The effects of earth rotation rate change on earthquakes were analized based on the earthquakes that are bigger than Richter scales 7.0 duaring the period of January 1st, 1973 and December 30th, 2012 around the world with LOD data from IERS. At the same time, the definitions of discrete function derivative and R were introduced to assess the seismic activity. The results indicate: 1) The probability of earthquake occuring at extreme points and zerovalue point is much larger than the probability of stochastic theory, which shows that rapid change or the rechange period of negative and positive values of earth rotation rate  would induce earthquakes. 2) The seismicity presents yearly period and nearly 6 500day period, which corresponds with LOD’s annual period and 18.6year period. These results have significant value in understanding and prediction of earthquakes.

On the basis of the absolute and relative gravity observation data (1998—2010）got from China Crustal Movement Observation Network gravity field and its dynamic variation of Xinjiang were obtained，and evolution characteristics of regional gravity field and its relation with  seismic activity before Ms6.6 earthquake in the junction of Xinyuan and Hejing in Xinjiang on June 30，2012 were analized.The results indicate that:1)there is a close spatial relationship between the gravity change and tectonic activity of Tianshan fault，and both of material migration and tectonic deformation accompanied by action of the fault，lead to surface gravity change.2)Dynamic images of regional gravity provide mobile gravity precursor in the progress of preparation and occurrence of Ms6.6 earthquake.especially location of the earthquake.3)Accumulated gravity sequential variation at Xinyuan measuring site reached 80×10^-8ms^-2，which indicates rapid increase of gravity value，and the earthquake occurred in the period duaring gravity increases rapidly.

The international Venus exploration programs and the Venus gravity field models established by the tracking data of international Venus probes are introduced, and then we propose that the SSTHL/SGGDopplerVLBI observation mode should be adopted which is preferable for the future Venus gravity gradiometry mission in China, and the satellite orbital altitude of 50-100 km is an optimal selection.

The geometric transformation model between the INSARLOS deformation and GPS displacement is established.The model is employed to convert GPS displacement into the deformation in LOS direction. Based on Delaunay triangulations and cubic interpolation,the errors existed in InSARLOS deformation are rectified.According to variable fault slip model,the fault parameters are acquired by accurate GPS observations.According to the characteristic that the displacement direction derived from coseism is accordant in certain range, the displacement direction derived from forward modeling was applied to realize the transformation from InSARLOS to threedimensional deformation.The result shows that the displacement field is consistent with actual damage of surface ruptures.

Current strain field in the eastern margin of the Tibetan Plateau was calculated with the triangle methodthe triangle method using observation data of the Continental Tectonic Environmental Monitoring Network of China during the period of 2009—2011, and calculated results were tested with Triangular shape factor and strain significance.The observation region is divided into interior, edge and boundary according to the obtained strain characteristics. Longmenshan fault zone of the eastern margin of the Tibetan Plateau has become a district wiyh high values of strain after Ms8.0 Wenchuan earthquake, which reflects the block relaxation and deformation adjustment after the earthquake. Xianshuihe fault is still a district with largest strain, and it’s very important to monitor it continuously.

The repeated gravity data and absolute gravity data were used to study vertical deformation of Taiyuan Basin. preliminary results show that the principal form of vertical deformation in Taiyuan Basin is subsidence, and its rate drops gradually. The subsidence rate in Taiyuan Basin is 21.8 mm/a from 2001 to 2006, caused by combination of inherited tectonic motion and groundwater exploration. The subsidence rate in Jinci area is 66mm/a since 2009, caused by inherited tectonic motion.

StackingInSAR technology was used to measure the average crustal deformation rate  of Linfen Basin and its boundary faults during 2003-2010.The results show that the inner part of Linfen Basin was subsiding and the mountains on both sides were rising relatively. The differential movement rate in middle section（YuliYucun）of Luoyunshan Fault was 4 mm/a, the rate in southern section（FanjiazhuangXiweikou）was 6 mm/a, and the rate around the Fault was about 2 mm/a, which corresponds with the result of the crossfault leveling at the same time. Within Linfen Basin, different tectonic units had different deformation patterns, the rate increased gradually from north to south. The rate in the northern part and in the southern part of HongtongLinfen Depression was 0-4 mm/a and 4-8 mm/a respectively. The Xiangfen Bulge was stable. The rate in the northeast part of Houma Depression was 4-8 mm/a, and 8-12 mm/a in the west part.

The vertical deformation rates of Shanxi Area during 1980—2010 got from leveling observation data and matched with geological structure units, the block movement rate of the Area（as a block）and the activity rates of boundary faults were obtained. Based on the data, the crust vertical changes and tectonic activity were studied in the past 30 years in the Area. The results show that the western region of Shanxi Area ascended, the eastern region descended, the Lvliangshan Block was tiltuplifting, the Taihangshan Block tilted, and Shanxi Graben tiltascended. The modern tectonic activity of Shanxi Graben system was faultedsubsidence with a rate of 2.68 mm／a，and activity rate of of block boundary faults were all normal faulting with the rates of1.14-4.33mm／a.

A fine 3dimensional element model of Tangshan fracture zone based on 2 seismic refraction and reflection profilings across Tangshan fracture zone was presented. GPS data and tectonic stress observing results were used to constrain the simulation. in the study areawere compared to insure the reliability of the simulation was verified by the comparison of focal mechanism and the element results. Analysis results of equivalent stress and shear stress on the XY plane indicate the seismogenic possibility of Tangshan fracture zone.

The major advantage of PSInSAR is inversion of ground surface slow change for a long time based on high coherent targets. Using 27 single look complex(SLC）Envisat ASAR images over Beijing area, the ground surface deformation field during 2006 2010 and its details were obtained with PSInSAR method . The results show that:（1）There were 7 subsidence centers: Shuangqiao, Langfang, Yanjiao, Zhangxizhuang, Yanggezhuang, Pinggu and Shahe. The subsidence rates of Shuangqiao and Langfang are largest, with a maximum rate of -27.19±5.66 mm/a and -24.65±5.14 mm /a respectively.（2）The largest subsidence region is Shuangqiao, its size is about 230 km2.（3）The subsidence of Beijing city rebounded continuously with a rate of 10.92±3.24 mm/a. That means that the recharge since the 1980s prevents ground subsidence.（4）The uplift in northern mountains suggests that the impact of regional tectonic movement has been continuously growing.

High frequency tremor wave were found in onesamplepersecond data observed by broadband tiltmeter, broadband seismometer and PET gravimeter. The analysis results of the data recorded in broadband titmeter and the meteorologic data during 2008—July 2011 indicate that the high frequency tremor wave(0.13-0.33 Hz）is related to typhoon and the sea wind with a speed of grade 6 or more, especially the sea wind in Taiwan Strait.

Analyzing results of observational data of water temperature gradient experiment in Wenquan Well, Xinjiang indicate that there are macrostability and microdifferency in water temperature changes. The macrostability can be represented by a Logarithmic equation, and the microdifferency manifests in dynamic changes in various depth. In addition, the tidal variations can be captured at the layer where water gradients changing violently.The differency mainly is related to hydro geological structure.

A computing method of the tunnel construction of universal segment through analyzing the design of the segment and the corrective route was presented.The method is established on the relation among the axes of the shield，segment and tunnel.The results of checking calculation for a project indicates that the method meets the requirements of design.

An efficient solution was proposed with which holes and gaps within point cloud of building facade can be quickly repaired. The first step is to quickly find distribution regularity of building facade components with interface mode.The second step is to determine position and type of the delected component according to component distribution regularity.The third step is to copy the reference point cloud model to the delected area with ICP algorithm.The last step is to merge the reference point cloud model into original point cloud. Experiments show that the solution can be used to more effectively process data with interface model than other methods.

The spatiotemporal distribution of the shear strain with steep drop is modeled numerically for a rock sample with a Zshaped fault by use of FLAC3D. Nine parameters related to shear strain with steep drop and released energy are calculated at different deformation stages.The results show that the failure and slip at the inflection points provide a possibility of the middle fault sliding, which causes the overall instability of the sample. The parameters related to the shear strain with steep drop reach peak value in the time close to instability and during the process of instability

Several inversion methods for wave periods based on altimetric data were described in the paper. The wave periods derived from with the methods were checked using ERA40 models. The results show that the accuracy with CS method is the lowest, the accuracy with CSG2 is the highest, and the accuracy with the other three methods(CSG1, Hwang and ZT）are similar to each others. Improvement of the accuracy with the latter four methods is due to use of the buoy measured data.

A method of weight determination to improve accuracy of calculating the station coordinate in free station establishment of industrial measurement,based on variance component estimation.The experimental results indicate that the amount of redundant observation influences the stability of the VCE results,and the accuracy of weight influences the reliability of the calculating parameters. Compared with traditional methods, the method can significantly increase the reliability of the calculating parameters.

Utilizing the advantages of getting augmentation information of singlestation modeling compared with doubledifferenced modeling,  the method of regional groundbased augmentation information with refined pseudorange of singlestation modeling method and the corresponding ambiguity fixing method are proposed for users in one epoch. Practical examples with the data from  CORS network of Jiangsu Province in China indicates that the methods can satisfy users' needs of instantaneous realtime precise point positioning.

Precise Orbit Determination (POD）of GNSS navigation constellation is the most key technology for guaranteeing the performance of GNSS navigation system. Basic methods，process and strategies to determine the orbits of GNSS navigation constellation were introducted. A set of software has been developed with which POD of GPS navigation constellation based on the data of 200 IGS stations during January is calculated. The result of comparing the orbits with IGS final orbit products shows that the orbits position accuracy of R，T，and N position components is 0.70，1.60，1.04 cm，and the 1DRMS is 1.23 cm，and it’s the same level as IGS final orbit products.

A new GM(1,1）model to analyze and predict the settlement of tunnel was established, based on Kalman filter. The analysis results of some examples show that the model can be used to eliminate efficiently the random error in original data so that the settlement curve was more smooth and more close to real settlement values than that by GM(1,1）model.

Based on the comparation of two kinds of GPS broadcast ephemeris parameters and their fitting algorithms，the influence of some changed parameters on broadcast ephemeris fitting algorithms is analizesd. The calculated results with the method using the orbit interpolated by GPS precise ephemeris indicate that the method is feasible.

A single epoch algorithm with irrespective of cycle was presented to be used to continuous dodeformation monitoring with a high sampling rate receiver. The feasibility of this approach is proved by experiments.

In deformation monitoring with GPS, gross errors often exist in displacement observations, for influence of cycle slip, multipath effect, diffraction and the geometry of satellites.The geometry of satellites is not an error source of positioning, it affects  positioning precision basically by the unstable normal equation. RPDOP(Relative Position Dilution Of Precision) of GPS precise positioning is incorporated into robust Kalman filter so that a reasonable weigh can be obtained. The result from practical GPS slide deformation monitoring application indicate that RPDOP based robust Kalman filter is effective.

Based on quantitative analysis for influence of tropospheric corrected model on the positioning precision in the Antarctic GNSS positioning, a tropospheric corrected model with real meteorological data is proposed to improve positioning precision in the same elevation cutoff angle. Correspondingly, under the condition of keeping the same positioning precision, the model can reduce elevation cutoff angle from 15 degrees to 5 degrees, which means that more lower elevational observations can be used to improve positioning precision and reliability significantly

The reliability of rate of continuous stations and mobile stations with the GPS long time series and the influence of rate of continuous station was analyzed, considering the different time lengths and starting phases, as well as the rate of mobile station, considering the influence factors of observation cycle and time span. The results show that horizontal rate of continuous station is within ±1mm/a based on 34 years’ fitting data,and the variation range of vertical velocity would still greater than ±3mm/a.

An improved searching algorithm of least independent closed loop and its forming conditions based on spanning tree and cotree were presented. Through program designing and verification with MATLAB, the goal of comparing and extract information of least independent loops automatically was realized. Results calculated with GPS control network data indicate that  the loop information meets all requirements of least independent close loops.

A method was presented  in which variance after total least square adjustment and the variance ratio test are used to be used for position of gross errors of common points and reducing its influence in coordinate transformation with BurseWolf model. The method combined by hypothesis test with weightreducing iteration can fit for the cases in which gross errors can not be eliminated directly. In this way, the influence of gross errors is reduced and the precision of coordinate transformation is improved. Some examples prove that this method is practical and effective.

In order to improve the reliability of underwater terrain matching navigation, a new method is proposed which is based on contour correlation matching and hypothesis test. In case a false matching is detected, the unscented transformation is used to rematch for merging the terrain observation data, then applies in the new correlation matching. Experiments prove that the method can improve precision and its reliability of matching positioning, and verify the effectiveness of the method.

The advantages, disadvantages and applications scopes of denoising methods are compared using wavelet filter, median filter and mean filter in the cases of different single noise pollution and mixing noise pollution using the GPS deformation monitoring data. Combining with the advantages of three filtering methods,  an improved composite filtering method is proposed, which the wavelet filter is used firstly, and then the median filter and mean filter methods are applied. Theoretical analysis and practical applications verify that denoising effect of the improved method is better than tranditional methods.

Estimation of sevenparameter model by noniterative method is relatively stable，but accuracy is lower; accuracy of estimation by iterative method is higher，but its iteration convergence is vulnerable to the initial value.A method estimating sevenparameter model by the combination of noniterative with iterative method was proposed.The new method resolves the problem that the iterative methods is unsuitable for large rotation angle.Convergence in estimation by the new method is fast and accuracy is higher，and it can be applied to cartesian coordinate with arbitrary conversion because of not required initial values.The checking calculation for some cases proves the method is correct and effective.

Application and feature of various single channel ICA algorithms (DEICA，WTICA，EMDICA and EEMDICA）are compared and analized. Results of the simulation experiments with them show practicality of the algorithms in the single channel signal separation，and feasibility in deformation analysis.

The normal triplefrequency cycleslip detection and repair algorithms, pseudorange/carrier phase combination and MW & Geometryfree combination, are studied. The detection threshold is given. Using Compass triplefrequency real data, with different sampling intervals, the detection effects of the two methods are analyzed in cases of cycle slips occurring on a single frequency and multiple frequencies. The results have shown that in the process of observation, the sampling interval should be set smaller than 10s which is more beneficial. And the MW Geometryfree combination is usually more effective.

A joint inversion method using cepstrum was presented for solving thin layers recognition problem in the frequencydomain inversion algorithm of layered media. In the case of lack of any priori information, it is possible to use trueficnitious cepstra method to recognizing thin layers, for the quefrency peaks indicate reflection interfaces. The opperating steps are as the followings:(1)moving the phase of the ficnitious part in the spectrum of the coefficient series. (2）adding the results to the real part of the spectrum and calculating the cepstrum of sum. (3）comparing the quefrency spikes on the cepstrum of the sum with that of trueficnitious cepstrum. Thus, one can detemines if the spike represents a thin layer.

The inconsistency of time recorded in Rinex file with real observation time was discovered in cycle slip detection of GPS reference station data，and the inconsistency leads to the difference between cycle slip.Two methods to solve the inconsistency was proposed.The first one is “combined cycle clip detection method”，and  the second one is “Observation epoch repair method”.Simulation calculating proves each of the methods is feasible and effective.

An unbiased estimation formula is derived from mathematic expectancy formula of residual matrix unit weight variance by penalized least square based on semiparametric model. Numerical test of 500 sample data indicates that the method is effective when there are system errors in the observations.

A method of preeliminating respectively varous types of estimated parameters was presented. In the process of solving precise orbit determination of LEOs, it is needed a large amount of time to deal a large number of clock parameters, ambiguity parameters and pseudostochastic orbit parameters. Experiments indicate that the method,can improve calculating efficiency.

In the widearea real time precise point positioning, the ultra short term prediction of satellite clock error is needed for real time solving. The effects of the different satellite type clock error prediction were analized, including linear model, quadratic polynomial model, grey model and quadratic polynomial model based on epoch difference. The results of analysis show that prediction accuracy by the four methods satisfy the needs of precise point positioning in the ultra short term prediction of satellite clock error. In ultra short term prediction with small specimen, prediction accuracy of clock error is irrelevant to satellite. It mainly depends on the small scale change of clock error in different period and the length of prediction time.

A method utilizing elevation difference between the simultaneous observation values by two mobile stations very close to each other was proposed to improve the precision of elevation difference observation. Multipath effect can be solidificated as systematic error through taking fixed observation time according to diurnal characteristic of multipath effect.Both of formula derivation and tests verify that VRS technology can realize the precision of elevation difference measurement reaches millimetre level precision by the method.

A method for determining precise orbit and clock offset of navigation satellites using zerodifference observation was presented in the paper. Experimental results based on real GPS observation show that when oneday observation of 30 stations all over the world is used, the precision of radial,tangential and normal direction is 0.032 m, 0.084 m, and 0.078 m respectively,and clock offset is 0.18 ns. By comparison of the doubledifference method, the method is easier to operate, and the condition needed is more simple(just one step). So the mthod has better application prospect in the navigation satellite precise orbit determination.

Helmert variance component estimation（VCE）can adjust the ratio of  GPS and BDS observations adaptively. However, in case of the fact that observations is contaminated by the gross error, this model will distort the positioning result. Robust Helmert VCE, based on the IGG model Ⅲ was presented  to use in the pseudorange single point position and phase relative position. The  results of experiments show that the robust Helmert VCE can restrain the influence of gross errors and adjust the ratio of  mixed observations reasonably, and improve the positioning accuracy and reliability.

The measurement results of T-1 absolute gravimeter showed abnormal values after the Japan Ms9.0 earthquake occurred in March 11th 2011. The deviation caused by the powerful aftershocks of the earthquake reached at as large as 656×10^-8 ms^-2. The disturbed vibration isolation is the main reason of the abnormal values，according to the analysis for dropping chamber，interferometer and vibration isolation.

Due to some defects such as complex operation and large errors of the traditional artificial wire alignment transducers，and under consideration of the needs in observation for dam deformation，a new style of mechanical wire alignment transducer has been developed.The test results indicate that observation accuracy and resolution of the new transducer reaches 0.05 mm and 0.02 mm respectively.Moreover，the new transducer has advantage of large measuring range，higher precision and easy to use.