This thesis is concerned with multi-target localisation and tracking in passive and Multiple-Input Multiple-Output (MIMO) radar. First, it is highlighted that in passive radar, the problem of location ambiguities has to be resolved for multi-target localisation. By considering a passive radar employing distributed Small Aperture Arrays (SAAs) as Receivers (Rxs), an algorithm utilising the signals received by the overall Large Aperture Array (LAA) is proposed for identifying the targets from a set of ambiguous locations provided by Direction Of Arrival (DOA) localisation algorithms.
Then, in bistatic MIMO radar, subspace-based algorithms are proposed in conjunction with an array manifold extender for multi-target localisation in a Coherent Processing Interval (CPI) where targets remain stationary over multiple Pulse Repetition Intervals (PRIs). With the removal of mutual target interference, the proposed algorithms support estimating and unambiguously associating the delay, DOA, Direction Of Departure (DOD), Doppler frequency, and complex path coefficient for different targets.
Next, the existence of adverse clutter is considered in monostatic MIMO radar. By exploiting the low-rank property of the clutter at the output of a Doppler-spatial manifold extender, a clutter subspace estimation approach is proposed for clutter suppression via null space projection. As a clutter pre-processor, the proposed approach enables the use of various localisation and tracking algorithms that rely on the white noise assumption.
Finally, for adaptively localising targets that are non-stationary across different CPIs, an H∞ approach is proposed for robust range, DOA, and velocity tracking. This is realised with the proposal of two manifold extenders that combine the slow-time and fast-time dimensions of the MIMO radar received signal. Importantly, the proposed H∞ approach is directly applicable to cluttered signals with unknown clutter statistics.