Perception Enhanced Deep Deterministic Policy Gradient for Autonomous Driving in Complex Scenarios

Autonomous driving has witnessed rapid advancement;however,ensuring safe and efficient driving in intricate scenarios remains a critical challenge.In particular,traffic roundabouts bring a set of challenges to autonomous driving due to the unpredictable entry and exit of vehicles,susceptibility to traffic flow bottlenecks,and imperfect data in perceiving environmental information,rendering them a vital issue in the practical application of autonomous driving.To address the traffic challenges,this work focused on complex roundabouts with multi-lane and proposed a Perception EnhancedDeepDeterministic Policy Gradient(PE-DDPG)for AutonomousDriving in the Roundabouts.Specifically,themodel incorporates an enhanced variational autoencoder featuring an integrated spatial attention mechanism alongside the Deep Deterministic Policy Gradient framework,enhancing the vehicle’s capability to comprehend complex roundabout environments and make decisions.Furthermore,the PE-DDPG model combines a dynamic path optimization strategy for roundabout scenarios,effectively mitigating traffic bottlenecks and augmenting throughput efficiency.Extensive experiments were conducted with the collaborative simulation platform of CARLA and SUMO,and the experimental results show that the proposed PE-DDPG outperforms the baseline methods in terms of the convergence capacity of the training process,the smoothness of driving and the traffic efficiency with diverse traffic flow patterns and penetration rates of autonomous vehicles(AVs).Generally,the proposed PE-DDPGmodel could be employed for autonomous driving in complex scenarios with imperfect data.

A Novel Quantization and Model Compression Approach for Hardware Accelerators in Edge Computing

Massive computational complexity and memory requirement of artificial intelligence models impede their deploy-ability on edge computing devices of the Internet of Things(IoT).While Power-of-Two(PoT)quantization is pro-posed to improve the efficiency for edge inference of Deep Neural Networks(DNNs),existing PoT schemes require a huge amount of bit-wise manipulation and have large memory overhead,and their efficiency is bounded by the bottleneck of computation latency and memory footprint.To tackle this challenge,we present an efficient inference approach on the basis of PoT quantization and model compression.An integer-only scalar PoT quantization(IOS-PoT)is designed jointly with a distribution loss regularizer,wherein the regularizer minimizes quantization errors and training disturbances.Additionally,two-stage model compression is developed to effectively reduce memory requirement,and alleviate bandwidth usage in communications of networked heterogenous learning systems.The product look-up table(P-LUT)inference scheme is leveraged to replace bit-shifting with only indexing and addition operations for achieving low-latency computation and implementing efficient edge accelerators.Finally,comprehensive experiments on Residual Networks(ResNets)and efficient architectures with Canadian Institute for Advanced Research(CIFAR),ImageNet,and Real-world Affective Faces Database(RAF-DB)datasets,indicate that our approach achieves 2×∼10×improvement in the reduction of both weight size and computation cost in comparison to state-of-the-art methods.A P-LUT accelerator prototype is implemented on the Xilinx KV260 Field Programmable Gate Array(FPGA)platform for accelerating convolution operations,with performance results showing that P-LUT reduces memory footprint by 1.45×,achieves more than 3×power efficiency and 2×resource efficiency,compared to the conventional bit-shifting scheme.

Multi-Scale X-Ray Imaging Technologies for Rechargeable Batteries

The rapid advancement in electric vehicles and electrochemical energy storage technology has raised the demands placed on rechargeable batteries.It is essential to comprehend the operational principles and degradation mechanisms of batteries across multiple scales to propel the research on rechargeable batteries for the next generation forward.Microstructure,phase information,and lattice of energy materials in both two dimensions and three dimensions can be intuitively obtained through the utilization of x-ray imaging techniques.Additionally,x-ray imaging technology is increasingly gaining attention due to its non-destructive nature and high penetrative capability,enabling in situ experiments and multi-scale spatial resolution.In this review,we initially overview the basic principles and characteristics of several key x-ray imaging technologies.Each x-ray imaging technology is tailored to specific application scenarios.Furthermore,examples of multi-scale implementations of x-ray imaging technologies in the field of rechargeable batteries are discussed.This review is anticipated to augment the comprehension of readers for x-ray imaging techniques as well as to stimulate the development of novel concepts and approaches in rechargeable battery research.

Experimental investigation on degradation mechanism of membrane electrode assembly at different humidity under automotive protocol

Humidity can affect the attenuation of MEA(membrane electrode assembly), however, the relationship between humidity and MEA decays is complex and ambiguous in realistic application. Herein, we design a simulating automotive protocol, performed on five single fuel cells under RH(relative humidity) 100%,RH 80%, RH 64%, and RH 40%, RH 10%, respectively, to study the relationship of MEA decays and humidity and suggest optimized humidity range to extend the durability. With the electrochemical impedance spectroscopy, cyclic voltammetry, X-ray fluorescence, X-ray diffraction, transmission electron microscope, X-ray photoelectron spectroscopy, the four degradation mechanisms about catalyst layer, including Pt dissolution, Pt coarsening, carbon corrosion and ionomer degradation, are observed. Pt coarsening and carbon corrosion are accelerated by higher water content at high humidity. Ionomer degradation and Pt dissolution are enhanced in low humidity. With the linear sweep voltammetry, ion chromatography,nuclear magnetic resonance, tensile test and scan electron microscope, chemical and mechanical degradation in proton exchange membrane are all observed in these five fuels. Chemical degradation, characterized by membrane thinning and more fluoride loss, occurred markedly in RH 10%. Mechanical degradation, characterized by the non-uniformity thickness and bad mechanical properties, is more pronounced in RH 100%, RH 80%, RH 64%. These two degradations are in a moderate level in RH 40%. The research suggests that the RH range from 64% to 40% is conductive to mitigate the degradation of MEAs operated in automotive applications.

Size effect on light propagation modulation near band edges in one-dimensional periodic structures

Periodic photonic structures can provide rich modulation in propagation of light due to well-defined band structures.Especially near band edges,light localization and the effect of near-zero refractive index have attracted wide attention.However,the practically fabricated structures can only have finite size,i.e.,limited numbers of periods,leading to changes of the light propagation modulation compared with infinite structures.Here,we study the size effect on light localization and near-zero refractive-index propagation near band edges in one-dimensional periodic structures.Near edges of the band gap,as the structure's size shrinks,the broadening of the band gap and the weakening of the light localization are discovered.When the size is small,an added layer on the surface will perform large modulation in the group velocity.Near the degenerate point with Dirac-like dispersion,the zero-refractive-index effects like the zero-phase difference and near-unity transmittance retain as the size changes,while absolute group velocity fluctuates when the size shrinks.

Uncovering the degradation mechanism induced by ion-diffusion kinetics in large-format lithium-ion pouch cells

Battery electrochemistry in an actual cell is a complicated behavior influenced by the current density,uniformity,and ion-diffusion distance,etc.The anisotropism of the lithiation/delithiation degree is usually inevitable,and even worse,due to a trend of big-size cell design,typically such as 4680 and blade cells,which accelerated a battery failure during repeat lithiation and delithiation of cathodes.Inspire by that,two big-size pouch cells with big sizes,herein,are selected to reveal the ion-diffusion dependency of the cathodes at different locations.Interestingly,we find that the LiCoO_(2) pouch cell exhibits ~5 A h loss after 120 charge-discharge cycles,but a 15 A h loss is verified in a LiNixMnyCO_(1-x)-yO_(2)(NCM) cell.Synchrotron-based imaging analysis indicates that higher ion-diffusion rates in the LiCoO_(2)than that in the LiNixMnyCO_(1-x)-yO_(2)is the determined factor for the anisotropic cathode fading,which is responsible for a severe mechanical issue of particle damage,such as cracks and even pulverization,in the cathode materials.Meanwhile,we verify the different locations at the near-tab and bottom of the electrode make it worse due to the ion-diffusion kinetics and temperature,inducing a spatially uneven electrochemistry in the big-size pouch cell.The findings give an in-depth insight into pouch cell failure and make a guideline for high-energy cell design and development.

缩水甘油酯型自固化环氧树脂的高效合成

以四甲基氢氧化铵五水合物(THP)作为催化剂,环氧氯丙烷(ECH)与2,5-二羟基对苯二甲酸(DHTA)为原料,制备了具有自固化特性的缩水甘油酯型环氧树脂。考察了温度、溶剂、催化剂等因素对制备过程的影响,用傅里叶变换红外光谱对反应过程中羧基转化情况进行了表征。结果表明,在90℃条件下,反应物与催化剂摩尔比n(DHTA)∶n(ECH)∶n(THP)=1∶17.5∶0.12时,DHTA内羧基可在2 h内完成酯化且温升较小。使用核磁碳谱及氢谱表征树脂,确认产物结构。使用凝胶渗透色谱对树脂相对分子质量分布进行表征,由于反应物平均官能度较低,产物内主要为小分子的二聚体及四聚体。该路线反应时间短、反应温升小、废液排放少且树脂稳定性较好,实现了具有工业基础的低介电缩水甘油酯型自固化环氧树脂的绿色高效合成。

三聚氰胺甲醛泡沫增韧改性研究进展

三聚氰胺甲醛泡沫是一种具有吸音、耐热、保温等性能的本征型阻燃泡沫,燃烧时火焰易自熄且不产生流滴,可用于建筑、农业、航空、交通运输和电子信息等领域,是一种发展前景良好的轻质材料。但是,由于柔性链段短、交联密度大,三聚氰胺甲醛泡沫脆性往往较大,限制了其实际应用。为此,从三聚氰胺甲醛树脂基体改性和泡沫结构调控2个方面综述了国内外关于三聚氰胺甲醛泡沫增韧改性的研究进展,分析了不同改性方法的原理及优缺点,并对增韧改性三聚氰胺甲醛泡沫的发展进行了展望,旨在为合成高性能的三聚氰胺甲醛泡沫提供有益的借鉴和指导。

Compound Danshen tablets downregulate amyloid protein precursor mRNA expression in a transgenic cell model of Alzheimer's disease Effects and a comparison with donepezil

After gene mutation, the pcDNA3.1/APP595/596 plasmid was transfected into HEK293 cells to establish a cell model of Alzheimer＇s disease. The cell model was treated with donepezil or compound Danshen tablets after culture for 72 hours. Reverse transcription-PCR showed that the mRNA expression of amyloid protein precursor decreased in all groups following culture for 24 hours, and that there was no significant difference in the amount of decrease between donepezil and compound Danshen tablets. Our results suggest that compound Danshen tablets can reduce expression of the mRNA for amyloid protein precursor in a transgenic cell model of Alzheimer＇s disease, with similar effects to donepezil.

Parameter Optimization of Interval Type-2 Fuzzy Neural Networks Based on PSO and BBBC Methods

Interval type-2 fuzzy neural networks(IT2FNNs)can be seen as the hybridization of interval type-2 fuzzy systems(IT2FSs) and neural networks(NNs). Thus, they naturally inherit the merits of both IT2 FSs and NNs. Although IT2 FNNs have more advantages in processing uncertain, incomplete, or imprecise information compared to their type-1 counterparts, a large number of parameters need to be tuned in the IT2 FNNs,which increases the difficulties of their design. In this paper,big bang-big crunch(BBBC) optimization and particle swarm optimization(PSO) are applied in the parameter optimization for Takagi-Sugeno-Kang(TSK) type IT2 FNNs. The employment of the BBBC and PSO strategies can eliminate the need of backpropagation computation. The computing problem is converted to a simple feed-forward IT2 FNNs learning. The adoption of the BBBC or the PSO will not only simplify the design of the IT2 FNNs, but will also increase identification accuracy when compared with present methods. The proposed optimization based strategies are tested with three types of interval type-2 fuzzy membership functions(IT2FMFs) and deployed on three typical identification models. Simulation results certify the effectiveness of the proposed parameter optimization methods for the IT2 FNNs.

Thermal Shock-Activated Spontaneous Growing of Nanosheets for Overall Water Splitting

Nanomaterials based on nickel foam(NF) have been widely applied in energy conversion and storage field.Traditional synthesis methods such as hydrothermal method which is dangerous and high cost limited the scalable developments.Herein,we report a fast,simple,and low-cost synthesis method of nanomaterials based on NF by Joule-heating and water soaking treatment.Thin carbon-coated CoS on NF(NF-C/CoS) was synthesized by Joule-heating for a few seconds with rapid cooling.And then,NF-C/CoS/NiOOH with core-shell heterostructure was fabricated by soaking treatment of NF-C/CoS in water on which NiOOH nanosheets grew spontaneously.The formation mechanism is proposed that the coordination complex precursor converts into C/CoS on NF driven by Joule-heating,and the nickel on the surface of NF is activated to form metastable nickel simultaneously.The metastable nickel reacting with water leads to the formation of NiOOH,and the induction of CoS makes NiOOH grow continuously.This synthesis technology provides a new route to manufacture NF-based nanostructures,and the as-fabricated NF-C/CoS/NiOOH exhibits great potential as electrocatalyst for oxygen evolution reaction and hydrogen evolution reaction.

A bifunctional perovskite oxide catalyst:The triggered oxygen reduction/evolution electrocatalysis by moderated Mn-Ni co-doping

ABO_(3)-type perovskite oxides(e.g.,LaCoO_(3))with flexible and adjustable A-and B-sites are ideal model catalysts to unravel the relationship between the electronic structure and electrocatalytic activity(e.g.,oxygen reduction/evolution reactions,ORR/OER).It has been well understood in our recent work that the secondary metal dopant at B-site(e.g.,Mn in LaMn_(x)Co_(1-x)O_(3))can regulate the electronic structure and improve the ORR/OER activity.In this work,the Mn-Ni pairs are employed as the dual dopant in LaMn_(x)Ni_(y)Co_(z)O_(3)(x+y+z=1)catalysts toward bifunctional ORR and OER.The structure-property relationships between the triple metal B-site(Mn,Ni and Co)and the electrochemical performance are particularly investigated.Compared to the individual Mn doping(e.g.,LaMnCoO3(Mn:Co=1:3)catalyst),the dual Mn-Ni doping significantly improves the ORR mass activity@0.8 V by 1.54 times;meanwhile,the OER overpotential@10 mA cm^(-2) is reduced from 420 to 370 mV,and the OER current density at 1.55 V is increased by 2.43 times.Reasonably,the potential gap between EDRR@-1 mA cm^(-2) and EDER@10 mA cm^(-2) is achieved as only 0.76 V by using the optimal LaMn_(x)Ni_(y)Co_(z)O_(3)(x:y:z=1:2:3)catalyst.It is revealed that the dual Mn-Ni dopant efficiently optimizes electron structures of the LaMnNiCoO_(3)(1:2:3)catalyst,which not only decreases the e_(g) orbital electron number,but also modulates the O 2 p-band closer to the Femi level,accounting for the enhanced bifunctional activity.

Construction of polysulfides defense system for greatly improving the long cycle life of metal sulfide anodes for sodium-ion batteries

Metal sulfides are promising anode materials for sodium ion batteries(SIBs)due to their high theoretical specific capacity and abundant source.Nevertheless,significant challenges,including large volume change,sluggish Na^(+)transport kinetics and polysulfides intermediates,have greatly affect their long cycle stability.Unfortunately,the majority of current studies only focus on the first two aspects,but lack of sufficient attention and insights into the effect of polysulfides intermediates.Here,a porous of CoS_(x)(P-CoS_(x))electrode material is fabricated as an example to investigate the influence of polysulfides on its cycling performance.The results show that polysulfides cause a slight loss of reversible capacity during the battery cycling,while the failure of the battery is due to its significant fluctuations in reversible capacity after extensive cycles.Detailed analyses demonstrate that the intense fluctuation in capacity originates from the faster growth of dendrites caused by the reaction of sodium polysulfides with sodium foil and/or the reaction of elemental sulfur with sodium foil to penetrate the separator,resulting in a local short circuit.To suppress these undesirable side reaction,N,S co-doped porous carbon tubes(N,S-PC)rich in C–S and C–N bonds have been added to adsorb polysulfides and alleviate their reaction with sodium foil.As a result,the capacity of the P-CoS_(x) electrode with N,S-PC(P-CoS_(x)/N,S-PC)remains stable without significant fluctuations for 1000 cycles,which is much better than that of the pure P-CoS_(x) electrode(intense fluctuation in capacity after 320 cycles).Our work offers insights into the crucial influence of polysulfides on the cycle performance of the P-CoS_(x) anode and provides a feasible strategy to prolong the cycle life of metal sulfide anode for SIBs.

Unraveling the reaction mechanism of low dose Mn dopant in Ni(OH)_(2) supercapacitor electrode

Mn doping is deemed as a promising strategy to improve the electrochemical performance of the a-Ni(OH)_(2)battery-type supercapacitor electrode.However,the internal structure evolution,the pathways and the dynamics of the proton/intercalated anion migration,as well as the functioning mechanism of Mn dopant to stabilize the layered structure during cycles remain unclear.Here,we unveil that irreversible oxidization of Mn^(3+)at the initial CV cycles,which will remain as Mn^(4+)in the NiO_(2)slabs after the first oxidization to effectively suppress the phase transformation fromα-Ni(OH)_(2)/γ-NiOOH toβ-Ni(OH)_(2)/β-NiOOH and further maintain the structural integrity of electrode.With a synergistic combination of theoretical calculations and various structural probes including XRD and^(2)H MAS solid state NMR,we decode the structure evolution and dynamics in the initial CV(cyclic voltammetry)cycles,including the absorption/desorption of hydrogen containing species,migration of intercalated anions/water molecules and the change of interlayer space.This present work elucidates a close relationship between doping chemistry and structural reliability,paving a novel way of reengineering supercapacitor electrode materials.

Visible light-driven oxidant-free dehydrogenation of alcohols in water using porous ultrathin g-C_(3)N_(4)nanosheets

Graphitic carbon nitride(g-C_(3)N_(4)) is a fascinating photocatalyst for solar energy utilization in photo-catalysis.Nevertheless,it often suffers from moderate photo-catalytic activity due to its low specific surface area and fast recombination rate of photogenerated electrons upon photo-excitation.Herein,we overcome the bottlenecks by constructing a porous g-C_(3)N_(4) nanosheet(PCNS)through a simple thermal oxidation etching method.Benefited from its porous layer structure,the obtained PCNS exhibits large specific surface area,efficient separation of photogenerated charge carriers,as well as high exposure of active sites.As a result,it is robust and universal in visible light-driven dehydrogenation of alcohols in water under oxidant-free condition.Almost quantitative yields(>99%)of various valuable carbonyl compounds were obtained over PCNS,while bulk g-C_(3)N_(4) was far less efficient.Moreover,the photo-catalyst was highly stable and could be facilely recovered from the aqueous system for efficient reuse.The easy preparation and excellent performance made PCNS a promising and competitive photocatalyst for the solar applications.

Tuning the phase evolution pathway of LiNi_(0.5)Mn_(1.5)O_(4) synthesis from binary intermediates to ternary intermediates with thermal regulating agent

Transition metal cation ordering is essential for controlling the electrochemical performance of cubic spinel LiNi_(0.5)Mn_(1.5)O_(4)(LNMO),which is conventionally adjusted by optimizing the high temperature sintering and annealing procedures.In this present work,multiple characterization techniques,including 6,7Li NMR,XRD and HRTEM,have been combined to trace the phase transformation and morphology evolution during synthesis.It has been illustrated that simultaneous formation of LiMn_(2)O_(4)(LMO)and LiNiO_(2)(LNO)binary oxides and their conversion into highly reactive LixNi^(3+)_(y)Mn_(3.5+)_(z)O ternary intermediate is a thermal dynamically difficult but crucial step in the synthesis of LNMO ternary oxide.A new strategy of modifying the intermediates formation pathway from binary mode to ternary mode using thermal regulating agent has been adopted.LNMO synthesized with thermal regulating agent exhibits supreme rate capability,long-cycling performance(even at elevated temperature)and excellent capacity efficiency.At a high rate of 100 C,the assembled battery delivers a discharge capacity of 99 mAh g^(-1).This study provides a way to control the formation pathway of complex oxides using thermal regulating agent.

Speed-assigned Position Tracking Control of SRM With Adaptive Backstepping Control

A novel speed-assigned method is applied to the position tracking control of switched reluctance motor(SRM).A speed control freedom can be drawn into the position control through speed assignment. Adaptive backstepping control is used to design the position controller for the SRM. The accuracy of position tracking of the SRM can be enhanced with speed assignment. A disturbance observer is further designed to enhance the estimation accuracy of the unknown load torque. Simulation results certify that the design scheme is right and effective.

Co-Processing Sewage Sludge in Cement Kiln in China

Sewage sludge is the by-product from municipal waste water treatment plant and is highly polluted. How to treat the SS in a solid environmental friendly way is strictly concerned in China. In this article, the SS situation and also the treatment methods in China have been introduced. The advantage and some related issues of co-processing SS in cement kiln have been discussed. The technical model and projects of Huaxin cement for co-processing SS in cement kiln also have been introduced.

Feature Selection for Image Classification Based on a New Ranking Criterion

In this paper, a feature selection method combining the reliefF and SVM-RFE algorithm is proposed. This algorithm integrates the weight vector from the reliefF into SVM-RFE method. In this method, the reliefF filters out many noisy features in the first stage. Then the new ranking criterion based on SVM-RFE method is applied to obtain the final feature subset. The SVM classifier is used to evaluate the final image classification accuracy. Experimental results show that our proposed relief- SVM-RFE algorithm can achieve significant improvements for feature selection in image classification.