This action might not be possible to undo. Are you sure you want to continue?

INTRODUCTION

1.1 BACKGROUND AND OBJECTIVE

The principal software functions executed in the Inertial Navigation System (INS) computer are the angular rate signal into attitude integration function (denoted as attitude algorithms), use of the attitude data to transform measured acceleration signal into a suitable navigation coordinate frame where it is integrated into velocity and integration of the navigation frame velocity into position (denoted as velocity and position algorithms). Thus three integration functions (INS algorithm) are involved: attitude, velocity, and position, each of which requires high accuracy to assure negligible error compared to inertial sensors accuracy requirements. Since acceleration from the accelerometers and angular rate from the rate gyros are generally susceptible to various measurement noise sources, the accuracy of position, velocity and attitude information degrades with time. Recent research has shown that the growth of numerical errors in Inertial Measurement Units (IMUs) navigation with time can be prevented by using extemal aiding sensors such as the Global Positioning System (GPS) The benefits of this integration are well known. In this integration, the GPS receiver can provide various combinations of velocity-type and position-type measurements. As a result, the long-term INS errors are corrected by fusing IMU measurements with the GPS measurements, and the shortterm faults of the GPS receiver can be effectively isolated by the INS. Since in INS algorithm, the accuracy of attitude information governs the accuracy of velocity and position estimation, the most researched topic for the INS integration functions has been primarily focused on the design of algorithms for the attitude integration function and initial alignment. Several

parameterizations have been used to represent the attitude, such as Euler angles, Direct Cosine Matrix (DCM), Quatemions, etc. The quatemions method is advantageous since it requires less computation, gives better accuracy, avoids singularity and the kinematics equation is bilinear. However, the quatemion 1

must obey a normalization constraint, which can be violated in many nonlinear quatemion filtering approaches. In fact, if the state vector is quatemion, the predicted quaternion mean should be calculated in the rotational space to preserve the nonlinear nature of unit quatemion. The Extended Kalman Filter (EKF) is widely used nonlinear filtering method for attitude estimation. The Unscented Kalman Filter (UKF) is an extension of the classical Kalman filter to nonlinear process and measurement models and has been proposed for attitude estimation, which is more robust than the EKF for large initial attitude-error conditions.

1.2 LITERATURE SURVEY Aerospace Avionics system – A modern Synthesis (George M.Siouris, 1993). In this book explains various coordinate system and transformation that are necessary in the design and analysis of inertial navigation systems and basic kinematics equations, which form the heart of all inertial navigation system. Attitude Reference Properties of Inertial Navigation Systems (J. L. Henry, JR.). This paper describes the most attitude reference systems function quite well under laboratory conditions but suffer considerable degradation in operational environments. A discussion is given of the excellent attitude reference information inherent in inertial navigation systems. Global Positioning Systems, Inertial Navigation, and Integration (M. S. Grewal, L. R. Weill, and A. P. Andrews, John Wiley Sons, Inc., 2001). This paper describes about inertial navigation system (INS) and GPS integration using kalman filtering approach. INS algorithm using Quatemion model for low cost IMU (X. Kong, Robotics and Autonomous Systems, vol. 46, no. 4, pp. 221-246, 2004.). This paper presents a generic inertial navigation system (INS) error propagation model that does not rely on small misalignment angles assumption. The modelling uses quaternions in the computer frame approach. Based on this model, an INS algorithm is developed for low cost inertial measurement unit (IMU) to solve the initial attitudes uncertainty using in-motion alignment.

2

Applying the Unscented Kalman Filter for nonlinear state estimation (Rambabu Kandepu, Bjarne Foss and Lars Imsland, 2007). This paper shows that the Unscented Kalman Filter (UKF) is superior alternative to the extended Kalman filter (EKF) when solving the nonlinear system. Kalman Filter Face-Off: Extended vs. Unscented Kalman Filters for Integrated GPS and MEMS Inertial (Naser El-Sheimy, Eun-Hwan Shin, Xiaoji Niu, Mobile MultiSensor Systems Research Group, Department of Geomatics Engineering, The University of Calgary, Canada). This paper describes about the implementation of EKF and UKF and compared their performance for integrated low cost inertial navigation systems.

1.3 ORGANISATION OF THESIS

In Chapter 2, a brief introduction about navigation systems is dealt. This gives inertial navigation systems normally used, and the values are taken with respect to the reference frame. In Chapter 3, a brief introduction about global positioning systems. This gives detail about GPS, signals, source of errors and its advantages. In Chapter 4, quaternion algebraic operation, representation of parameter, notation and ins algorithms with quaternion approaches are discussed. In Chapter 5, a brief introduction about unscented kalman filters and discuss about ukf algorithms implementation. In Chapter 6, a briefly explain about INS-GPS integrations and what are advantages over the integrated navigations are discussed. In Chapter 7, the simulated experimental results are discussed and compute the attitude, velocity and position errors. In Chapter 8, the conclusion and future scope of the project are discussed.

3

while strapdown system is just what the name implies: the platform holding the IMUs is attached firmly to the vehicle. while the latter one measures the rotation rate with respect to the inertial frame.1 INTRODUCTION Inertial navigation is a system of dead-reckoning navigation in which the instruments in the craft determine its acceleration and by successive integration. Throughout this report we will refer to the navigation system‟s frame of reference as the body frame and to the frame of reference in which we are navigating as the global frame. 2. different types of mechanism are used. as shown in Figure 2.CHAPTER 2 INERTIAL NAVIGATION SYSTEM 2.. The former one senses the inertial accelerations. obtain its velocity and displacement.2 INERTIAL SYSTEM CONFIGURATIONS Nearly all IMUs fall into one of the two categories outlined below. The system is entirely self-contained and can be used both on earth. The main idea of inertial navigation is based on the acceleration integration. The difference between the two categories is the frame of reference in which the rate-gyroscopes and accelerometers operate. Stable platforms provide angular motion isolation from the vehicle. under sea and space. In order to match the sensitive axes of accelerometers with certain navigation coordinate frame. 4 . accelerometers and gyroscopes. i. an inertial navigation system (INS) comprises two sets of inertial measurement units (IMU).e. The second integration yields the vehicle position increments relative to the initial point.1. The first integration provides the velocity increment. In general. There are two different concepts for the navigation frame simulation: the stable platform and the strapdown system configuration.

2. 5 . Note that it is necessary to subtract acceleration due to gravity from the vertical channel before performing the integration.2. To calculate the position of the device the signals from the platform mounted accelerometers are double integrated. The platform mounted gyroscopes detect any platform rotations.3. The stable platform inertial navigation algorithm is shown in Figure 2. as shown in Figure 2. To track the orientation of the device the angles between adjacent gimbals can be read using angle pick-offs.1 The body and global frames of reference 2.1 STABLE PLATFORM SYSTEMS In stable platform type systems the inertial sensors are mounted on a platform which is isolated from any external rotational motion.Figure 2. This is achieved by mounting the platform using gimbals (frames) which allow the platform freedom in all three axes. These signals are fed back to torque motors which rotate the gimbals in order to cancel out such rotations. hence keeping the platform aligned with the global frame. In other words the platform is held in alignment with the global frame.

To track position the three accelerometer signals are resolved into global coordinates using the known orientation. as determined by the integration of the gyro signals.2.4. and therefore output quantities measured in the body frame rather than the global frame.2 STRAPDOWN SYSTEMS In strapdown systems the inertial sensors are mounted rigidly onto the device. The global acceleration signals are then integrated as in the stable platform algorithm. To keep track of orientation the signals from the rate gyroscopes are „integrated‟.2 Stable platform IMU Figure 2.Figure 2.3 Stable platform inertial navigation algorithms. This procedure is shown in Figure 2. 2. Stable 6 .

and body frame (b-frame). 2.3 COORDINATE SYSTEMS The coordinate systems used in this thesis are inertial frame (i-frame).5 Coordinate reference frames 7 . Strapdown systems have reduced mechanical complexity and tend to be physically smaller than stable platform systems. Figure 2. Figure 2. Earth-fixed frame (e-frame).platform and strapdown systems are both based on the same underlying principles. These benefits are achieved at the cost of increased computational complexity. As the cost of computation has decreased strapdown systems have become the dominant type of INS. navigational frame (n-frame).4 Strapdown inertial navigation algorithm.

and the axes coincide with vehicle-fixed ro11pitch. which is the fundamental plane). Conventions typically depend on the particular vehicle.heading coordinates when the vehicle is level and headed north. Axes of the n-frame is point in the directions north. The n-frame is used for local navigation purposes. The benefit of the east-north-up (END) system is that altitude increases in the upward. The b-frame is an orthogonal axis set which is fixed onto the vehicle body and rotates with it. use of the attitude data to transform measured acceleration signal into a suitable navigation coordinate frame where it is integrated into velocity and integration of the navigation frame velocity into position. It is formed by fitting a tangent plane to the geodetic reference ellipse at a point of interest.The i-frame is fixed with the centre of the Earth as the origin and axes which are nonrotating with respect to the fixed stars with its x-axis points toward the vernal equinox direction (also known as the First Point of Aries or the vernal equinox point). Thus three integration functions (INS algorithm) are involved: 8 . and the y-axis completes the right-handed orthogonal frame. however. the z-axis axis points in the direction of the North pole and the y-axis completes the right-handed system (note that the x-axis and y-axis are on the equator. The e-frame coincides with the i-frame at the origin and z-axis but rotates with the Earth rate. The advantages of NED system are that the direction of a right tum is in the positive direction with respect to a downward axis.4 INS ALGORITHMS Inertial Navigation System (INS) computes angular rate signal into attitude integration function. 2. down in that order. the x-axis points in the direction of the Earth's prime meridian. east.

4.Figure 2. can be updated as follows n k qb((k ) 1) n k 1) qb((k 1) b k qb((k ) 1) (2. the digital attitude quaternion integration model. 2.1 ATTITUDE ALGORITHMS The attitude quaternion dynamic equation is given by . qbn . which can be expressed in terms of rotation vector as follows: 9 . q 1 2 *b nb q 1 (q) 2 b nb (2.2) (2.1) Referring this equation.6 The Basic principle INS Block diagram.3) n qbn((kk)) qn((kk )1) qbn((kk) 1) b k where qb ((k ) 1) is attitude quaternion that accounts for the b-frame rotation relative to inertial space from its orientation at time tk−1 to its orientation at time t k.

n k The attitude quaternion qn((k ) 1) . Refer the above Equation.5 0. fb is the specific force measured by the accelerometer and ωz xy denotes a skewsymmetric matrix function with components of the angular velocity ωxy in z-frame.5 k cos 0. that accounts for the n-frame rotation relative to inertial space from its orientation at time tk to its orientation at time tk−1. 2.7) Where Cnb (q) is the attitude-quaternion matrix transpose of Cbn (q). Then the velocity navigation equation that derived by the Coriolis equation.5 k (2. in above equation.5 0.6) Where k is the n-frame rotation vector. hence they need to be transformed to the navigation frame in order to be used in the navigation equation fN fn fE fD q fb q 1 Cb n ( q ) f b (2. the digital velocity integration model.5 k 0.q b ( k 1) b(k ) sin 0.5 k (2.4.5) k k n ie n en k 1 2 tk (2. can be also expressed in terms of a rotation vector as follows: n k qn ((k ) 1) sin 0. 10 .4) k where φk is the b-frame rotation vector.8) where vn is the kinematic velocity expressed in n-frame.2 VELOCITY ALGORITHMS This means that the specific forces from the accelerometer triad are obtained in the body frame. becomes Vn Cb ( q ) f n g n (2 ( ie ))V n (2. gn is the gravity vector expressed in nframe.5 k 0.5 k cos 0.

10) and vb are the compensated outputs of the gyro and the accelerometer. The position evolution can be expressed as the following geodetic coordinates differential equations of latitude ϕ. longitude λ and altitude above the Earth (height) h. respectively. in the North-East-Down (NED) coordinates: vn Rm h Rn .9) The velocity increment due to the specific force vn fk I 0. fk V n g / cork in the above equation can be computed as The gravity and Coriolis correction term follows: Vgn/cork g n (2 ( n in ))vn k 1 2 tk (2. (2. can be written as V nk V nk 1 V nk V nk 1 V n fk V n g / cork (2.4.12) (2.13) vE h cos h n u znv n (2. 2 11 . The general position p n can be updated from above equation. as follows pkn pkn 1 pkn n pkn 1 uznvkn 1 tk 2 (2.5 k 1) Cbn((kk 1) (q ) vn can be computed as follows: fk vn fk 1 2 1 k k vb fk 1 2 where k k 1 vb fk 1 2 vb fk (2.14) Where n u zn is unit vector upward along the geodetic vertical (the z axis of the n-frame).vn.3 POSITION ALGORITHMS The INS computes position vector by integrating velocity vector equations.11) 2.15) Where vkn 1 is the velocity at the midway.

These are the residual errors exhibited by the installed gyros and accelerometers following calibration of the INS. If an unaided INS is used.2.5 ERRORS IN THE INS Most INS errors are attributed to the inertial sensors (instrument errors). Doppler heading sensor or air-data dead reckoning systems. a constant gyro output without angular rate presence. three velocity errors. Gravity model can also cause some errors.1 Sensor generated errors in the INS Errors in the accelerations and angular rates lead to steadily growing errors in position and velocity components of the aircraft. pitch and heading errors a constant offset in the accelerometer output that changes randomly after each turn-on.1 Alignment errors Accelerometer bias or offset Accelerometer scale factor error Nonorthogonality of gyros and accelerometers roll. These errors have to be modelled accordingly. due to integration. These are called navigation errors and there are nine of them { three position errors. 12 . results in an acceleration error proportional to sensed acceleration. The acceleration due to gravity varies from place to place along the earth and also with height. results in an angular rate error proportional to the sensed angular rate random noise in measurement Gyro drift or bias (due to temperature changes) Gyro scale factor error Random noise Table 2. the axes of accelerometer and gyro uncertainty and misalignment. It is for this reason that the INS is usually aided with either GPS. these errors grow with time. The dominant error sources are shown in Table 2. two attitude errors and one heading error.

7 DISADVANTAGES OF INS Bias drift problems (accuracy deteriorate with time) Requires initial and In-flight calibration and alignment Requires knowledge of gravity 13 . no external electromagnetic signals (non-jammable) A very high data output rate (much bigger than GPS) Works in all environments (indoor.Inertial sensors for strapdown systems experience much higher rotation as compared to their gimballed counterparts. underwater. Rotation introduces error mechanisms that require attitude ratedependent error compensation. underground) Excellent performance in high dynamic environment Exhibits relatively low noise 2.6 ADVANTAGES OF INS Autonomous system. 2.

Air Force develops. and maintains health and status of the satellite constellation. It tracks the GPS satellites. allowing even better performance than is possible using the basic GPS civilian service. and adjust the satellite clocks. the control segment. and integrity. These improve signal availability. and the user segment. The military service is available to U. GPS satellites provide service to civilian and military users.S. and operates the space and control segments. far into the future. The control segment consists of worldwide monitor and control stations that maintain the satellites in their proper orbits through occasional command maneuvers. The space segment consists of a nominal constellation of 24 operating satellites that transmit one-way signals that give the current GPS satellite position and time. uploads updated navigational data. The user segment consists of the GPS receiver equipment. 14 . and timing (PNT) services. The civilian service is freely available to all users on a continuous. worldwide basis. accuracy.CHAPTER 3 GLOBAL POSITIONING SYSTEMS 3. It has proven its dependability in the past and promises to be of benefit to users.1 OVERVIEW The Global Positioning System (GPS) is a U. maintains. and allied armed forces as well as approved Government agencies.-owned utility that provides users with positioning.S. The outstanding performance of GPS over many years has earned the confidence of millions of civil users worldwide. A variety of GPS augmentation systems and techniques are available to enhance system performance to meet specific user requirements. The U. navigation. which receives the signals from the GPS satellites and uses the transmitted information to calculate the user‟s threedimensional position and time. This system consists of three segments: the space segment. throughout the world.S.

000 miles above us. They are constantly moving.1 The GPS satellite system 15 . Figure 3. making two complete orbits in less than 24 hours. Small rocket boosters on each satellite keep them flying in the correct path. Each satellite is built to last about 10 years. GPS satellites are powered by solar energy. Transmitter power is only 50 watts or less. Here are some other interesting facts about the GPS satellites (also called NAVSTAR.2 THE GPS SATELLITE SYSTEM The 24 satellites that make up the GPS space segment are orbiting the earth about 12. A full constellation of 24 satellites was achieved in 1994.000 pounds and is about 17 feet across with the solar panels extended. Department of Defense name for GPS): The first GPS satellite was launched in 1978. Replacements are constantly being built and launched into orbit.000 miles an hour. when there's no solar power. A GPS satellite weighs approximately 2.3. These satellites are travelling at speeds of roughly 7.S. They have backup batteries onboard to keep them running in the event of a solar eclipse. the official U.

The GPS system uses a built-in model that calculates an average amount of delay to partially correct for this type of error. as it identifies which satellites it's receiving. Signal multipath . This increases the travel time of the signal. which is constantly transmitted by each satellite. This part of the signal is essential for determining a position.4 SOURCES OF ERRORS Factors that can degrade the GPS signal and thus affect accuracy include the following: Ionosphere and troposphere delays . meaning they will pass through clouds. thereby causing errors. The pseudorandom code is simply an I. ephemeris data and almanac data. Receiver clock errors . Ephemeris data.3 GPS SIGNALS GPS satellites transmit two low power radio signals. A GPS signal contains three different bits of information . it may have very slight timing errors.This occurs when the GPS signal is reflected off objects such as tall buildings or large rock surfaces before it reaches the receiver. contains important information about the status of the satellite (healthy or unhealthy). The signals travel by line of sight. code that identifies which satellite is transmitting information. glass and plastic but will not go through most solid objects such as buildings and mountains. current date and time. Each satellite transmits almanac data showing the orbital information for that satellite and for every other satellite in the system.a pseudorandom code.42 MHz in the UHF band. You can view this number on your GPS unit's satellite page. The almanac data tells the GPS receiver where each GPS satellite should be at any time throughout the day. Therefore.The satellite signal slows as it passes through the atmosphere.D.3. 3.A receiver's built-in clock is not as accurate as the atomic clocks onboard the GPS satellites. Civilian GPS uses the L1 frequency of 1575. 16 . designated L1 and L2.

Also known as ephemeris errors.Selective Availability (SA) is an intentional degradation of the signal once imposed by the U. electronic interference. This happens only rarely today. Satellite geometry/shading . As a result.This refers to the relative position of the satellites at any given time.The more satellites a GPS receiver can "see. Intentional degradation of the satellite signal .5 ADVANTAGES AND DISADVANTAGES The experienced GPS users learn to optimize data gathering by understanding system strengths and weaknesses. underwater or underground. Spatial and tabular data are collected simultaneously. which significantly improved the accuracy of civilian GPS receivers. Technology has reduced the effect of multipath and GPS data gathering capabilities are being strengthened. Over the past few years GPS maturity has eliminated many system disadvantages. Poor geometry results when the satellites are located in a line or in a tight grouping. Department of Defense. however. SA was intended to prevent military adversaries from using the highly accurate GPS signals. The government turned off SA in May 2000. Advantages Position accuracy is superior to conventional methods. Buildings.S. Number of satellites visible . 3. 17 . terrain. The user must be informed. these are inaccuracies of the satellite's reported location. before 1993 the full 24 GPS satellite constellation was not in place. For example." the better the accuracy. GPS units typically will not work indoors. periods existed where four satellites were not available for accurate positioning. or sometimes even dense foliage can block signal reception.Orbital errors . about GPS advantages and disadvantages. Coordinate systems and reference datums can be easily changed in the field and in the processing software. causing position errors or possibly no position reading at all. Ideal satellite geometry exists when the satellites are located at wide angles relative to each other.

Feature visual inspection is possible while gathering data. 18 . Multi-path reflective signals can make data inaccurate. Disadvantages Requires training and retraining as technology changes. Data gathering is possible 24-hours a day. Data collection costs are lower than conventional methods. GPS is unaffected by weather. Requires careful attention to system configuration and data collection standards and procedures. seven days a week.GIS conversion is simple. Heavy foliage and thick branched trees can attenuate and/or block satellite signals. Urban canyon buildings can block satellite signals.

a vector is treated equivalently as its corresponding quaternion with a vanishing scalar part.3) The norm of a quaternion is defined as ║q║ ═ q ο q* where q* ═ [s. v1 + v2] λq ═ [λs.CHAPTER 4 QUATERNIONS 4. where s is a scalar and v is a three dimensional vector. ─v] is the conjugate Where rο and rn are two quaternions with vanishing scalar parts formed by r ο and rn .4) [s1s2 ─ v1 • v2 . sin (θ/2) n] relating a fixed vector expressed in the original frame rο with the same vector expressed in the new frame rn by rn ═ q* ο rο ο q . If not specifically stated. there is a quaternion Q ═ [cos (θ/2). The kinematic equation of quaternion is 2q o ON q q N ON (4. quaternions satisfy the following operations. rο ]. (4.respectively.s1v2 + s2v1 +v1 × v2] (4.1) (4.1 INTRODUCTION Quaternion was invented by Hamilton as an extension of the complex number A definition of quaternion is given as q ═ [s. q1+q2 ═ [s1 + s2.v].5) 19 . For example. and the small number of parameters allows efficient implementation. the lack of trigonometric functions. By definition. λv] q1 ο q2 quaternion. The multiplication between quaternions is defined as ═ (4. For the frame rotation about a unit axis n with an angle θ. Usually quaternion implementation is preferred in updating the attitude as the linearity of the quaternion differential equations. rο ═ [0.2) Where λ is a scalar.

In such a device. γ are respectively This axis of rotation of makes angle μx. Each axis is placed in some rotational setting. μy.y. and z axes. Using the gimbles visualization. β. there will be a whole family of Euler angels that will describe the exact rotation. 4.1) By using Euler theorem (x. One way to visualize the operation of Euler angles is using a set of gimbles. z) co-ordinate system is rotated through the angle μ about fixed axis. z′) axes also. we can also see that for some set singular of rotations. v] 20 (4. we may expect that there is a representation for rotations that uses only three numbers.Where is the angular rate vector of the frame N relative to the frame O expressed in the frame N. y. Starting from this configuration. avoiding any singular or special configurations.3 UNIT QUATERNIONS Quaternions give us a way of representing rotations using four numbers. y′. suppose one sets up the gimbles so that the inner z axis and outer x axis are lined up. 4. Euler xyz angles are one such representation. Euler angles do not offer us a “unique” representation. Thus. there are three sets of axes as shown in Figure 4.7) . if one increases the rotation setting on the inner axis. For example. It can be shown that this can achieve any rotation. This will describe some rotation from the initial configuration.1. one can exactly undo this by decreasing the rotation setting on the outer axis. (4. but there are no funny places. which makes angle α. and can be written as q ═ [q0 q1 q2 q3] q ═ [s. Coordinating system (x. This situation is called gimble lock. μz with (x′. Euler xyz angles represent an arbitrary rotation by specifying a sequence rotations about the x. A quaternion is simply a fourtuple of real numbers. where there are three parameters but only two effective degrees of freedom. z′) is moving in some arbitrary manner however both coordinate systems have same origin (Fig 4. y.6) A quaternion can be written as a scalar s together with a 3 element coordinate vector v. z) fixed in space while the coordinate system (x′. y′. It is one more number than minimal.2 EULER ANGLES From the above discussion.

If we know Initial roll.)*cos (φ/2.8) 4. and μ = (μx 2+ μy 2+ μz 2)1/2. μx.) q2 = cos (ψ /2.)*sin (φ/2.)*cos (φ/2.)*cos (θ/2. μy. μy.) q3 = sin (ψ /2.) (4.)*sin (θ/2.1 Quaternion Angle Representations Parameter (μ.)*cos (φ/2.)*sin (φ/2.).)*cos (φ/2. pitch.)*sin (θ/2. yaw we can find initial quaternion q0 = cos (ψ /2.)*cos (θ/2.)*sin (φ/2.). μz are components of the rotation angle vector μ.)*cos (θ/2.A unit quaternions ˆq has unit norm where we define the norm as ^q q 0 2 q12 q 2 2 q32 (4.) q1 = cos (ψ /2.)*sin (θ/2.4 QUATERNION REPRESENTATION z′ z λ y′ y x x′ Figure 4. μz) are used in quaternion representation θ=q0+ q1i +q2 j +q3 k A quaternion is a four–element vector where μx.)*sin (θ/2.9) 21 .)*cos (θ/2.cos (ψ /2.sin (ψ /2.)+ sin (ψ /2.)+ sin (ψ /2.)*sin (φ/2.

1 INTRODUCTION The Unscented Kalman Filter was proposed in 1997 as alternative to Extended Kalman Filter. i 1. 5.. The n-dimensional random variable x is approximated by 2L+1 weighted sigma point as.. A set of points “sigma points” are chosen such that their mean and covariance are x and Pxx respectively. that it is too difficult to tune and only reliable for systems which are almost linear on the time scale of the update intervals. X0 Xi Xi L x x x ( (L ( (L ) Pxx ) i .2 UNSCENTED TRANSFORMATION The unscented transform is a method for calculating the statistics of a random variable which undergo nonlinear transformation. When the system become highly nonlinear.1) ) Pxx ) i L ..CHAPTER 5 UNSCENTED KALMAN FILTER 5. In this new approach ”Unscented Transformation” is used to parameterize mean and covariance which is founded on the intuition that it is easier to approximate a Gaussian distribution than it is to approximate an arbitrary nonlinear function or transformation.L L 1. For example consider a random variable x with dimension L through a nonlinear function y = f (x). These points are applied in the nonlinear function y = f (x) to get the y and Pyy .2 L (5. It is important to note that the sigma points are not chosen in random but rather according to some deterministic algorithm. and x has covariance Pxx and mean x . i 22 .. There are two drawbacks in EKF. EKF is less efficient in estimation.

. 2 ( L K ) L is a scaling parameter and determine the spread of the sigma points around x .3) The mean is given by the weighted average of the transformed points.5) 5. There can be two different implementations. The method instantiate each sigma point through the function y = f (x) resulting in a set of transformed sigma points. The default values for these parameters are 10 4 . 2L Pyy i 0 Wi c [Yi y ][Yi y ]T (5.W0m W0c Wi m / (L / (L Wi c ) ) (1 2 ) (5. i 1.3 THE UKF IMPLEMENTATION This method claims to have superior performance than EKF and directly comparable to second order Gaussian filter and do not restrict to assume that the distributions of noise sources are Gaussian. non-augmented and augmented UKF.L. mean and covariance as Yi f (Xi ) (5.4) The covariance is the weighted outer product of the transformed points.2) 1/ 2( L ). 2L y i 0 Wi mYi (5. ß = 2 and k = 3 ..2L Where Wi m and Wi c are the weights of the mean and covariance calculation associated with ith point. In augmented UKF the state and covariance matrices are augmented with process and 23 . k is secondary scaling parameter and ß corresponds to the prior knowledge of the distribution of X (for Gaussian distributions. ß=2).

This method results in better state estimation when the noise is non-additive but the estimation process will become more computational intensive due to the increase in number of states. . The time update and measurement update are as follows. Time Update Figure 5. U (t )) w(t ) h( X (t ). h is the nonlinear measurement equation. X (t ) Z m (t ) f ( X (t ). U (t )) v (t ) (5.6) Where f is the nonlinear state equation. Non-augmented UKF is implemented in this thesis. Considering the nonlinear state and measurement equations are.measurement noise components. w and v represent process and measurement noise.1 Comparison of mean propagation in EKF. U is the input vector. UKF and Sampling 24 .

k / k 1 ) according to the equations (5. i 0.. Then these sigma points are propagated through the non-linear state equations (5.2 L) are derived from the updated state and covariance matrices where L is the dimension of the state. i 0. where L is the number of states.7) 2L Pk / k 1 i 0 Wi c [ X ki 1/ k 1 Xk 1/ k 1 ][ X ki 1/ k 1 Xk 1/ k 1 ]T (5.8) Measurement Update As in time update.11) .6) to obtain the updated state vector and covariance matrix ( X and k /k 1 P xx .9) The predicted measurement and measurement covariance are calculated with weighted sigma points as follows 2L Zk i 0 Wi m i k (5. The sigma points are propagated through the observation function h as i k h( X ki / k 1 )....17)...10) 2L Pzz .2 L (5.i 0.5) respectively as 2L X i 0 i Wi m X k 1/ k 1 (5.4) and (5.k i 0 Wi c [ i k Z k ][ i k Z k ]T 25 (5.2 L) are calculated from the previous known ( xk 1/ k 1 ) of state vectors according to the equations (4.The time update is as explained in unscented transform where a set of 2L+1 sigma points ( X ki 1/ k 1 . a set of 2L+1 sigma points ( X ki / k 1 .

k i 0 Wi c [ X ki / k 1 X k / k 1 ][ i k Z k ]T (5.14) P /k k P /k k 1 (5.15) 26 .k P .The state-measurement cross covariance matrix is obtained as 2L Pzz . updated state and covariance are respectively Kk 1 P .k Kk (5.13) X k /k X k /k 1 K k (Z k Z k ) T Kk Pzz .k xz zz (5.12) and UKF Kalman Gain.

Ship. Most of the accuracy is determined by the equipment used in the INS and GPS. Most of the applications will strive for the following two characteristics: Continuous and reliable navigation determination (most often position.1 INTRODUCTION Integrated INS and GPS systems are not a new phenomenon. velocity and attitude) Acceptable accuracy level and the possibility to keep the accuracy over time INS can provide the continuous and reliable navigation determination.CHAPTER 6 GPS AND INS INTEGRATION 6. velocity and attitude errors by using the GPS measurements as update to the position and 27 . The accuracy needed can vary a lot between each application. This thesis investigates the INS and GPS integration using unscented kalman filtering approach. but have been used the last couple of decade to a wide span of different systems. and implements it in a matlab program. but their errors are increasing over time due to the integration algorithm they use. Especially the performance from different inertial sensors can vary a lot and low-cost INS systems may result in errors up to many hundreds of meters (and even thousands of meters!) in just a few minutes of standalone mode. The main reason for integrating INS and GPS is therefore done in order to get a system that can achieve both of the above mentioned characteristics. 6. Navigation systems for autonomous car and aircraft may require sub-meter level accuracy while others like car navigation systems only needs 10-30 meter of accuracy in order to achieve its goal. Instead GPS can be used as an aiding system in order to minimize the errors over time by updating the position and velocity as often as possible.2 GPS/INS INTEGRATION The integration of GPS and INS is an efficient way of limit the INS derived position. aircraft and submarine navigations systems are just some of the most common applications.

Therefore GPS-only accuracy is improved on by the integrated solution.velocity whenever it is available. Figure 6. Further the INS can be used to identify and correct GPS carrier phase cycle slips. jamming.1 Integrated INS/GPS Block Diagram 6. Optimal mixing of the INS and the GPS information reduces the effect of GPS errors. velocity and attitude accuracy especially during and after GPS data outage using unscented kalman filtering (UKF) techniques. and accuracy. reliability. It is autonomous and does not rely on any other external aids or on visibility conditions and maintains the availability of navigation solution during GPS outages due to interference.3 ADVANTAGES OF GPS/INS INTEGRATION GPS Aided INS systems have some real advantages in terms of output rate. 28 . The INS provides the full navigation (6 degrees of freedom) state without differentiation. etc. This thesis analyzes an integration of GPS and INS in order to increase the position. The 6 degrees of freedom refer to 3 translational and 3 rotational degrees of freedom.

1 Comparison GPS to INS 29 . Comparison GPS to INS: Advantage Errors are bounded Low cost Disadvantage Low data rate Susceptible to Jamming (intentional.GPS signals could be used to determine accelerations by differentiation or attitude by techniques. unintentional) GPS INS High data rate Both translational and rotational information Self Contained (not susceptible jamming) Unbound errors as a time increase High cost for the quality Table 6.e. The INS provides the navigation solution in real time (i. without latency) at rates higher than may be achievable from a GPS receiver.

1 VELOCITY ERROR COMPARISION 30 .CHAPTER 7 RESULTS 7.

31 .

the weighted mean computation for quaternions is derived in rotational space as a barycentric mean with renormalization and a multiplicative quaternion-error is used for predicted covariance computation of the quaternion because it represents the distance from the predicted mean quaternion. The updates are performed using quaternion multiplication which guarantees that quaternion normalization is maintained in the filter.CHAPTER 8 CONCLUSION 8. 32 . 8. As an additional development. it is modeled as a rotation vector. the square root version of the UKF using for the integrated INS navigation and GPS because it claims to be more stable and less computational intensive. Differing from other quaternion models. velocity and position accuracy can be achieved using the proposed approach for large attitude errors.1 SUMMARY Since the behavior and the accuracy of an INS navigation algorithm are strongly influenced by INS errors. Since the quaternion process noise resulted in increase of the uncertainty in attitude orientation.2 FUTURE SCOPE The information gathered and experiences gained are useful in completing the implementation of INS GPS integration. Numerical simulation results showed that attitude. this thesis develops an integrating Inertial Navigation aided by GPS measurements for large attitude errors using quatemion parameterization of the attitude. velocity and position.

August 2007 [2] SMITH.W. J. \ A Kalman Filter for the Integration of a Low Cost INS and an attitude GPS. and A." Institute of Geodesy and Navigation.. 401_414. R. S. McGrawHill... Andrews..D. V. Eissfeller. 39.S. M. J. ppI20-123.” in Proceedings of the IEEE American Control Conference. [6] Farrell. “The scaled unscented transformation. 2000.." Robotics and Autonomous Systems. Voronov. Kuipers. Proc.. \ Low Cost INS/GPS Integration : Concepts and Testing. [10] Jack B. 477–482. CDROM. USA. John Wiley Sons. Inertial Navigation. no. Virginia: AIAA Education Series.. [13] Yi Y. K. R. TX." Institute of Navigation National Technical Meeting. 46. 2nd ed. [4] Kaplan. Cannon. J. Global Positioning Systems. E. and Integration. The Global Positioning System & Inertial Navigation.V. 1108–111. “A new method for the nonlinear transformation of means and covariances in filters and estimators.G. `Developments in inertial navigation'. 2002. G. Princeton University Press (1999). [7] Wolf. S. pp.. Understanding GPS: Principles and Applications. "INS algorithm using Quatemion model for low cost IMU. Inc. Evanston IL. 3. 2182-2191. of Navigation. An introduction to inertial navigation. J. R. and H. [11] S.A. no. Fort Worth. 2000. California. 2004.. Uhlmann. A. University of Cambridge. and Barth. vol. Rogers.. [9] X.. J. Sept 1986. G. 1996. Woodman. nayak. Weill. of ION GPS 2006.REFERENCES [1] Oliver J. [5] M. Reston. 4. Kong. O. pp. Applied Mathematics in Integrated Navigation Systems. Munich. p. Grejner-Brzezinska (2006b). Tightly-coupled GPS/INS Integration Using Unscented Kalman Filter and Particle Filter. Quaternions and Rotation Sequences.. B. pp. and D. pp. 1998. 221-246. P. 2003. F. [3] R. 33 .. Artech House. Julier. Hein. M.” IEEE Transactions on Automatic Control. Lachapelle.E. MA. Grewal. Durrant-Whyte.. M. [8] Salychev.. [12] S. Julier. vol. L. 45. 2001. Germany.

[19] Farid Gul. 2.” in XIV Congress Brasileiro de Auto. Mobile Multi-Sensor Systems Research Group. pp. Luoyang. Alberta. Fang Jiancheng. The University of Calgary.28. [16] E.. L. 34 . China. vol. 2006. Xiaoji Niu. and Dynamics.. “Attitude determination algorithms.[14] L. Crassidis. 26. Unscented Kalman Filters for Integrated GPS and MEMS Inertial”. 2006.-H. Waldmann. Canada. International Conference on Mechatronics and Automation June 25 . 2002.. "Sigma-point kalman filtering for intagrated GPS and inertial navigation. Calgary.” Journal of Guidance.. [18] Naser El-Sheimy. “Kalman Filter Face-Off: Extended vs. Shin. “Attitude error representations for kalman filtering. May 2005. GPS/SINS Navigation Data Fusion Using Quaternion Model and Unscented Kalman Filter. 2. 750-756. 42. The University of Calgary. 2367–2373. vol.” UCGE Reports Number 20219. and the accuracy of terrestrial navigation with strapdown inertial sensors. Control. 2003. pp. Department of Geomatics Engineering. [15] F. no. Eun-Hwan Shin. Markley. [17] J. computational complexity. Anwar Ali Gaho. Canada. “Estimation Techniques for Low-Cost Inertial Navigation." IEEE Transactions on Aerospace and Electronic Systems. no. pp. 311– 317.

Sign up to vote on this title

UsefulNot useful- 1024_3
- NMEA Revealed
- Global Positioning System
- MX 400 GPS U Manual
- Automotivenavigationsystem 130331084710 Phpapp01 (1)
- Navigational Equipments.docx
- 9.Route Navigation Through Global Positioning System
- Chapter 2 GPS Fundamentals
- RNAV Manual
- Navigation System by Using Gis and Gps
- The Gps Modernization Program Impact in Civil Users
- rinex302
- GPS
- What Is GPS
- Lec8 GPS Intro
- Gca gps
- Global Positioning System Report by Hamza
- Sv8 Preface
- gps 2222
- gnssbooklet[1]
- NMEA Revealed
- MX400_OME
- UNIT 4 GPS
- Global Positioning System
- What Can the Gps Device Portable Product Accomplish for Your Walker
- Gps, Gagan & Lbs
- GNSSINS Integration Methods
- RINEX 2.11
- Chapt11 Gps
- RNAV Trining Manual
- Document

Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

We've moved you to where you read on your other device.

Get the full title to continue

Get the full title to continue reading from where you left off, or restart the preview.

scribd