We use a computationally efficient technique of Logistic Continuous Wavelet transform (CWT) to analyze patent data for Switzerland, Germany, USA, and Brszil for the period 1980-2000. We found that patent growth dynamics follows the dynamics of innovation system synergy in the framework of Triple Helix model of innovations where observed non-linear actors' interactions are provided by biased information exchange between heterogenious actors. Suggested approach reveals the latent trend structure in patent and innovation dynamics and may help policymakers identify the potential drivers of patent and innovation activity and form informed policy for boosting innovation development. The paper also privides a foundation for future research in differnt fields studying complex systems of interacting heterogenious agents.