Target tracking and state estimation

Target tracking and state estimation

An established communications company sought ways to improve performance of a laser-based gimbal-mounted communications on-the-move product line. The DARPA funded program was intended to push the performance of existing systems by an order of magnitude. In particular, the product’s tracking ability and the ability to recover quickly from periods of outages were major program focuses.

IJK Controls built a detailed simulation that modeled everything from stabilized gimbal and target dynamics, video tracker noise to motor saturation and possible atmospheric effects.

IJK Controls next designed a simple controls solution with a position tracking loop and rate-aiding feedforward. However, an accurate target state estimate was key to meeting performance goals. IJK Controls formulated an Extended Kalman filter that optimally combined range information with tracker offset information in an accurate model of the host and target dynamics and state. The redundant sensor information was fused in a dynamically correct filter to produce an estimate useful for the control loops. The resulting filter equations were translated to real-time code and verified with “software in-the-loop” simulation: the actual flight-ready code was plugged in as the estimator/controller block in the simulation.

Without any changes to the hardware, IJK Controls was able to vastly improve performance in key metrics. IJK Controls delivered a technical white paper with filter formulation and development, production software with API, a system simulation and test plans. IJK Controls also provided flight test support.

IJK applied well-developed theory and experience to a new application, providing performance improvement on a tight schedule. The system involved integrating with two different inertial navigation systems, two different gimbal designs and a suite of custom made optical sensors. The customer was able to focus on complex systems issues, confident that the key target tracking, estimation and controls component would perform to expectations.

Keywords: target tracking, guidance, target estimation, Extended Kalman Filter, kinematics and dynamics, simulation, software in-the-loop, SimuLink, C++, controls, pointing/tracking/stabilization