Search

Real-time detecting objects on captured images on UAV (Unmanned Aerial Vehicle) platforms, rather than barely transmitting images back to supporting equipment for post-processing, is a core requirement for advanced UAV applications. However, due to limited computing capacity and memory of UAV platforms, it is very challenging to deploy real-time detection models on them. In addition, there are more small objects in aerial images, which makes it more difficult to detect accurately. To solve these problems, this paper brings dense connection to Yolo(You Only Look Once)v3 network, and proposes Yolo-LiteDense model. The backbone of Yolo-LiteDense is densely connected, which improves the performance of feature extraction. Then, we enforce channel pruning to Yolo-LiteDense model by pruning less informative channels with less scaling factors. After pruning, parameters and weight size of the model are compressed significantly, and inference time is also shortened. Evaluation results on VisDrone2018-DET show that parameters and weight size of Yolo-LiteDense are 83% and inference time is 30% less than Yolov3-SPP with comparable average precision. In addition, this paper also proposes the lighter version of Yolo-LiteDense, Yolo-DenseNano. Parameters and weight size of Yolo-LiteDense are 70% less than Yolov3-tiny with 2.68 times greater average precision.

Recently, automatic visual data understanding from drone platforms becomes highly demanding. To facilitate the study, the Vision Meets Drone Object Detection in Image Challenge is held the second time in conjunction with the 17-th International Conference on Computer Vision (ICCV 2019), focuses on image object detection on drones. Results of 33 object detection algorithms are presented. For each participating detector, a short description is provided in the appendix. Our goal is to advance the state-of-the-art detection algorithms and provide a comprehensive evaluation platform for them. The evaluation protocol of the VisDrone-DET2019 Challenge and the comparison results of all the submitted detectors on the released dataset are publicly available at the website: http: //www.aiskyeye.com/. The results demonstrate that there still remains a large room for improvement for object detection algorithms on drones.

A 300 V to 600 V 100 kW SiC MOSFET based one-cell switched tank converter (STC) is developed as a bidirectional dc-dc power transfer stage between the vehicle battery and the DC-link side of the vehicle dc-ac inverter. A continuous half-load 50 kW and short-period full-load 100 kW operation is targeted. Working principles of the proposed topology are analyzed. Design of the key components such as SiC MOSFET power modules, AC resonant capacitor and inductor is presented. A 100 kW prototype has been assembled and tested. An energy-efficient test platform is designed. The power density of the main power processing part is around 41.7 kW/L. The tested peak and full-load efficiencies are about 98.7% and 97.35%, respectively. The thermal performance has also been evaluated. Both the tested electrical and thermal results are consistent with the theoretical design.

This paper compares three different dc-dc topologies, i.e. boost converter, three-level flying capacitor multilevel converter (FCMC)and one-cell switching tank converter (STC)for a 100 kW electric vehicle power electronic system. This bidirectional dc-dc converter targets 300 V-600 V voltage conversion. Total semiconductor loss index (TSLI)has been proposed to evaluate topologies and device technologies. The boost converter and one-cell STC have been fairly compared by utilizing this index. The simulation results of a 100 kW one-cell STC working at zero current switching (ZCS)mode have been provided. A 100 kW hardware prototype using 1200 V 600 A SiC power module has been built. The estimated efficiency is about 99.2% at 30 kW, 99.13% at half load, and 98.64% at full load. The power density of the main circuits is about 42 kW/L.

This paper presents a 100kW one-cell switched-tank converter (STC) for electric vehicle (EV) application. A new evaluation method that evaluates different converter topologies has been proposed in this paper to show the advantages of the STC over the boost converter and 3-level flying capacitor multilevel (FCML) converter. Both non-interleaved (1-phase) and interleaved (2-phase and 3-phase) operation of the STC have been analyzed. The analytical study shows that it is difficult to achieve the optimum design of the passive components such as input and output capacitors in 1-phase converter because of the high RMS current flowing through them. This means the passive components need to be over-designed in order to meet the current stress requirement. For instance, the designed capacitance of input capacitor is several times of the required value, which leads to bulky capacitor size. Therefore, this paper evaluates the potentials of using 2-phase and 3-phase interleaved operation to address this issue. Two operation modes, zero-voltage switching (ZVS) mode and zero-current switching (ZCS) mode, are evaluated to show the ZCS operation mode is more suitable for the presented converter with interleaved operation. By using the interleaving concept, the predicted 100kW 3-phase interleaved converter can achieve 60% size reduction based on the 1-phase converter design. And the predicted power density of the 3-phase interleaved STC can achieve 115kW/L power density. Simulation results are provided to validate the theoretical analysis. Both 1-phase and 3-phase 100kW prototypes under developing are shown in this paper.

A new Switched Tank Converter (STC) with Partial Power Voltage Regulation (PPVR) is introduced for electric vehicle applications. It offers a flexibly adjustable conversion ratio for a wide-range battery voltage. The unregulated stage is modularized with resonant switched capacitor converters. All device voltage stresses are equal to the input voltage. The efficient unregulated stage processes the bulk of system power and the less efficient regulated stage processes a small amount of power. Thus, high overall efficiency can be achieved. A new index called Semiconductor Loss Index (SLI) is defined as a function of semiconductor die area to evaluate different power topologies. It is helpful to analyze the impact of each switch die area, different operated output power, switching frequency, and conversion ratio on the total device power loss. With the same total semiconductor die area, the proposed converter can achieve less than 1/3 device total power loss compared with boost converter. A 4-kW 1200V output converter prototype with 200V~400V input voltage range is developed, which combines a 6 times conversion ratio STC with a PPVR buck converter. The overall full-load efficiency of designed 4-kW converter can reach to 97.71%. Simulation, prototype and experiment results are presented to verify the validity of the proposed converter.

Multilevel modular resonant switched-capacitor converter can achieve either zero-current switching (ZCS) or zero-voltage switching (ZVS) by utilizing different converter control strategies. This paper presents a comprehensive way to compare the root mean square (RMS) value of current flowing through switching devices in both ZCS operation and ZVS operation. The study shows that with appropriate converter parameter design, the ZVS operation allows the RMS value of switch current at most 10% lower than that in ZCS operation. Therefore, the converter operating at ZVS mode has the potential to achieve higher efficiency comparing to the converter that operates at ZCS mode due to less semiconductor conduction loss. Furthermore, the ZVS operation can reduce the power loss due to MOSFET output capacitance. A 6x converter with 54V input voltage, 9V output voltage and 600W power rating is used as an example to show the detailed design procedure. Simulation results are provided to verify the theoretical analysis. Also, a 600W lab prototype that has 6 to 1 voltage conversion ratio has been built to verify the theoretical analysis.

In this paper, an new converter that includes isolated composite resonant multilevel converter (ICRMC) that uses composite converter concept and partial power voltage regulator (PPVR) has been proposed for telecommunication application. The proposed converter can achieve high efficiency at nominal operating point since it takes minimum effort to regulate the output voltage at this point. A comparative study shows that with the proposed two operation modes of ICRMC, the proposed converter has the best capability to maintain lowest total semiconductor power stress among the existing state of the art solutions when the input voltage varies from 36V to 60V. Furthermore, zero current switching (ZCS) can be always achieved on the switching devices in ICRMC under different operating points. A 600W converter has been simulated to validate the theoretical analysis. The estimated peak efficiency can be as high as 97.65%. Fully debugged prototype and experimental results are provided in this paper.

This paper presents an instantaneous pulse power compensator (IPPC) concept to achieve ultra-high density with zero voltage ripple and zero dc-link capacitor for high-power-density single-phase inverter application. Compared with traditional 2ndorder method with various power decoupling topologies, IPPC proposed a totally new theoretical analysis from a pulse-power point of view. For dc-link voltage ripple evaluation, traditional theoretical analysis is not suitable for small-capacitance power decoupling applications, however, pulse power analysis can be applied in the whole range. From a view of state-of-the-art on power decoupling methods, a comparison figure with voltage ripple and dc-link capacitance is given to show the limitation of traditional power decoupling method and improvement of IPPC method on minimizing dc-link capacitance. One case study is presented with full-bridge topology with simulation results. Furthermore, one Silicon Carbide (SiC) based inverter and the IPPC based circuit are built. Experimental results verify the proposed idea very well.

This paper presents two novel single-phase resonant multilevel modular boost inverters based on resonant switched capacitor cells and a partial power processed voltage regulator. Compared with other multilevel boost inverters applied in PV systems, one remarkable advantage of the proposed topologies is that the bulky AC filtering inductor is replaced by a smaller-size one in the partial power processed buck converter. Constant duty cycle PWM method is attractive for the multilevel inverter controller design. GaN Enhancement Mode Power Transistors help both the modular resonant switched capacitor cells and the full-bridge unfolder be realized in a small size with high power density. The clamp capacitors in the resonant switched capacitor cells effectively alleviate the switch voltage spikes. These two inverter topologies are analyzed and simulated in PLECS. Simulation results verify the validity of boost inverter function. Stress analysis shows that the inverter has relatively small total normalized switch conduction power stress and total normalized switch stress ratio. Relative total semiconductor chip area comparison results reflect that the proposed topology achieves more efficient semiconductor utilization compared with typical non-resonant multilevel modular switched capacitor boost inverters. Test results indicate that the proposed topology can be used for single-phase non-isolated PV boost inverter applications with small ground leakage current, high voltage conversion ratio, small volume and potential high efficiency.

This paper presents a smart Composite Modular Power Delivery Architecture (COMPDA). One example for this architecture is applied in DC-DC 48V to 12V data center circuit shown here. COMPDA has many good features, such as voltage clamping for all devices, low total device stress under wide range ratio. It also has good modular characteristics. Each device has the same voltage and current stress. The principle operation is that all modular cells are connected in-series to share input high voltage and in-parallel to pump output high current, which can easily achieve ultra-high power density and high efficiency. Zero current switching (ZCS) can be applied in the proposed converter. A 4X unregulated DC-DC COMPDA with ZCS is studied for data center application. Finally, a GaN based prototype is built and tested to verify the theoretical analysis. The proposed converter can achieve the peak efficiency at 98.55% under ultra-high switching frequency 253 kHz. Furthermore, another GaN based high-density 6X boost COMPDA based prototype is also presented here. The peak efficiency can achieve 98.9% at 4kW power rating.

Switched-tank converter (STC) concept can leverage the advantages of wide bandgap devices. And it is especially suitable for GaN/SiC devices. This paper presents an improved STC with separate resonant branches and on-time control method to achieve efficiency improvement. Since the STC relies on resonant operation to reach high efficiency, the problems caused by unmatched resonant tanks in different resonant loops need to be studied and solved. The study shows that in most severe case, problems caused by the unmatched resonant tanks not only compromise the efficiency of STC, but also cause undesired damage to the STC. As a result, old simple control method that uses only two complementary control signals (50% duty cycle) is not good enough anymore. Hence, an improved STC as well as adaptive on-time control method have been proposed to ensure soft-switching and high efficiency operation of the STC. A 1.2kW prototype with 6:1 conversion ratio has been built to verify the effectiveness of the proposed control method. The experimental results show that with the proposed control, efficiency of the improved 6x STC is increased by 0.5% at 900W, which means power loss is reduced by 16.77%. A peak efficiency of 98.71% has been achieved.

This paper presents an ultra-efficient composite modular power delivery architecture (COMPDA) that is able to be used for DC-DC, DC-AC, AC-DC, and AC-AC applications. The proposed architecture has very good module structure, the devices in each module have the same voltage and current stress. When it is used for high conversion ratio dc-dc applications, the device voltage stress is the same with the low voltage side, and the device current stress is the same with the low current side. Compared with the buck, flying capacitor multilevel or modular multilevel dc-dc converters, the proposed architecture has much lower total device power or current stress, which means a significant semiconductor chip area reduction. During the operation, the modules are connected in-series to sustain the high voltage and connected in-parallel to pump the high current. By properly design the resonant tank in the COMPDA, all the switching devices could achieve zero current switching or zero voltage switching. A partial power regulator could be integrated in the proposed architecture, where only a minimum amount of power needs to be processed in order to achieve the output voltage fine regulation. By properly control the operation of modules, the COMPDA can also be used as a high boost ratio DC-AC inverter or high step-down ratio AC-DC rectifier while keeping the soft-switching features, which are specially suitable for solar farm and data-center application. When apply the COMPDA to single-phase DC-AC or AC-DC applications, only a minimum amount dc-link capacitor is needed since the needs of energy storage is minimized, and there is no high voltage dc-link requirement anymore. In order to show the benefits of the COMPDA, one special unregulated case derived from the proposed architecture has been used for 48 V–8 V DC-DC data center application, which is also named switched-tank converter. Finally, a GaN based prototype is built and tested to verify the theoretical analysis. The proposed converter can achieve the peak efficiency at 98.55% under switching frequency 253 kHz and power density is ∼750W/inch3.

This paper presents a new indicator of SiC MOSFET gate oxide degradation based on Miller plateau. The physical mechanism of Miller plateau shift with gate oxide electric field is first analyzed. The relationship between Miller plateau and ambient temperature is then explored by theoretical analysis. The electro-thermal simulation is conducted in LTSpice to verify the Miller plateau shift with ambient temperature. Besides, 20 groups of High Electric Field (HEF) acceleration tests are conducted with V gs stress amplitude of 25 V, 30V, 35V, 40V and stress duration of 10, 40, 70, 85, 100 hours. 5 SCT2120AF SiC MOSEFTs from Rohm are stressed in each group. After ageing tests, the stressed devices are used to verify dynamic characteristic change in the designed double pulse test platform. After 100-hour HEF tests with 40V V gs stress, Miller plateau shift can reach up to 1.5V. Finally, comparison is made among Miller plateau, threshold voltage and gate resistor turn-on energy. Analysis shows that Miller plateau can be used as an indicator of SiC MOSFET gate oxide degradation with detectable amplitude shift as well as inherent gate driver integration and online monitoring characteristics.

With the popularity of mobile smart devices, location-based services have been more widely used. Bichromatic Reverse k Nearest Neighbor (BRkNN) queries have become a hotspot in spatial-textual databases domain. In this paper, we extend the concept of traditional BRkNN method to process the object with multi-level tags in some specific scenes, and we propose a new type of query, called Maximized Bichromatic Reverse k Nearest Neighbor with Multi-Level Tags queries (MaxBRkNN-MLT), to find the optimal position of object with multi-level tags in the spatial-textual database. Unlike traditional methods, the number of the results of the MaxBRkNN-MLT query is maximized, which can cross the great divide between space and text. The query method proposed in this paper has a wide range of application scenes. For example, in the advertising industry, advertisers expect to find an optimal position, so that the ads with a given tag can attract the most users. Finally, experiments show that the MLT method has better query precision and execution efficiency than the baseline approach.

In this paper, a resonant switched-capacitor dc-dc converter is proposed for data center application. The proposed converter possesses features such as high efficiency, high power density and light-weight. Zero current switching (ZCS) can be achieved with the resonant operation, which allows the converter operating under high efficiency. Proper switching device selection and in-depth power loss analysis on different devices have been performed to support better design of the proposed converter. Simulation results are provided in this paper to validate operation principle of the proposed converter. A GaN based prototype with nominal 450W, 54V input and 9V output has been built to verify the theoretical analysis. The maximum power rating of the designed prototype is 600W. Power density of the prototype reaches 750W/in3. When the prototype operating at 253kHz, it achieves 98.55% peak efficiency.

This paper presents a multilevel modular step-down DC-DC switched-capacitor converter that utilizes soft switching resonant operation. Previous versions of traditional multilevel modular switched-capacitor dc-dc converters (MMSCC) have a low switching device voltage rating. Moreover, capacitor voltage stress increases proportionally as the voltage conversion ratio increments. Significant advantage of proposed converter in this paper is high voltage conversion ratio with capacitor voltage stress always being only two times of the low side voltage. Additionally, the proposed circuit (SWRMC) shares all of the advantages from the (MMSCC) type converters such as soft switching, low voltage device rating, bidirectional operation, high conversion ratio, extremely high or low environment temperature operation, basic control algorithm and modular structure. In comparison with the magnetic core based converter the converter proposed in this paper is found to be superior due to the simple structure, soft switching, low conduction loss and 97% or higher efficiency.

This paper presents comparison of GaN based switched-tank converter (STC) and traditional cascaded voltage divider (CVD) dc-dc converter for data center applications. The operation principle of two kinds of converters are presented. And simulations with Saber are both conducted to verify the proposed comparison, respectively. The system volume and efficiency are fully compared via GaN based device under the same current stress condition. It is shown that under the same condition, STC can achieve 1% higher efficiency at full load and over 2% higher efficiency at light load when compared with CVD. Furthermore, STC can achieve ∼37% power-density improvement when compared with CVD. Finally, GaN based STC prototype is built and experimental results verify the proposed idea very well.

This paper presents a novel modular multilevel switched-capacitor based resonant converter (MMSCRC) to achieve dc-dc power conversion function. In comparison to existing multilevel modular switched-capacitor circuits, the proposed switched-capacitor resonant converter (SCRC) possess voltage regulation capability. Its modular structure accomplishes voltage conversion function and proposed closed-loop phase-shift control regulates the output voltage. Meanwhile, zero-voltage switching (ZVS) operation can be accomplished. By increasing switching frequency of wide bandgap device and reducing inductance in the circuit, magnetic components can be minimized. Hence, high power density and high efficiency of the proposed converter can be achieved. This paper includes circuit operation, steady-state characteristics as well as in-depth analysis for ZVS operation. Simulation results are provided to verify the operation principle of the proposed MMSCRC. A 600W lab prototype with 6:1 conversion ratio has been built. Experimental result are also provided to validate the theoretical analysis.

This paper shows a combinational decoupling method to minimize dc-bus capacitance and improve power density for DC to AC applications, especially for single-phase inverters. The proposed method is combined with in-series structure and in-parallel structure together with power equations analyzed. The detailed control strategy is given. The simulations are conducted via PLECS software to prove the theoretical proposal. Furthermore, a 2kW prototype is built. The power density of this prototype is ∼55W/inch3. The decoupling capacitance is minimized ∼2/3 that of traditional series voltage compensator method. The experiments are further finished and the efficiency of the prototype is ∼ 95%. Besides, the issues for 2nd order decoupling method is investigated and an instantaneous pulse decoupling method is discussed in this paper. A brief comparison of these two methods are summarized.

This paper presents a modular multilevel dc-dc converter (MMC) with zero-current switching (ZCS) feature and its analysis for medium-voltage direct current (MVDC) applications. The main focus of this paper is analyzing the overall efficiency of the proposed MMC and experimentally verifying the theoretical analysis on operating principle. Furthermore, in order to achieve high efficiency, the proposed MMC has the capability to maintain ZCS operation at different load conditions. In this paper, the proposed MMC with maximum 200kW power output is used to perform the efficiency analysis. Simulation uses the same parameters that are used to analyze efficiency and displays waveforms when the MMC operates at 66kW. At last, a 200W lab prototype has been built. And operating characteristics of the MMC are validated through experimental results.

This paper presents an innovative resonant multilevel modular bidirectional dc-dc switched-capacitor converter based on GaN switching devices. Proposed converter achieves the following advantages; soft switching, low voltage device rating, boost or buck mode, common ground, high conversion ratio, variety of environment temperature operation, simple control algorithm and modular structure. The proposed single-wing ladder resonant multilevel converter is composed of COG temperature compensating ceramic capacitors along with GaN switching devices that can handle high switching frequency hence, high power density. Converter proposed in this paper has superior advantages over the traditional multilevel modular switched capacitor dc-dc converters due to its modular structure, high efficiency, common ground, low voltage device stress, and simplicity of the control algorithm.

This paper presents high-power-density inverter with pulse current injection power decoupling method. In order to reduce the dc-link capacitor significantly, dc-link current is fully analyzed with double Fourier method in theory. Compared with the traditional 2nd harmonic power decoupling method, the pulse current injection method considers not only 2nd order harmonic but also the higher order harmonics. As a result, it can further reduce the dc-link capacitor value. The relationship between the dc-link capacitor value and different order harmonics has been investigated and analyzed. It shows that when dc-link capacitor value is reduced to one critical value, traditional 2nd harmonic power decoupling method is limited. Finally, one novel pulse current injection method is proposed and verified by simulation.

Co-training is a semi-supervised learning paradigm that trains some classifiers and let them label some unlabelled instances for each other during the learning process. One challenge of the co-training style algorithm is to train an initial weakly useful predictor when the number of labeled instances is very limited. In this paper, we use Teaching-to-learn and Learning-to-teach strategy, which each the ground truth is regarded as a "teacher". So that the entire labeled process is guided by the teachers and selected from simple instances to more difficult ones. In the teaching-to-learn step, teachers select the highest evidence instances based on distance and local density, which can automatically, spotted and excluded outliers from analysis. In the learning-to-teach step, teachers reversely learn from the co-training feedback to properly select the highest evidence instances for the next iteration. Thus, we propose a framework based co-training style algorithms to efficiently improve the performance of them. Moreover, we can tackle with real-world applications in which may have very few labeled training instances, eventually only have one labeled instance for each label. Finally, we give specific algorithm for efficient solving in different tasks and outperforming in UCI datasets, also achieve better performance in content-based image retrieval.

The field of biomarkers in cancer research has recently gained widespread interest, for its potential to improve diagnosis accuracy, prognosis, and make cancer treatments to be more personalized. However, the detection of multi-tumor markers method still has many problems, such as limited applicability, need to different model for different tumor markers, a simple series or parallel method cannot effectively take advantage of different tumor markers. This paper proposed a data fusion method for multi-tumor markers, which can be adapted to different scene. It can effectively use the different markers to give an adjuvant diagnosis. With the new markers continue to be found, we can provide guidance for the combined detection.

This paper presents a high-efficiency and low cost linear LED driver with concave current control method for low power applications. In order to achieve a high efficiency of the linear LED driver, control methods with three different current shapes, namely, traditional flat and convex current shape, proposed concave current shape, have been discussed and compared via theoretical analysis. It is concluded that the proposed concave current control method has the highest efficiency among these three methods for different input voltage conditions. The design considerations have been provided. Finally, experimental results verify the proposed method.

In this paper, in order to improve the in-series power decoupling circuit, an in-series and -parallel method is proposed to further reduce the extra capacitor without performance penalty. Based on DC-side in-series method and AC-side three-leg method, a novel way is given to not only reduce the DC-link capacitor but also to decrease the extra capacitor. Thus, it could further make the system high power density and long lifetime. Simulation results are presented to show the performance of this proposed method.

This Paper presents a new modular multilevel converter (MMC) concept to achieve high ratio DC-DC power conversion purpose. This converter can be deployed in High-voltage direct current (HVDC) system, or it can be used as a Solid-state Transformer (SST). The proposed converter is flexible, reliable and it possesses scalability and simplicity. The conversion ratio can be regulated by the phase-shifted control, without any hardware modification. Using the MMC concept, this converter can be implemented with lower voltage rating switching devices instead of much higher voltage rating switches, which is usually unavailable. Two kinds of submodules are outlined for building the MMC. By utilizing inductance distributed on wire/trace and equivalent series inductance (ESL) of the capacitors, no inductors are needed to build the converter. Resonant operation between wire/trace inductance, ESL of capacitors and submodule capacitors assists the converter work under ZCS mode for higher efficiency. The operation principle as well as simulation results of a 20kW converter are presented in this paper.

This paper presents a low cost multilevel modular switched capacitor dc-dc converter for high power and high voltage conversion ratio application with reduced device count. The switching devices number of the traditional multilevel modular switched capacitor dc-dc converter is increasing fast with the increase of conversion ratio. Large number of switching devices means more gate drive circuit, more complex and higher cost. The proposed circuit has low device number, low device voltage stress and current stress. The proposed circuit also inherits the modular structure, simple control, low conduction loss, high switching frequency operation capability, and soft switching features from the traditional multilevel modular switched capacitor dc-dc converter. Therefore, a high voltage conversion ratio switched capacitor dc-dc converter with low cost and simple structure can be built using the proposed circuit. Simulation results are provided to validate the operating principle and features of the circuit.

This paper proposes a novel method of multi-beam laser heterodyne measurement for the angle. Based on Doppler effect and heterodyne technology, loaded the information of the angle to the frequency difference of the multi-beam laser heterodyne signal by sinusoidal frequency modulation of the oscillating mirror, which move as simple harmonic motion, this method can obtain many values of the angle after the multi-beam laser heterodyne signal demodulation simultaneously. Processing these values by weighted-average, it can obtain the angle accurately. This novel method is used to simulate measurement for incident angle of standard plane mirror by MATLAB, the obtained result shows that the relative measurement error of this method is just 0.49%.

This paper puts forwards a novel method of second harmonic of multi-beam laser heterodyne measurement for the thickness of the glass, which is based on the combination of Doppler effect and heterodyne technology, loaded the information of the thickness to the frequency difference of the multi-beam laser heterodyne signal by frequency modulation of the oscillating mirror, which is in the light path. Heterodyne signal frequency can be obtained by fast Fourier transform, and can obtain values of the thickness accurately after the multi-beam laser heterodyne signal demodulation. This method is used for simulate values of different thickness of the glass through numerical simulation of MATLAB and the obtained result shows that maximal relative error is less than 0.02%.

There are many some advantages of wet flue gas denitrification technology, such as simple process equipment, low operating temperature, less energy consumption and low cost. This article introduces the research progress of NO x removal by wet process technologies domestic and abroad, including alkaline liquor absorption process, acid absorption process, oxidation absorption process, reduction absorption process and complexation absorption process, and analyzes the advantages and disadvantages of different methods. The article puts forward the proposal to the development direction of wet flue gas purification technology in China.

The CO 2 removal technology are introduced from Post-combustion, Pre-combustion separation and Oxy fuel-combustion. And storage methods(geological storage, oceanic storage) are also introduced. The advantages and disadvantages of CO 2 removal technology and storage methods are discussed. It shows there is a sea of troubles for us to develop the technology.

Whether does dioxin emissions of Municipal Solid Waste(MSW) incineration can effectively control for environment protection and the human body health, but also related to the sustainable development of the domestic waste treatment and the key of utilization for energy from solid waste. This paper thoroughly discusses the main froming mechanism of dioxin formation in thermal processes and the reduction measures of its emissions. The parameters influencing the quantity of dioxins formed and the dependence of their quantity on the conditions of combustion are highlighted. Furthermore, the methods of reducing dioxin emissions from combustion processes are discussed. Finally proposing several measures in control and purification dioxins technology based on its formation mechanism.

A new method which can estimate the null time in the ellipsometry is developed in order to measure precisely the thickness and refractive index of thin film. Ten images are taken consecutively at the null time during the measurement, and ten sets of (P i , A i ) value for the images are obtained. The gray scale matrices of different images are analyzed and transformed, then the real values of the (P i , A i ) for the sample are determined. Distribution of the thickness and the refractive index of the sample are presented. The measurement accuracy for thickness and refractive index are 1nm and 0.01, respectively.

The optical local oscillator power which makes the maximized signal-to-noise ratio is analyzed. Aiming at photovoltaic detector, the optimum optical local oscillator power is researched, and the theoretic value of optimum optical local oscillator power is acquired, and takes some experimental research basing on the theory.

A method for measuring thickness of glass is proposed based on laser heterodyne measurement and Doppler measurement, the largest error of simulation results is 0.8%.