads within the same process is minimized by pointers. We investigate the performance of NTW-MT on a reaction-diffusion model for the transmission of calcium waves in a neuron. Calcium plays a fundamental role in the second messenger system of a neuron. However, the mechanism by which calcium waves are transmitted is not entirely understood. Stochastic models are more realistic than deterministic models for small populations of ions such as those found in apical dendrites. To be more precise, we simulate a stochastic discrete event model for calcium wave propagation on an unbranched apical dendrite of a hippocampal pyramidal neuron. We examine the performance of NTW-MT on this calcium wave model and compare it to the performance of (1) a process based optimistic simulator and (2) a threaded simulator which uses a single priority (SQ) queue for each thread. Our multi-threaded simulator is shown to achieve superior performance to these simulators.