In order to better control intelligent systems, they can be equipped with physical sensors to know their states. However, this can be costly, leading to cumbersome systems and sometimes obtaining certain information through physical sensors is simply impossible. To remedy this problem, we can use so-called software sensors to deduce information about systems from that provided by essential physical sensors. Software sensors represent an important branch of artificial intelligence (AI), which thanks to computer advances and the multitude of information has become an effective way to solve societal, environmental and human issues. In addition, the systems affected by AI tools become complex because dynamic modelling is often impossible. The objective of the Syroco project is to provide a solution to modeling, using the design of observers (software sensors) by taking robots as intelligent systems. He is interested in two important applications of the ImViA laboratory dealing with complex systems. The first application is the manipulation of objects (potentially deformable) based on multimodal information (modal vision, tactile). The second is medical: endoscopy and microrobotics for targeted delivery of medical treatments.