We propose a novel mechanism to induce self-propulsion and to manipulate ensembles of micron-sized particles with varying composition (e.g., Janus-type, or porous) trapped at solid-liquid interfaces. The physical origin of the process is related to so-called light driven diffusioosmosis introduced recently by our two groups S. Santer (experiment, Potsdam) and O. Vinogradova (theory, Aachen). During irradiation of a solution of azobenzene containing surfactants with focused light, there is a formation of local hydrodynamic flow at a solid/liquid interface. The corresponding hydrodynamic forces are sufficient to swiftly clean the illuminated area from particles trapped at the interface. In preliminary experiments we could verify that when the colloids are turned into Janus particles, their self-propulsion can be initiated in the solution of azobenzene containing surfactant under global/homogeneous illumination with UV light. Additionally, we have found that aggregates of porous but unmodified particles show a pronounced tendency to repel under UV illumination forming a pattern with a unimodal inter-particle separation. The precise physico-chemical mechanisms underlying these phenomena remain, however, unclear. In our project we shall combine theory and experiment in order to understand how the properties of particle surfaces (Janus modification or porosity) in combination with those of the solid substrate (patterning, heterogeneity and anisotropy of, e.g, the zeta-potential) under varying environmental conditions (illumination wave length, salt and surfactant concentration) affect the particle individual motion and collective motion of particle ensembles. With the methodology to be developed in our project we want to establish light-driven hydrodynamics as a useful and versatile tool for investigating collective motion of self-propelled particles and aggregation.