Battery forensics is a growing research field that is becoming increasingly important with the introduction of hybrid-electric and electric vehicles. The need to correctly diagnose battery condition and predict signs of early failure is well recognised. Many presently used techniques are only applicable to laboratory situations where sensitive measurement is required or where complicated mathematical approaches are needed to assess battery condition. Advanced techniques are explored, such as extended Kalman filtering, to identify the challenges associated with analysis of multi-cell battery modules. Energy-recycling hardware is developed that is capable of efficiently cycling energy to and from cells connected in a series configuration. Switching a supercapacitor-bank-based energy store between series and parallel configurations, coupled with a bidirectional switch-mode power-supply, ensures that maximum energy is retained during the analysis cycle. Extended Kalman filtering (EKF) applied to three different battery models was used to quantify the internal component values of the battery equivalent circuits. The bulk-surface model was determined to be the most appropriate for the Toyota Prius battery modules as the EKF predicted component values converge to stable values, and the recovered voltage trace has a low error. However, the computational complexity when considering 12 series-connected NiMH cells, with their individual component variation with state-of-charge and state-of-health, make the EKF approach unviable. The data harvested during the energy recycling is used to calculate a new effective capacitance measure which relates directly to battery state-of-health. Not only is there a direct relationship between effective capacitance and state-of-health, but the (Q,V) coordinate of maximum effective capacitance on the charge-voltage plane, captured during battery discharge, is able to distinguish clearly between ordinary ageing and catastrophic cell failures.