In this paper, the issue of \(\mathcal {L}_{2}\) - \(\mathcal {L}_{\infty }\) filtering design is investigated for a class of continuous-time Markov jump piecewise-affine systems (MJPWASs). Owing to the randomness of both mode switching and PWA partition switching, the states of both the MJPWASs and the MJPWA filter may be located at different regions. To cope with this “non-synchronous” operation, a novel Markov partition transitions strategy is developed to describe the random position information of regions between the MJPWASs state and the designed filter state. Then, a reachable region path of the MJPWASs state and the filter state, recording only the paths where the switching occurs, is determined to remove the “dead” regions under the Markov partition transitions scenario. Furthermore, by employing the mode-region-dependent Lyapunov function, sufficient conditions are obtained such that the related MJPWA filtering error system is stochastically stable with a guaranteed \(\mathcal {L}_{2}\) - \(\mathcal {L}_{\infty }\) performance index. A pattern of changes on the performance index is captured when adjusting the element values of the partition transition probability matrices (TPMs). Finally, two examples including a tunnel diode circuit system are worked out to show the effectiveness and advantage of the proposed filter design approach.
