Vibration energy harvesting at micro-scale is considered essential to enable low-power, maintenance free and long-lasting wireless sensor networks (WSNs). However, many problems still limit the efficiency of current vibration-driven generators: narrow bandwidth, low power density, MEMS scaling, inconsistent vibrating sources. The scientific objective of this research project is to overcome the present limitations by proposing an innovative dynamical concept. This is based on the exploitation of multiple-mass velocity amplification and nonlinear dynamical features. An high-efficient micro vibration energy harvester will be implemented by using electrostatic (and/or electromagentic) conversion method. The prototype will be realized by using silicon-base MEMS fabrication facilities at disposal of the hosting institute. In terms of power density and frequency bandwidth performances, this technology is expected to demonstrate an improvement by more than a factor four compared to traditional harvesting generators. Intellectual property is also highly anticipated. The training objective of the Fellow Researcher will be to complement and improve his skills in MEMS manufacturing and characterization by attending at formal and informal technical courses and practicing with the resources provided by the hosting institute. The results will be communicated by the Fellow Researcher in conferences and journal papers. A patent application is also envisaged.