Adaptive Square-Root Cubature–Quadrature Kalman Particle Filter for satellite attitude determination using vector observations

• A novel adaptive square-root cubature–quadrature particle filter is developed.
• Population size is adaptively selected to maintain 95% confidence level.
• The q-method is used for filter initialization to reduce the computation burden.
• Viability, accuracy and robustness are demonstrated via a Monte Carlo simulation.
• Comparative Analysis of results with other nonlinear filters is performed.

A novel algorithm is presented in this study for estimation of spacecraft׳s attitudes and angular rates from vector observations. In this regard, a new cubature–quadrature particle filter (CQPF) is initially developed that uses the Square-Root Cubature–Quadrature Kalman Filter (SR-CQKF) to generate the importance proposal distribution. The developed CQPF scheme avoids the basic limitation of particle filter (PF) with regards to counting the new measurements. Subsequently, CQPF is enhanced to adjust the sample size at every time step utilizing the idea of confidence intervals, thus improving the efficiency and accuracy of the newly proposed adaptive CQPF (ACQPF). In addition, application of the q-method for filter initialization has intensified the computation burden as well. The current study also applies ACQPF to the problem of attitude estimation of a low Earth orbit (LEO) satellite. For this purpose, the undertaken satellite is equipped with a three-axis magnetometer (TAM) as well as a sun sensor pack that provide noisy geomagnetic field data and Sun direction measurements, respectively. The results and performance of the proposed filter are investigated and compared with those of the extended Kalman filter (EKF) and the standard particle filter (PF) utilizing a Monte Carlo simulation. The comparison demonstrates the viability and the accuracy of the proposed nonlinear estimator.