Itzik Klein

A joint filter for formation tracking

An implementation of a joint filter (JF) for unmanned aircraft formation missions is proposed. Such a filter utilizes the fact that velocity of formation members is almost the same in its core structure. In that manner, an advantage of a JF is in situations in which one of the targets is not detected, yet its state can still be estimated.

Feature article: Control theoretic approach to gyro-free inertial navigation systems

In general, an inertial navigation system (INS) consists of a navigation computer and an inertial measurement unit (IMU). Given initial conditions and IMU measurements, the INS provides the position, velocity, and orientation of its carrying platform. The inertial sensors, namely, the accelerometers and gyroscopes (gyros), are part of the IMU. A classical IMU architecture has three accelerometers (to measure specific force) and three gyros (to measure angular velocity) arranged in orthogonal triads.

Analytic Solution of ECV Filter with Position and Velocity Measurements

This paper deals with target tracking of the exponentially correlated velocity filter with position and velocity measurements. Analytic solutions of covariance and gain are presented by closed-form formulas, thus demonstrating the structure of the solution and the performance of the filter. The asymptotic behavior of the filters covariance and gain is presented. The results and the asymptotes are presented in analytic and graphical forms for various values of the measurement noise level and targets correlation time.

GNSS/INS Fusion with Virtual Lever-Arm Measurements

The navigation subsystem in most platforms is based on an inertial navigation system (INS). Regardless of the INS grade, its navigation solution drifts in time. To avoid such a drift, the INS is fused with external sensor measurements such as a global navigation satellite system (GNSS). Recent publications showed that the lever-arm, defined as the relative position between the INS and aiding sensor, has a strong influence on navigation accuracy. Most research in this field is focused on INS/GNSS fusion with GNSS position or velocity updates while considering various maneuvers types. In this paper, we propose to employ virtual lever-arm (VLA) measurements to improve the accuracy and time to convergence of the observable INS error-states. In particular, we show that VLA measurements improve performance even in stationary conditions. In situations when maneuvering helps to improve state observability, VLA measurements manage to gain additional improvement in accuracy. These results are supported by simulation and field experiments with a vehicle mounted with a GNSS and an INS.

Composite measurement from asynchronous LOS and unknown sensor position

Target tracking with multiple asynchronous passive sensors is addressed. In this paper, a method is derived for forming composite measurements when the LOS measurements from multiple sensors are asynchronous and the position of one of the passive sensors is unknown. Using the composite measurements, the target position, velocity and unknown sensor location are estimated. In addition, analytical general observability conditions are derived.

Analytic steady-state solution of fine alignment with velocity measurements

Fine alignment process is used after the initialization procedure to improve the accuracy of the attitude to a desired value. In stationary conditions, zero velocity updates are usually used to achieve that goal. In this paper, we derive closed form solutions for the error state covariance of the fusion process. With such closed form solutions at hand, insight can be gained into the parameters involved in the fine alignment process. In addition, a numerical simulation is used to verify the analytical expressions.

Zero Δv Solution to the Angles-Only Range Observability Problem during Orbital Proximity Operations

During orbital proximity operations, research has shown that angles-only navigation during coasting flight suffers from a lack of range observability. To circumvent this deficiency, previous research has required a prior information on the target geometry or the implementation of special translational maneuvers. This paper shows that the range observability problem during coasting flight can be solved by properly including the offset of the camera from the vehicle center-of-mass in the problem formulation, and by applying appropriate vehicle rotations. Range observability without translational maneuvers (zero Δv) or a priori knowledge of the target geometry is clearly demonstrated using a pseudo 6 degree-of-freedom simulation. Results for v-bar station-keeping, flyby orbits, and circumnavigation (football) orbits are presented.