Turbo-TSWLS: enhanced two-step weighted least squares estimator for TDOA-based localisation

The two-step weighted least squares (TSWLS) estimator is a well-known linear approach for TDOA-base localisation. It has been applied to various applications and is also widely used as a benchmark in subsequent studies. In this reported work, it is revealed that the approximation in the weighting matrix of the second-step WLS is a crucial factor to the performance of TSWLS, and the performance of TSWLS is improved in a simple and computational attractive way through reference anchor selection and co-ordinate rotation. The enhanced method is called turbo-TSWLS. Simulation results show that turbo-TSWLS has a much higher noise threshold before performance breaks away from the CRLB and the bias can be reduced effectively.